Richard Fuller, Donald Gallant, Donald W. Goo-Alcoholism & PDF

RECENT DEVELOPMENTS IN
ALCOHOLISM
VOLUME 13
ALCOHOL AND VIOLENCE
RECENT
DEVELOPMENTS IN
Edited by
MARC GALANTER
New York University School of Medicine
New York, New York
Associate Editors
HENRI BEGLEITER, RICHARD DEITRICH,
RICHARD FULLER, DONALD GALLANT,
DONALD GOODWIN, EDWARD GOTTHEIL,
ALFONSO PAREDES, MARCUS ROTHSCHILD,
and DAVID VAN THIEL
Assistant Editor
DEIRDE WINCZEWSKI
An Official Publication of the American Society of Addiction Medicine

and the Research Society on Alcoholism.
This series was founded by the National Council on Alcoholism.
ALCOHOLISM
VOLUME 13
ALCOHOL AND VIOLENCE
Epidemiology
Neurobiology
Psychology
Family Issues
KLUWER ACADEMIC PUBLISHERS

NEW YORK / BOSTON / DORDRECHT / LONDON / MOSCOW
eBook ISBN: 0-306-47141-8
Print ISBN: 0-306-45358-4
©2002 Kluwer Academic Publishers

New York, Boston, Dordrecht, London, Moscow
Print ©1997 Kluwer Academic / Plenum Publishers

New York
All rights reserved
No part of this eBook may be reproduced or transmitted in any form or by any means, electronic,
mechanical, recording, or otherwise, without written consent from the Publisher
Created in the United States of America
Visit Kluwer Online at: http://kluweronline.com

and Kluwer's eBookstore at: http://ebooks.kluweronline.com
Editorial Board
Chair: Chair Emeritus and Founder:
James D. Beard, Ph.D. Charles S. Lieber, M.D.
Irina Anokhina, Ph.D. Adrienne S. Gordon, Ph.D. Yasuhiko Mizoi, M.D.
Howard C. Becker, Ph.D. Kathleen A. Grant, Ph.D. Sara Jo Nixon, Ph.D.
Marlene O. Berman, Ph.D. Consuelo Guerri, Ph.D. Roger Nordmann, M.D., Ph.D.
Stefan Borg, M.D. Victor Hesselbrock, Ph.D. Stephanie S. O’Malley, Ph.D.
Michael E. Charness, M.D. Paula L. Hoffman, Ph.D. Adolf Pfefferbaum, M.D.
Allan C. Collins, Ph.D. Hiromasa Ishii, M.D. Tamara J. Phillips, Ph.D.
Christopher L. Cunningham, Ph.D. Thomas R. Jerrells, Ph.D. Barry J. Potter, Ph.D.
Philippe A.J. De Witte, Ph.D. Harold Kalant, M.D., Ph.D. Mikko Salaspuro, M.D.
Michael J. Eckardt, Ph.D. Ting-Kai Li, M.D. Jalie A. Tucker, Ph.D.
V. Gene Erwin, Ph.D. John M. Littleton, Ph.D. Joanne Weinberg, Ph.D.
Daniel Flavin, M.D. Barbara S. McCrady, Ph.D. Gary S. Wand, M.D.
H. Werner Goedde, M.D. Robert O. Messing, M.D. James R. West, Ph.D.
Research Society on Alcoholism American Society of

Addiction Medicine
President:
Ivan Diamond, M.D., Ph.D. President:
David E. Smith, M.D.
Vice President:
Edward P. Riley, Ph.D. President-elect:
G. Douglas Talbott, M.D.
Secretary:
Tina Vanderveen, Ph.D. Secretary:
Marc Galanter, M.D.
Treasurer:
Victor Hesselbrock, Ph.D. Treasurer:
James W. Smith, M.D.
lmmediate Past President:
R. Adron Harris, Ph.D. Immediate Past President:
Anne Geller, M.D.
Publications Committee Chair:
James D. Beard, Ph.D.
This page intentionally left blank.
Contributors
Nancy Asdigian, Family Research Laboratory, University of New Hampshire,
Durham, New Hampshire 03824
Ronet Bachman, Department of Sociology, University of Delaware, Newark,

Delaware 19716
D. Caroline Blanchard, Bekesy Laboratory of Neurobiology, Department of
Anatomy and Reproductive Biology, University of Hawaii at Manoa, Hono-
Iulu, Hawaii 96822
Robert J. Blanchard, Department of Psychology, Bekesy Laboratory of Neu-
robiology, University of Hawaii at Manoa, Honolulu, Hawaii 96822
Jenia Bober Booth, Laboratory for the Study of Addictions and UCLA Drug
Abuse Research Center, West Los Angeles Veterans Administration, Los
Angeles, California 90073
Brad J. Bushman, Department of Psychology, Iowa State University, Ames,
Iowa 50011-3180
Cheryl J. Cherpitel, Alcohol Research Group, Western Consortium for Public
Health, Berkeley, California 94709
Mark A. Cohen, Owen Graduate School of Management, and Vanderbilt
Institute for Public Policy Studies, Vanderbilt University, Nashville, Tennes-
see 37212
James J. Collins, Health and Social Policy Division, Research Triangle Insti-
tute, Research Triangle Park, North Carolina 27709-2194
Joseph F. DeBold, Department of Psychology, Tufts University, Medford,
Massachusetts 02155
Richard A. Deitrich, Department of Pharmacology, Alcohol Research Center,
University of Colorado Health Sciences Center, Denver, Colorado 80262
M. Elena Denison, Laboratory for the Study of Addictions and UCLA Drug
Abuse Research Center, West Los Angeles Veterans Administration, Los An-
geles, California 90073
vii
viii Contributors
James H. Derzon, Vanderbilt Institute for Public Policy Studies, Vanderbilt

University, Nashville, Tennessee 37212
William R. Downs, Center for the Study of Adolescence, University of

Northern Iowa, Cedar Falls, Iowa 50614
Annemoon M. M. van Erp, Department of Psychology, Tufts University,

Medford, Massachusetts 02155
Richard K. Fuller, Division of Clinical and Prevention Research, National

Institute on Alcohol Abuse and Alcoholism, Bethesda, Maryland 20892-7003
Edward Gottheil, Department of Psychiatry and Human Behavior, Thomas

Jefferson University, Philadelphia, Pennsylvania 19107
Ellen F. Gottheil, Department of Psychiatry and Behavioral Sciences, Univer-

sity of Washington Medical School, Seattle, Washington 98195
J. Andy Henrie, Department of Chemistry, University of Hawaii at Manoa,

Honolulu, Hawaii 96822
J. Dee Higley, Laboratory of Clinical Studies, Primate Unit, National Institute

on Alcohol Abuse and Alcoholism, Poolesville, Maryland 20837
Glenda Kaufman Kantor, Family Research Laboratory, University of New

Hampshire, Durham, New Hampshire 03824
Larry A. Kroutil, Health and Social Policy Division, Research Triangle Insti-
W. Vernon Lee, Penn Recovery Systems, Philadelphia, Pennsylvania 19104-1953
Markku Linnoila, Laboratory of Clinical Studies, Division of Intramural Clin-

ical and Biological Research, National Institute on Alcohol Abuse and Alco-
holism, Bethesda, Maryland 20892-1256
Mark W. Lipsey, Vanderbilt Institute for Public Policy Studies, Vanderbilt

Eugene Maguin, Research Institute on Addictions, Buffalo, New York 14203
Susan Ehrlich Martin, Prevention Research Branch, Division of Clinical and

Prevention Research, National Institute on Alcohol Abuse and Alcoholism,
Rockville, Maryland 20892-7003
Contributors ix
Klaus A. Miczek, Departments of Psychology, Psychiatry, and Pharmacol-

ogy, Tufts University, Medford, Massachusetts, 02155
Brenda A. Miller, Research Institute on Addictions, Buffalo, New York 14203
MarIee Moore-Gurrera, Health and Social Policy Division, Research Triangle

Institute, Research Triangle Park, North Carolina 27709-2194
Alfonso Paredes, Laboratory for the Study of Addictions and UCLA Drug
Abuse Research Center, West Los Angeles Veterans Administration, Los An-
geles, California 90073
Judith Roizen, Institute of Population Studies, University of Exeter, Exeter

EX4 6DT, England
E. Joyce Roland, Health and Social Policy Division, Research Triangle Insti-
Walter Tornatzky, Department of Psychology, Tufts University, Medford,

Massachusetts 02155
Matti Virkkunen, Department of Psychiatry, Helsinki University Central

Hospital, Helsinki 00180, Finland
Stephen P. Weinstein, Department of Psychiatry and Human Behavior,

Thomas Jefferson University, Philadelphia, Pennsylvania 19107
Helene Raskin White, Center for Alcohol Studies, Rutgers University, Pis-
cataway, New Jersey 08855-0969
David B. Wilson, Vanderbilt Institute for Public Policy Studies, Vanderbilt

Errol Yudko, Department of Psychology, University of Hawaii at Manoa,

Honolulu, Hawaii 96822
Preface
From the President of the Research Society on Alcoholism
On behalf of the Research Society on Alcoholism, I am pleased to introduce

this thirteenth volume of Recent Developments in Alcoholism about alcohol and
violence. Current concepts are presented in well-organized sections that fo-
cus on epidemiology, neurobiology, psychology, and family issues. It is be-
coming increasingly clear that age, gender, socioeconomic circumstances, and
genetics affect aggressive behavior and vulnerability to alcoholism. This vol-
ume contains up-to-date discussions of these issues. Indeed, the information
presented here will help all alcohol researchers to identify biological and
social factors that contribute to the comorbidity of alcoholism and aggression.
The editors and associate editors should be congratulated for bringing togeth-
er such important information. This volume will be a valuable resource for
investigators and therapists alike.
Ivan Diamond M.D., Ph.D.

President, Research Society on Alcoholism
From the President of the American Society of Addiction Medicine
The American Society of Addiction Medicine is honored to continue its co-

sponsorship with the Research Society on Alcoholism of Recent Developmen ts
in Alcoholism. The topic of alcohol and violence is a particularly timely one,
given the growing number of studies that are examining the relationship
between the two. These studies are consistent with the hypothesis that alco-
hol can increase aggression and contribute to both domestic and criminal
violence. Intoxicating blood levels of alcohol have been found to be especially
prevalent in those injured in fights and assaults. Despite this growing body of
evidence, there still is a great deal to be learned about what causes alcoholics
to become aggressive and violent. Additional resesarch is also needed relative
to self-directed violence and suicide in relation to alcohol. This volume on
alcohol and violence ranges from epidemiology and neurobiology to psychol-
ogy and family issues and will make an important addition to the body of
knowledge relative to this complex phenomenon.
David E. Smith, M.D.

President, American Society of Addiction Medicine
xi
Contents
I. Epidemiology
Richard K. Fuller, Section Editor
Overview
Richard K. Fuller
Chapter 1
Epidemiological Issues in Alcohol-Related Violence

Judith Roizen
1. Defining the Problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
1.1. Introduction: The Six Dilemmas . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
1.2. Definitions of Violence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
1.3. Measurement of Alcohol Use and Alcohol Problems . . . . . . . . 14
2. An Overview of Methodological Problems in Research on Alcohol
and Violence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
2.1. Event-Based Research . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
2.2. Studies of the General Population . . . . . . . . . . . . . . . . . . . . . . . . . 17
3. Evidence on Alcohol and Rape . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
3.1. Event-Based Research on Alcohol and Rape . . . . . . . . . . . . . . . . 18
3.2. Studies of Prison Offenders . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
3.3. Rape in the General Population . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
4. The Recent Work of Pernanen . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
4.1. Pervasiveness of Alcohol . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
4.2. Differential Risk of Alcohol-Involved Violence . . . . . . . . . . . . . . 32
4.3. Selected Findings on Alcohol and Violence from Pernanen's
Work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
4.4. Contributions of Pernanen's Recent Work to Theoretical
Debate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
4.5. Alcohol and Severity of Choice of Acts and Outcomes . . . . . . 34
4.6. Indiscrimination in Acts of Violence and Alcohol . . . . . . . . . . . 34
5. The Future of Epidemiological Research on Alcohol and Violence 36
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
xiii
xiv Contents
Chapter 2
The Relationship of Alcohol to Injury in Assault Cases
Susan Ehrlich Martin and Ronet Bachman
1. Research on Alcohol in Human Violence . . . . . . . . . . . . . . . . . . . . . . . 42
1.1. Correlational Studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
1.2. Experimental Studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
1.3. Limitations of Existing Studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
2. The NCVS Study . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
2.1. Sample . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
2.2. Measures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
2.3. Analytic Procedures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
2.4. Findings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
2.5. Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
3. Future Research Directions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54
Chapter 3
Alcohol and Spouse Abuse: Ethnic Differences
Glenda Kaufman Kantor
1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
2. Drinking and Violence in Ethnic Groups . . . . . . . . . . . . . . . . . . . . . . . 59
2.1. Problem Drinking in African-American Men . . . . . . . . . . . . . . . . 59
2.2. Wife Abuse in African-American Families . . . . . . . . . . . . . . . . . . 60
2.3. Moderators of Alcohol-Related Wife Assaults in African-
American Families . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60
2.4. Problem Drinking in Hispanic-American Men . . . . . . . . . . . . . . 61
2.5. “Machismo” Drinking and Wife Assaults . . . . . . . . . . . . . . . . . . . 62
2.6. Ethnic Comparisons of Wife Abuse . . . . . . . . . . . . . . . . . . . . . . . . 62
2.7. Wife Assault Patterns among Hispanic Americans . . . . . . . . . . 63
2.8. Drinking and Wife Assaults in Multiethnic Groups . . . . . . . . . . 63
2.9. Moderators of Alcohol-Related Wife Assaults in Hispanic-
American Families: Acculturation . . . . . . . . . . . . . . . . . . . . . . . . . . 64
2.10. The 1992 National Alcohol and Family Violence Survey . . . . . 66
2.11. Evidence from Community and Clinical Samples of
Multiethnic Groups . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71
3. Summary and Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72
3.1. Examining the Theoretical Framework . . . . . . . . . . . . . . . . . . . . . 72
3.2. Does Alcohol Influence Wife Beating Equivalently across
Ethnic Groups? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72
Contents xv
3.3. The Importance of Race and Social Structure . . . . . . . . . . . . . . . 73

3.4. The Importance of Gender Roles . . . . . . . . . . . . . . . . . . . . . . . . . . 74
3.5. The Importance of Culture and Acculturation . . . . . . . . . . . . . . . 74
3.6. Implications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75
Chapter 4
Longitudinal Perspective on Alcohol Use and Aggression
during Adolescence
Helene Raskin White
1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81
2. Explanatory Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82
3. The Extent of Alcohol Use and Aggression in Adolescence . . . . . . . 84
4. The Association between Alcohol Use and Aggression
in Adolescence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88
4.1. Acute Incidents of Alcohol-Related Aggression . . . . . . . . . . . . . 88
4.2. Associations between Patterns of Alcohol Use and
Aggression . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90
4.3. Developmental Trends . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92
5. A Common-Cause Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95
6. Preventing Alcohol Abuse and Aggression . . . . . . . . . . . . . . . . . . . . . 97
7. Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99
Chapter 5
Alcohol and Violence-Related Injuries in the Emergency Room
Cheryl J. Cherpitel
1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105
2. Prevalence Estimates of Blood Alcohol Concentration . . . . . . . . . . . . 106
3. Self-Reports of Alcohol Consumption Prior to Injury . . . . . . . . . . . . 109
4. Drinking Patterns and Alcohol-Related Problems . . . . . . . . . . . . . . . . 111
5. Regional Comparisons of ER and Coroner Data . . . . . . . . . . . . . . . . . 113
6. Limitations to ER Studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116
xvi Contents
II. Neurobiology
Richard A. Deitrich, Section Editor
Overview
Richard A. Deitrich
Chapter 6
Emerging Themes in Preclinical Research on Alcohol and Aggression
Errol Yudko, D. Caroline Blanchard, J. Andy Henrie,
and Robert J. Blanchard
1. Alcohol Effects on Aggression and Violence . . . . . . . . . . . . . . . . . . . 124
2. Effects of Aggression, Violence, and Other Stressors on Alcohol
Use and Abuse . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 125
3. Does Administration of Alcohol Increase Aggression
in Animal Models? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 126
4. Alcohol and the Recipient of Attack . . . . . . . . . . . . . . . . . . . . . . . . . . . . 127
5. Does Social Stress Enhance Voluntary Alcohol Consumption (VAC)? 128
6. Stress and Substance Abuse of Other Aggression-Impacting
Substances . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 130
7. The Relationship of Aggression to the Predisposition to VAC . . . . 131
8. Is an Anxiolytic or Inhibition-Reducing Property of Alcohol an
Important Component of the Relationship between Stress,
Voluntary Alcohol Consumption, and Aggression? . . . . . . . . . . . . . . 132
9. Relationship of Preclinical Studies of Alcohol Effects to Human
Alcohol-Aggression Phenomena . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133
References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 134
Chapter 7
Alcohol, GABA,-Benzodiazepine Receptor Complex, and Aggression
Klaus A. Miczek, Joseph F. DeBold, Annemoon M. M. van Erp,
and Walter Tornatzky
1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139
2. Alcohol and Aggressive Behavior in Animals: Ethological Analysis 141
2.1. Individual Differences in Alcohol Effects on Aggressive
Behavior . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 141
2.2. Animal Models of Alcohol Self-Administration and
Aggression . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 150
3. Pharmacological Evidence for Interactions between Alcohol
and γ-Aminobutyric Acid-A Receptors and Aggression . . . . . . . . . . 154
3.1. Benzodiazepine Receptor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 154
Contents xvii
3.2. Alcohol Interactions with Neurosteroids . . . . . . . . . . . . . . . . . . . 159

4. Neurochemical Mechanisms for Alcohol–GABAA Interactions
and Aggression . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 162
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 164
Chapter 8
Serotonin in Early-Onset Alcoholism
Matti Virkkunen and Markku Linnoila
1. Subgrouping Alcoholics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 173
2. Type 2 Alcoholism and Early-Onset Alcoholism . . . . . . . . . . . . . . . . . 174
3. Antisocial Personality Disorder and Early-Onset Alcoholism . . . . . 175
4. Is Early-Onset Aggressivity Specifically Conducive to the
Development of Early-Onset Alcohol Abuse? . . . . . . . . . . . . . . . . . . . 176
5. Psychobiology of Early-Onset Alcoholism . . . . . . . . . . . . . . . . . . . . . . 177
6. Early-Onset Alcoholism and Antisocial Personality Disorder
Are Associated with Reduced Brain Serotonin Turnover Rate
and Transmission . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 178
6.1. Serotonin Turnover Rate (CSF 5-HIAA) . . . . . . . . . . . . . . . . . . . 178
6.2. Serotonin Receptor Challenges in Early- versus Late-Onset
Male Alcoholics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 179
6.3. Abnormal Tryptophan-Large Neutral Amino Acid Ratio
in Plasma . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 180
6.4. Platelet Monoamine Oxidase . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 180
6.5. Uptake of Serotonin to Platelets . . . . . . . . . . . . . . . . . . . . . . . . . . . 182
6.6. Molecular Genetic Findings Associated with Early-Onset
Alcoholism . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 183
6.7. Brain Neuroimaging Findings in Alcoholism . . . . . . . . . . . . . . . . 184
7. Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 184
References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 184
Chapter 9
A Nonhuman Primate Model of Excessive Alcohol Intake:
Personality and Neurobiological Parallels of Type I-
and Type II-Like Alcoholism
J. Dee Higley and Markku Linnoila
1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 192
2. Why Study Nonhuman Primates? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 193
3. Difficulties Producing a Nonhuman Primate Model
of Alcohol Abuse. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 193
4. Human and Nonhuman Primate Parallels in Alcohol Use . . . . . . . . . 194
5. Methodology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 194
6. Biological and Behavioral Measures . . . . . . . . . . . . . . . . . . . . . . . . . . . . 195
xviii Contents
7. Studies of Interindividual Differences . . . . . . . . . . . . . . . . . . . . . . . . . . 196

8. Studies Investigating Features of Type I Alcoholism . . . . . . . . . . . . . 196
9. Serotonin and Type II Alcoholism . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 199
10. Reduced CNS Serotonin Functioning as a Long-Term Enduring
Trait . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 200
11. Impaired CNS Serotonin Functioning and High Alcohol
Consumption . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 201
12. Serotonin and Violence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 202
13. Reduced CNS Serotonin Functioning and Antisociallike Behavior 204
14. Low CSF 5-HIAA as a Risk Factor for Social Ostracism
and Early Mortality . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 206
15. CSF 5-HIAA, Frontal Cortex Serotonin Functioning, and
Impulsive Behaviors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 207
16. Etiology of Low CSF 5-HIAA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 208
16.1. Genetic Influences . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 208
16.2. Environmental Influences . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 208
17. Summary and Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 211
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 212
III. Psychology
Alfonso Paredes, Section Editor
Overview
Alfonso Paredes
Chapter 10
Effects of Alcohol on Human Aggression:
Validity of Proposed Explanations
Brad J. Bushman
1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 227
2. Prototypical Methods of Administering Alcohol and Measuring
Aggression in Experimental Studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . 229
2.1. Administration of Alcohol . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 229
2.2. Measurement of Aggression . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 230
3. Explanations of Alcohol-Related Aggression . . . . . . . . . . . . . . . . . . . . 233
3.1. Physiological Disinhibition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 233
3.2. Expectancy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 234
3.3. Indirect Cause . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 234
4. Present Review . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 234
Contents xix
5. Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 235
5.1. Literature Search Procedures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 235
5.2. Criteria for Relevance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 235
5.3. Coding Frame . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 236
5.4. Meta-Analytic Procedures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 236
6 . Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 236
6.1. Sex Differences in Intoxicated Aggression . . . . . . . . . . . . . . . . . . 236
6.2. Measurement of Aggression . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 237
6.3. Validity of Proposed Explanations of Alcohol-Related
Aggression . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 237
7 . Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 239
8. Appendix . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 241
8.1. Participant Characteristics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 241
8.2. Experiment Characteristics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 241
8.3. Primary Study Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 242
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 242
Chapter 11
Is There a Causal Relationship between Alcohol Use and Violence?
A Synthesis of Evidence
Mark W. Lipsey, David B. Wilson, Mark A. Cohen,
and James H. Derzon
1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 245
1.1. The Nature of the Causal Question . . . . . . . . . . . . . . . . . . . . . . . . 246
1.2. The Complexities of Empirically Demonstrating a Causal Link 248
2. Experimental Approaches . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 251
2.1. Animal Studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 251
2.2. Human Studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 253
3. Correlational Approaches . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 261
3.1. Individual-Level Studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 262
3.2. Macro-Level Studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 273
4. Overall Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 277
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 278
Chapter 12
Alcohol and Cocaine Interactions and Aggressive Behaviors
M. Elena Denison, Alfonso Paredes, and Jenia Bober Booth
1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 283
2. Alcohol and Violence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 284
3. Cocaine and Violence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 284
xx Contents
4. Drug Use Status and Deviant Behaviors: Results of the Study . . . . 286
4.1. Characteristics of the Study Sample . . . . . . . . . . . . . . . . . . . . . . . . 286
4.2. Data Collection Methodology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 288
4.3. Cocaine Use Patterns . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 288
4.4. Other Drug Use . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 289
4.5. Use of Alcohol . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 290
4.6. Socially Disruptive or Deviant Behavior during Adolescence 291
4.7. Social Deviance and Violence during Adulthood . . . . . . . . . . . . 293
4.8. Interactions of Cocaine and Alcohol Use . . . . . . . . . . . . . . . . . . . 296
4.9. Violence in Cocaine Addicts according to Amount
of Alcohol Used . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 297
5. Summary and Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 298
6. Comments and Research Implications . . . . . . . . . . . . . . . . . . . . . . . . . . 301
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 302
IV. Family Issues

Edward Gottheil and Ellen F. Gottheil, Section Editors
Overview
Edward Gottheil and Ellen F. Gottheil
Chapter 13
When Women Are under the Influence: Does Drinking or Drug Use
by Women Provoke Beatings by Men?
Glenda Kaufman Kantor and Nancy Asdigian
1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 315
2. What Are the Theoretical Mechanisms? . . . . . . . . . . . . . . . . . . . . . . . . 317
2.1. Theories of Victimization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 317
2.2. Theories of Intoxicant-Induced Aggression . . . . . . . . . . . . . . . . . 318
2.3. Mechanisms of Intoxicant-Induced Victimization . . . . . . . . . . . . 318
3. How Common Is Drinking by Both Parties? . . . . . . . . . . . . . . . . . . . . 322
4. Empirical Studies of Wife Assault and Intoxication
by Female Victims . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 323
4.1. Evidence from General Population Surveys . . . . . . . . . . . . . . . . . 324
4.2. Evidence from Clinical Populations . . . . . . . . . . . . . . . . . . . . . . . . 327
5. Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 329
5.1. The Role of Women’s Intoxication in Husband-to-Wife
Violence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 329
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 332
Contents xxi
Chapter 14
How Far Have We Come? A Critical Review of the Research
on Men Who Batter
W. Vernon Lee and Stephen P. Weinstein
1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 337
2. Causes of Domestic Violence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 338
3. Characteristics of Male Batterers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 339
3.1. Personality Behavioral Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 340
3.2. Social Learning Approaches . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 345
3.3. Psychodynamic Explanations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 346
4. Substance Abuse and Battering . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 347
5. Treatment Interventions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 350
6. Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 353
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 353
Chapter 15
Alcohol‚ Drugs‚ and Violence in Children’s Lives
Brenda A. Miller, Eugene Maguin, and William R. Downs
1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 357
2. Perpetrator’s Substance Use/Abuse and Physical and Sexual
Abuse of Children . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 358
2.1. Parental Substance Problems and Perpetration of Child
Physical Abuse . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 358
2.2. Perpetrator AOD Problems and the Sexual Abuse of Children 360
2.3. Explanations for the Link between Perpetrators' Substance
Problems and Violence toward Children . . . . . . . . . . . . . . . . . . . 362
3. Childhood Victimization as Antecedent to Later Substance Use . . . 365
3.1. Childhood Physical Abuse Prior to Substance Use . . . . . . . . . . 365
3.2. Child Sexual Abuse Prior to Substance Use . . . . . . . . . . . . . . . . . 367
3.3. Explanations of the Linkage from CSA and Parent-to-Child
Violence to AOD Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 369
4. Methodological Considerations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 372
4.1. Definitions and Measurement of Violence toward Children . . 372
4.2. Controversies Regarding Childhood Memories of Violence . . . 375
4.3. Ethical and Legal Issues . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 376
5. Future Directions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 378
5.1. Considerations for Research Designs and Data Analyses . . . . . 378
5.2. Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 379
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 380
xxii Contents
Chapter 16
Issues in the Linkage of Alcohol and Domestic Violence Services
James J. Collins, Larry A. Kroutil, E . Joyce Roland,
and Marlee Moore-Gurrera
1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 387
2. Role of Alcohol in Domestic Violence . . . . . . . . . . . . . . . . . . . . . . . . . . 388
3. Explanation of the Alcohol-Domestic Violence Relationship . . . . . . 390
4. Examples of Current Linkage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 392
5. Linking Services . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 394
6. Models for Service Linkage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 396
7. Next Steps . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 400
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 401
Contents of Previous Volumes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 407
Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 423
I
Epidemiology
Richard K. Fuller, Section Editor

Overview
Richard K. Fuller
It is both appropriate and timely that this volume of Recent Developments in

Alcoholism is devoted to alcohol and violence, since violence is regarded by
many as the most important problem facing our society today. Daily, we hear
or read stories of domestic violence, child abuse, homicides, and gang vio-
lence; and the public perception is that alcohol often has an important role in
these events. Epidemiological studies of perpetrators and victims of interper-
sonal violence indicate that alcohol has often been consumed shortly prior to
the violent act. However, drinking is a common social activity for most adult
Americans, particularly young males who are those most likely to commit a
violent act; and “ therefore it is to be expected that drinking will often occur
proximate to violence.”1 A key question is whether alcohol use is coincidental
to aggression or whether it plays a causal role in aggression. The consistency
of the relationship between alcohol and aggression across many studies sug-
gests that it is a real association and not spurious. Despite many studies, our
understanding of the precise role of alcohol in aggression and violence is in its
infancy.
At this point in time, it appears that alcohol is one of several factors
leading to violence.1 The other types of factors contributing to aggression
appear to be biochemical, genetic, psychological, and environmental in na-
ture. Other sections of this volume discuss these factors. Another key ques-
tion is how does the mix of all these factors increase aggression in some
individuals but not in others?2
The public health model may be useful in answering this question. The
premise of the public health model is that alcohol-related problems arise from
an interaction of individual (host) variables, alcohol (agent), and environmen-
Richard K. Fuller • Division of Clinical and Prevention Research, National Institute on Alcohol
Abuse and Alcoholism, Bethesda, Maryland 20892-7003.
Recent Developments in Alcoholism, Volume 13: Alcoholism and Violence, edited by Marc Galanter.
Plenum Press, New York, 1997
3
4 I • Epidemiology
tal factors. Host variables include traits and life experiences that influence the
individual's vulnerability to the effects of alcohol. Alcohol (the agent) varies
by type, pattern of drinking, and availability. Environmental factors include
interpersonal interactions, social milieus, cultural norms surrounding both
the use of alcohol and aggressive behavior, and factors regulating a person's
exposure to alcohol. Epidemiology is one of the disciplines that generates
knowledge for the public health model.
In Chapter 1, Dr. Roizen discusses that epidemiological studies of alcohol
and violence can be divided into two categories: event-based studies and
general population studies. Event-based studies are based on samples of
perpetrators or victims of the violent event. In her discussion of event-based
studies, Dr. Roizen focuses on the role of alcohol in one type of violent act,
i.e., rape. An example of a general population study is the one described in
Chapter 2, by Drs. Martin and Bachman. General population studies use
samples drawn from a community, county, state, region, or country to deter-
mine the prevalence of alcohol-related violent behavior. Dr. Roizen’s discus-
sion indicates that both of these types of studies have limitations.
While alcohol is frequently consumed prior to acts of violence, the lack of
a comparison group in some event-based studies makes the interpretation of
the results difficult, if not meaningless. Another methodological problem for
event-based studies is the selection of an appropriate control group. Is the
person who lives next door an appropriate “control” for a habitual child
molester?3
Other methodological criticisms of event-based studies include use of
samples of convenience, which results in sample selection bias, and subjects
in these studies may not be representative of all violent offenders, victims, or
their drinking patterns. Furthermore, since these studies are cross-sectional,
they do not allow causal inferences about the relationship between alcohol
and aggression.
As Roizen points out in Chapter 1, population surveys also have limita-
tions. Population samples, even good ones, may miss the small segment of
the population given to violence. A limitation to both types of studies is that
those interviewed may have selective recall or be reluctant to admit to deviant
acts. Dr. Roizen selects one population study to discuss in detail. She selected
the study done in Thunder Bay, Ontario, by Kai Pernanen because of its
methodological rigor.
Another limitation of epidemiological studies is that if there are a number
of variables related to aggression, the strength of the role of alcohol in the web
of causation leading to aggressive behavior is not testable unless those other
variables are also measured.
Like Dr. Roizen, the authors of the other chapters in Section I also discuss
the limitations of the studies described in their chapters. While these limita-
tions should lead to caution in drawing causal inferences, there is much
valuable information provided by the epidemiological studies reviewed in the
chapters in this section. They show that alcohol is prevalent in violent acts
I • Overview 5
and they further our understanding of alcohol's role in aggression and vio-
lence by focusing on specific issues.
Kai Pernanen, who has made a life-long study of alcohol and violence,
has called for data on the role of alcohol in the escalation of aggressive behav-
ior, e.g., from verbal threats to physical violence.4 In Chapter 2, Drs. Martin
and Bachman test several hypotheses posited by Pernanen4 by examining the
role of alcohol in the escalation of hostile interactions. They review the litera-
ture and use the National Crime Victimization Survey (NCVS) to provide new
information on the escalation from threat to assault without injury to assault
causing injury. The NCVS is a large, nationally representative sample of the
US population 12 years and older and obtains information on crimes includ-
ing incidents not reported to the police.
In Chapter 3, Dr. Kaufman Kantor evaluates ethnic differences in the role
of alcohol in spouse abuse. She examines the interplay among ethnicity,
structural factors (e.g., poverty), and cultural factors (e.g., “machismo drink-
ing” in Hispanic culture).
Dr. Raskin White, in Chapter 4, discusses the relationship of alcohol-
related violence and ethnicity among adolescents. There have been few stud-
ies of alcohol and violence in youth per se. Dr. Raskin White reviews several
models for explaining the role of alcohol in violent acts committed by adoles-
cents and arrives at some surprising conclusions.
In Chapter 5, the final chapter of this section, Dr. Cherpitel examines the
role of alcohol and drinking patterns in violent and nonviolent admissions to
emergency departments. She reviews studies that used probability samples of
all injured patients representative of those served by these facilities.
Studies such as those presented in the section contribute to our under-
standing of the role of alcohol in aggression and violence. This epidemiologic
knowledge, when integrated with the knowledge derived from other disci-
plines, can provide the basis for designing prevention strategies to reduce the
violence that is related to alcohol use.
References
1. Collins JJ, Messerschmidt MA: Epidemiology of alcohol-related violence. Alcohol Health Res
World 17:93-100, 1993.
2. Gordis E: Alcohol, aggression, and injury. Alcohol Health Res World 17:91, 1993.
3. Roizen J: Issues in the epidemiology of alcohol and violence, in Martin SE (ed): Alcohol and
Interpersonal Violence: Fostering Multidisciplinary Perspectives (DHHS NIH Publication No.
93-3496). Washington, DC, US Government Printing Office, 1993, pp 3–36.
4. Pernanen K: Alcohol in Human Violence. New York, Guilford Press, 1991.
1
Epidemiological Issues in Alcohol-

Related Violence
Judith Roizen
Abstract. Epidemiological research on alcohol and violence exhibits a number of methodologi-

cal limitations. This is the case whether it is event based (i.e., based on samples of victims and/or
perpetrators of violence) or based on samples of the general population. The chapter identifies
some of the limitations that confront researchers, policymakers, and other end-users of the
research. The methodological issues are illustrated by exploring one type of violent event—
rape—and one genera1 population study—Kai Pernanen's research on alcohol-related violence in
a Canadian community. It is argued that epidemiological research would benefit from further
qualitative research on the natural history of violent events.
1. Defining the Problem

1.1. Introduction: The Six Dilemmas
In this chapter and the longer work from which it is drawn,1 I review a
number of studies on alcohol and violence that come under the umbrella of
epidemiological research. I take a very broad view of what is meant by “epide-
miological,” often looking at small populations and at analyses rarely carried
out by an epidemiologist or with the rigor of epidemiological research at its
best. The work reviewed here is among the best empirical research on alcohol
and violence from North America. These are studies of different populations
that contribute to our knowledge of the distribution and correlates of alcohol-
related violence. The chapter focuses on alcohol use in violent events rather
Judith Roizen • Institute of Population Studies, University of Exeter, Exeter EX4 6DT, England.
7
than the chronic alcohol problems of those who are violent or the relationship
between alcohol use and abuse and criminal careers. Table I shows the range
in percentages of alcohol-present cases in studies based on violent events
and, for comparative purposes, other untoward and serious events. The
width of the ranges in the proportion of alcohol-present cases in different
studies is the result of a number of factors. These include variable definitions
of alcohol use and the violent behavior itself, inconsistent attention to alcohol
in the event, and small sample sizes. The fact that there are few definitive
studies in this area and that studies are of uneven quality means that a close
look at each study reviewed is needed rather than the more usual concise
review of many studies.
Readers seeking to draw conclusions about alcohol and violence from
epidemiological research will find themselves caught in a number of dilem-
mas. First, despite decades of research on these problems and although alco-
hol use often precedes violence, we still know little about alcohol's role in
violent behavior. Much of the evidence on which judgment will depend
comes from data collected for entirely other purposes, such as data collected
in police reports or emergency room intake forms. Yet, purposive research is
expensive and there is very little theoretically guided empirical work to build
on. Even after many decades of research on alcohol and violence, Pernanen2
has asserted,
For the time being, we still need a much firmer empirical foothold, in order
to assess the validity of the relationship between alcohol use and violence
in potentially less biased samples of violence episodes and of actors in
these episodes than those available in official documents. We need infor-
mation on the potential role of alcohol in the choice of different types of
violent acts and in escalations in seriousness of aggression and physical
violence, as well as in the use of indiscriminate aggression in partial or
total obliviousness to the nature of the victim, the setting, and the general
social context.
In relation to a social problem as important as alcohol and child abuse,
Leonard and Jacob have concluded,3
A final difficulty worth noting is simply the paucity of literature attempt-
ing to examine this issue. Few studies have been conducted and most of
these have methodological problems. . . . Additionally, these few child
abuse studies are frequently concerned with only one or two specific forms
of child abuse, thus rendering comparisons between studies or conclu-
sions regarding one specific form of abuse difficult to make.
We know that an alcohol presence in violent events does not necessarily
mean that alcohol affected the behavior of any of the participants. And more
than half of violent crimes and other incidents of violence do not involve
alcohol use by the victim or the offender. Further, as is the case in much
epidemiological research, the precise mechanism for a relationship between
the independent and dependent variables is not known, and there is no gen-
1 • Issues in Alcohol-Related Violence 9
Table I. Summary of Studies Reporting Alcohol Presencea at the Time

of the Event (in percent)
Number of studies Range
Casualty
Accidents (nontraffic)
Fatal
Aviation 15 0.7–44
Drowning 14 12–80
Fire/burns 19 9–83
Falls 8 17–70
Work 1 15
Other accidents 7 9–45
Coroners’ studies 13 14–64
Nonfatal
Fire/burns 7 12–62
Falls 3 13–25
Work 2 1–16
Other accidents 5 21–83
Emergency room/trauma studies 3 23–63
Traffic accidents
Fatal
Drivers 33 32–64
Passengers 8 16–49
Pedestrians 26 21–83
Motorcycle 8 25–63
Drivers
Single vehicle 19 41–72
Multivehicle 15 18–51
Responsible
All fatal accidents 6 45–75
Multivehicle accidents 3 31–44
Nonresponsible 3 7–12
Nonfatal
Drivers 6 3–25
Crime
Arrested populations
Homicide offenders 13 28–86
Assault offenders 3 24–37
Robbery offenders 3 7–72
Sex offenders 18 13–60
Homicide victims 29 14–87
Assault victims 5 25–60
Robbery victims 2 12–16
Sex victims 5 6–40
Prison populations
Offenders 17 14–100
Suicide
Attempters 6 30–70
Completers 13 18–66
continued
Table I. Continued
Number of studies Range
Casualty
Accidents (nontraffic)
Fatal
Family abuse
Marital violence (men’s drinking) 6 6–57
Marital violence (women’s drinking) 2 10–27
Child abusers/neglecters 1 13
Child molesters 6 32–54
a
Studies use measures such as BACs, police reports of driving, witness reports, self reports.
eral agreement about which alcohol effects might be operating. More is writ-
ten about the possible contributions alcohol might make to violent and crimi-
nal behavior than is written from research that attempts to establish whether
there is an empirical relationship and what that relationship might be. Alco-
hol’s presence is often considered presumptive of a causal relationship.
The second dilemma that we face is the lack of cumulation in work on
alcohol and violence generally and in important specific areas such as alcohol
and rape or family violence. Research is scattered among disciplines, journals,
and countries. If one could characterize an area of research as very “pre-
paradigmatic,”4 this would be it. The task in reviewing this work is to try to
glean findings from work that springs from little or no common base. The
process of gleaning results from disparate studies of uneven quality means
that there cannot be the usual overview. We can learn something from these
studies only by taking a pointillist view, observing small parts in relation to
the whole.
The third dilemma that we confront in relation to research in this area is
that social research in the last two decades or so has become increasingly
complex. Looking for multiple causes of attitudes and behavior and using
multivariate methods for examining these potential causes have become part
of the stock-in-trade of the social scientist. Behavior was ever this complex,
but it is now recognized that we are no longer looking for a single or direct
cause of complex behavior. Good research of the last 15 years acknowledges
this in design and analysis, but the consequences are rarely explored. First,
the messiness involved in interpreting multivariate findings means that there
will be no simple or single consequence for policymakers. Correlatively, this
raises the question of how research on social problems should be divided
among administrative agencies and research groups.
For example, over the last two decades, as those looking at alcohol prob-
lems were slowly coming to grips with the multivariate causes of untoward
behavior, drugs became more frequently implicated in many of the behaviors
that we were seeking to understand. In Collins’ 1981 book5 on alcohol and
crime, drugs other than alcohol played a small part in our analyses. Drug use
is now present in violent behavior, especially criminal behavior, to a degree

that makes it questionable whether it is sensible to look at alcohol and vio-
lence apart from other drugs. The work of the Drug Use Forecasting group
shows that 59% of arrestees for violent crimes had been using drugs, often in
conjunction with alcohol, in the days prior to the offense. A good case can be
made that it is not just criminal violence that shows this drug presence but
much other violent behavior as well. R. Room (personal communication,
1993), however, has argued against including drugs routinely in research on
alcohol and crime on the grounds that the “alcohol will get lost” due to the
often greater attention to drug problems where both are under investigation.
Perhaps this need for the separation of in-depth investigation of alcohol and
drug problems in part reflects the fact that administrative control over re-
search and policy on alcohol and drugs is divided among different agencies
with differing agendas. But it is also symptomatic of the increasing difficulty
we have in handling multivariate explanations of social problems.
The fourth dilemma, related to the third, arises because we live in a
multivariate world in which our improved methods of social analysis have
capabilities beyond what the data will usually support. In part this is because
it is generally easier to develop new analytic methodologies than to find new
ways of measuring behavior. It is, in part, linked to the allocation of prestige
in disciplines. As Stinchcombe has argued,6
the higher the prestige of a piece of sociological work, the less people [who
are analyzed in it] are sweaty, laughing, ugly or pretty, dull at parties, or
have warts on their noses. . . . If we range theories from the prolix fashion
of Herbert Blumer—who knows how people will define the situation and
consequently what they will do—to the lean and spare rational actors
models that allow us to use maximization mathematical methods to speci-
fy at least one feature of the behavior exactly (e.g., what the net profit will
be), it is the theories that are most divorced from blood, sweat and tears
that have highest prestige.
It is a conclusion of many who write on alcohol and violence (e.g., Pernanen,2

Collins,7 Roizen8) that we need to know a great deal more about what actually
happens in violent situations: Who does what to whom and for what reasons.
This means systematic, in part, qualitative, studies to find out how people
actually act in situations that result in violence. There is often little prestige in
this and nothing exotic in looking at the natural history of events that affect
the people next door.
The fifth dilemma is that the police, courts, and medical professionals
need to make judgments about alcohol’s role in violence at a time when we
actually know relatively little about it. Murphy et al.,9 examining the relation-
ship between substance abuse and child abuse on behalf of the agencies
concerned with child protection in Boston, frustratedly argued,
Orme and Rimmer’s 1981 review[10] of the research on alcoholism and
child abuse concluded that the studies done up until that time had failed to
provide the empirical data necessary to support the association between
alcoholism and child abuse. . . . Although from a scientific point of view it

is important to maintain this methodological skepticism, it is equally im-
portant to note that from a practical point of view, courts, protective work-
ers, and clinicians are called upon to make decisions about the welfare of
children even when definitive evidence about the impact of factors like
substance abuse is not available. It is important to keep in mind that the
majority of the previous studies as well as prevailing legal and clinical
opinion agree that untreated, serious substance abuse plays a clear role in
increased levels of risk for child mistreatment.
They go on to assert, despite the limited empirical evidence, that
Substance abuse has been so clearly and consistently associated with child
mistreatment that the Boston Juvenile Court, like other family courts, now
accepts serious, untreated substance abuse as prima facie evidence of parental
inability to adequately care for a child. [emphasis added]
However, these families often have many other problems apart from their
history of substance abuse.
The last dilemma is that although some may argue that the contribution
of alcohol to violent behavior is “less than meets the eye,”7 the problem of
explaining the very great proportion of violent acts of all kinds that have
alcohol and intoxicated actors present remains (see Parker11). Presently, we
can explain neither to what degree alcohol is effectively involved in these
events nor why an alcohol presence is so prevalent.
The epidemiological research on alcohol and violence is large, diverse,
and poorly integrated. This chapter uses two approaches in assessing the role
of alcohol in violent behavior from an epidemiological perspective. Studies of
a single category of violent behavior—rape—are discussed using different
research windows based on different study populations. The same exercise
can be carried out in relation to other violent behaviors.1 Another approach is
to review in detail a single epidemiological study of alcohol and violence, in
this case, Pernanen’s Alcohol in Human Violence,2 in order to illustrate many of
the key issues in epidemiological research that will need to be addressed in
the next decade.
Any review of research on alcohol and violence must make a choice
between a broad overview of many studies and a detailed look at a few. The
importance of detailed analysis can be illustrated by an example of a review
paper that discusses studies reviewed in this chapter. Antonia Abbey,12 in a
review article, ”Acquaintance Rape and Alcohol Consumption on College
Campuses: How are they Linked,” uses two studies to establish that there is a
link between these two behaviors. (These studies, Koss and Dinero13 and
Muehlenhard and Linton,14 are reviewed in the second section of this chap-
ter.) Abbey’s review devotes only 14 lines to the actual evidence for the asso-
ciation. Three lines are devoted to Koss and Dinero.13 They read, “Alcohol
use at the time of the attack was one of the four strongest predictors of the
likelihood of a college woman’s being raped.” But this 1989 article used typical
alcohol use of women as the alcohol measure, not alcohol use at the time of
event, and it proved to be a fairly weak predictor. Alcohol use by men is

found to be a risk factor in Koss and Dinero.13 They measured alcohol use at
the time of the event on the part of men, but this factor is buried in myriad
other risk factors and is undefined. The remaining 11 lines are devoted to
Muehlenhard and Linton’s14 study of 635 psychology students on a college
campus. On this thread of evidence, alcohol begins to be perceived as a cause
of acquaintance rape.
1.2. Definitions of Violence

Violent behavior, as well as drinking behavior, covers an enormous num-
ber of different acts. Looking at only a single type of violent act, such as
assault, a number of physically and socially different acts are implicated: the
threat of assault, assault with a deadly weapon, assault accompanied by
physical injury. The same objective act may be characterized as directed
against a spouse, a child, or in war. Violent acts can also be typologized by
how they are subjectively perceived. Perhaps the single most important typol-
ogy of violent acts is achieved by dividing those that are legal from those that
are not. These may be the same objective acts with the same physical and
emotional consequences for the victim but may never come to the attention of
the police or welfare agencies.
Pernanen15 acknowledged the difficulty in aggregating all violent acts in
his 1976 review of alcohol and aggression. By separating instrumental crimes,
such as crimes for gain, from others, he sought some explanatory simplicity:
I will almost exclusively deal with noninstrumental and interindividual
crimes of violence. The emphasis will be on homicide partly because it is
an easily definable category of crime and thus there is the least possible
definitional variation between cultures and jurisdictions. Homicides are
definitely interindividual. A proportion of homicides are, however, instru-
mental for various reasons and one criterion is not optimally fulfilled.
“Assaults,” he argued, “are probably the most noninstrumental category of
violent crimes.” He noted, however, “If robbery, rape and arson were in-
cluded [in an analysis] just because they are classified as violent crimes for
nonscientific purposes, the explanatory accounting would have been ex-
tremely complex and more often misleading than not”
In the last decade, proportionately more homicides are instrumental,
especially those with some drug involvement, and therefore even they in-
volve an extension of the explanatory framework. In his recent empirical
work, Pernanen2 defined violence operationally by specific acts of physical
violence, measured at three behavioral levels: actual physical harm, threats of
violence, and witnessing violence. To be counted as an act of violence, “the
assailant must clearly have shown the intention to hurt, or shown that he/she
gave higher priority to reaching some other instrumental goal than to avoid
hurting the respondent.”
The focus of most research on alcohol and violence, especially criminal

violence, has been on noninstrumental, expressive acts of violence because it
is the (often unstated) belief of investigators that these are more likely to be
related to alcohol use. This is changing with the development of a body of
work on nonviolent criminal offending that contributes to our understanding
of alcohol and violence by illustrating the many nonviolent behaviors that
show a considerable alcohol presence (see, for example, Cordilia,16 Petersilia
et a1.,17 Ladouceur and Temple18). Other dimensions of violence that should
be but rarely are used in assessing the relationship of alcohol and violence
include the intensity of violent acts, duration in time, the rate of violent
episodes in a time period, and the physical consequences of a single vio-
lent act.
1.3. Measurement of Alcohol Use and Alcohol Problems

Just as there are a number of types of violent acts and ways of measuring
them, there are a large number of ways of measuring alcohol use. These
include blood or urine alcohol levels, self-reports of quantity and frequency of
drinking, drinking problems, types of beverages, congener contents of these
beverages, observer reports of drinking, speed of drinking, and alcoholism.
There is, in addition, variation in the cultural climate, temporally and geo-
graphically, in which drinking occurs and the alcohol-specific norms that will
affect drinking behavior.
There is a wide range of effects attributed to alcohol. These include effects
on coordination, eye movements, cognition, and judgment. There are also
“expectancy” effects: behavior may change when someone thinks they have
been drinking or when they think others have. Within the literature there is
considerable debate over the importance of pharmacological and cultural ef-
fects, a debate that sometimes borders on the ideological.
In analyzing alcohol and violent events, we are typically concerned with
distinguishing the acute effects of alcohol from the chronic or long-term ef-
fects. Thus, we separate out the use of alcohol in the event from the alcohol
problems of those involved in the event. In addition, we consider separately
those who are defined by their alcohol use and problems, that is, alcoholics.
Much of the research on alcohol, crime, and other violence in the last 15
years is far better than that which was reviewed in the wide-ranging review of
alcohol, casualties, and crime carried out by Aarens et al.,19 in 1977. The
epidemiological research on drinking patterns and problems is working its
way into the literature. Nonetheless, there remain many methodological
problems connected with the measurement of drinking. A blood alcohol mea-
surement must be taken on a person within a few hours after drinking has
occurred. Self-reports of alcohol use may involve some element of deviance
disavowal. Police may ignore women's drinking because they do not expect
them to be drinking heavily. Not all members of a sample will have an alcohol
measure taken, leading to possible biases in the alcohol-present subsample.
The time order of these behaviors is not always clear: violent behavior may
cause drinking, both by the victim and the offender. (These methodological
problems and other aspects of the measurement of drinking behavior and a
discussion of alcohol effects can be found in Aarens et al.,19 Greenberg,20 and
Roizen.1)
The complexity of the relationship between alcohol and violence, even
from an epidemiological perspective, is captured by Pernanen.21 In this exer-
cise, he proposes that we consider all possible measurements of alcohol as a
set and then consider all violent acts as a set:
Formally, all possible relationships between the elements of the sets would
be represented by the Cartesian product of those sets: {alcohol use ×
{violent acts}. In addition, [there will be] some interactive combination of
elements in the alcohol use variables . . . contemplating this way of repre-
sentation may make us more sensitive to the indeterminateness of much of
the discussion in this area.
(I have substituted “violent acts” for “crime” in this quotation.) We are, then,
engaged in the examination and evaluation of the research on some hundreds
of possible empirical relationships.
2. An Overview of Methodological Problems in Research

on Alcohol and Violence
Methodological and conceptual problems that arise in the definition and
measurement of violence and alcohol have been briefly discussed in the previ-
ous sections. This section outlines some of the other important methodologi-
cal problems and constraints. (There are a number of comprehensive meth-
odological critiques of research on crime and alcohol, including Pernanen,15
Greenberg,20 and Roizen and Schneberk.22) This chapter focuses on event-
based studies and studies of the general population, each of which has differ-
ent methodological problems.
2.1. Event-Based Research

By event-based research we mean samples of people to whom a serious
event has occurred (e.g., victims of rape or assault) or samples of people who
have initiated such an event (e.g., rapists or assaulters). For our purposes, we
are looking at the amount of alcohol consumed before these events or the
frequencies and kinds of alcohol problems these people have.
Perhaps the single most important methodological failing in event-based
studies is the lack of, or an inappropriate, comparison group. Thus, in eval-
uating the alcohol problems of a sample of battered women, it is essential to
know the level of alcohol problems in a sample of women comparable on
other variables. Since it is often the case that event-based samples do not have
comparison groups, distributions of alcohol problems in a general population

sample are sometimes used. However, the cases in the events sample may
differ on many other characteristics, making a general population sample
inappropriate. Where comparison groups do exist, they are often convenient
to the researcher rather than appropriate. Emergency room (ER) studies of
trauma, for example, will use other types of ER patients. When the purpose of
the research is to measure drinking problems, it may be questionable to
include in a comparison group women in labor, victims of heart attack, and
those suffering from surgical problems, all of whom are relatively unlikely to
have been drinking.
Elsewhere, Aarens et al.19 have argued that attempts to find comparison
groups for events that involve intentional behavior, as most violent acts do
fully or in part, are difficult if not impossible. A comparison group must be
based on ceteris paribus criteria. It is questionable whether these criteria can be
established for someone who has murdered his wife or shot someone in a
robbery. Is the person who lives next door a reasonable “control” for someone
who habitually assaults children? Assessing and controlling for the degree of
intentionality in violent behavior is a problem that needs to be addressed in
any study of violent behavior.
A second problem with event-based samples is that they are a highly
selective subgroup of all cases of the occurrence of the event, with perhaps
the single exception of homicide victims, most of whom are eventually discov-
ered. Women who are victims of domestic violence may only come to public
notice because they have nowhere else to go. This is more likely to be the case
for poor women than those who are wealthy. Severely battered women may
come to an ER, while others only slightly less injured nurse themselves at
home. Prison offenders will have been through the highly selective processes
of the courts, including plea bargaining and diversion.
Event samples typically include the ”worst cases.” Only a small propor-
tion of rape victims, for example, ever report their rape. These reported cases
are the ones that gain public attention in some way. Often these worst cases
have multiple social, economic, and personal problems and many live on the
fringes of society. For this reason much of the possible variation in important
explanatory variables is attenuated. Disproportionate numbers in these sam-
ples are poor, ill, use drugs, and are poorly educated. (See, as a dramatic
example of these multiple problems, the review of Barnard et al.23 in a later
section.)
Last, much of the data collected on events comes from intake and evalua-
tion forms that are meant for other purposes, such as police reports, ER
intake, and initial interviews with women seeking shelter. They are not pur-
posefully drawn questionnaires. Correlatively, often the data analysis is in the
hands of someone who is “interested in the problem” but is not skillful in the
analysis of the often complicated data.
The methodology of the study of events is underdeveloped and a signifi-
cant contribution to the study of alcohol and violence (or indeed other serious
events) would be made by further work in this area. Pernanen’s recent work is
a good beginning.
2.2. Studies of the General Population

We are here concerned only with those methodological constraints on
general population surveys that are relevant to studying substance abuse and
violence or other untoward events. The single most important constraint is
that in most social surveys, even large ones, there will be too few cases of
serious events such as violent behaviors or victimizations to justify the costs
of including the relevant questions. This problem becomes even more acute
when it is a relationship that is under investigation, such as the relationship
between alcohol and violence. Related to this is the fact that neither drinking
patterns and problems nor violent behavior are randomly distributed in the
population. Looking at the joint relationship may involve a biased subset of
relevant cases.
General population samples, even very good ones, miss large numbers of
people; indeed, this is true even of censuses. These missing individuals are
likely to be (or so we may think) those who have many of the problems in
which we are interested. Thus, thinking in terms of Venn diagrams, we may
have a large overlap between event samples and the general population;
alternatively, we may have little or no overlap. That is, it is possible that a
general population survey may miss altogether those most given to serious
violence, although the work of Straus and Gelles24 suggests that this is not
always the case. If extreme cases of the dependent variable, such as criminal
behavior, are undersampled in the general population survey, suspected risk
factors may appear to be relatively weak when in fact they are of considerable
importance (see, for example, Greenfield and Weisner25). One of the impor-
tant and unaddressed questions in the research on the epidemiology of vio-
lence is the degree to which there is a continuum of violent behaviors or
whether there is a sharp disjunction, with extreme acts of violence being
qualitatively different from other violence.
In this chapter I am looking in part at the epidemiology of “events”
described in general population surveys—events that may occur to a rela-
tively few people—in contrast to attitudes toward violence, which might
characterize the whole of a sample. Thus, a fourth problem, which is in part
described by Pernanen,2 can be stated as follows: Although the sample of
”events” from a general population survey is less selective than in event-
based samples, even these are not random samples of events. There is selec-
tive recall and, as argued above, the events that find their way into a general
population sample may well be a biased sample of all events. The fact that in
many cases the (retrospective) period from which these events are drawn
extends back in time many years creates a problem of its own. The types of
violent events in recent years may be of a different nature to those that
occurred 20 years ago. Patterns of violence and its modes of expression
change. Thus, the distribution of types of recent events may differ from those
that occurred to people some time ago but are still the most recent event they
experienced. Furthermore, without other types of events (e.g., weddings,
birthday parties) than violent events for comparison, it is impossible to say
with any certainty what the effects an independent variable such as alcohol
may have. A factor that must be accounted for in both event-based and gener-
al population surveys on problems of the type under investigation here is the
reluctance of some people to admit to acts that are deviant and that conse-
quently they may seek to disavow or reinterpret.
3. Evidence on Alcohol and Rape

3.1. Event-Based Research on Alcohol and Rape
Research drawn from data on arrested populations largely explores the
immediate situational characteristics of criminal events rather than long-term
personal, social, or economic problems of offenders or victims. The principal
foci of this research are violent “index” crimes according to the Uniform
Crime Report; that is, crimes against persons, such as robbery, rape, assault,
and homicide. The most well-considered event-based research on these spe-
cific crimes follows the basic design of the initial work of Wolfgang26 on
homicide. This design has been used in several subsequent studies of homi-
cide, and at least one study modeled after Wolfgang is found among those of
rape, robbery, and assault. In these studies the focus is on the characteristics
of the case as a whole rather than the characteristics of victims or offenders.
The data sets include a wide range of variables: ethnicity of victims and
offenders, alcohol use of victims and offenders, previous criminal record of
offenders, temporal patterns, spatial patterns, degree of violence, method,
motive, and various observations concerning victim-offender relationships.
Alcohol use is included as a single variable in these studies but is often only
covaried with some of these other variables. These studies have influenced
more recent victims studies, which continue to be an important source of data
on alcohol and violence.
The quality of these studies depends, in large part, on the quality of the
police records. Some of the studies reviewed here have been reviewed else-
where by Roizen and Schneberk22 and Roizen27: Only the better studies are
discussed here, with an emphasis on the United States. The ranges of alcohol
estimates in these studies are shown in Table I. Looking only at the better
studies has the effect of narrowing the range of estimated alcohol presence in
criminal events. It also allows us to dispose of studies that fail to meet even
minimum scientific standards.
Forcible rape is defined in the Uniform Crime Report28 as “the carnal
knowledge of a female forcibly and against her will” and has been redefined
to include males in some states. Assaults or attempts to commit rape by force
or threat of force are also included; however, statutory rape (without force)
and other sex offenses are excluded. In 1975, the rate of rapes was 51 per
100,000 women in the United States; in 1988, it was 73 per 100,000. This varied
from 83 in large cities to 36 in rural areas. In 1988, 52% of the known rapes
were cleared. Forty-three percent of rape arrestees were under the age of 25;
53% were white and 46% black. Rape is perhaps the most underreported
index crime, although report rates have grown as support for victims has
increased and attention has been brought to the problem (see US Dept. of
Justice, Law Enforcement Assistance Administration,29 and Bureau of Justice
Statistics,30 for government estimates of victimization and underreporting).
Arrest leads to a conviction in only a small proportion of cases. Dietz31 esti-
mated that only 16% of reported rapes led to a conviction, and nearly a
quarter of these were for lesser offenses (see also Clark and Lewis32). Thus,
offenders found in captured populations probably differ from the universe of
rapists.
There is a wide range of alcohol involvement reported in studies of rape
as shown in Table II. The Selling33 study is noteworthy because it gives a self-
reported alcohol measure, which is unusual in samples of arrestees (see also
Visher34 for self-reported alcohol use by arrestees for all violent crimes). The
level of reported alcohol use by offenders in these studies more closely ap-
proximates the estimates of self-reported alcohol use prior to the most recent
Table II. Empirical Studies: Rape Offenders and Victims
Percent Percent
Author, date, alcohol alcohol
location Sample offender victim Alcohol measure
Selling (1940) 100 cases male 43 — Combination

Detroit MI33 sex offenders self-reports
and police
reports
Shupe (1954) 42 apprehended 50 — Urine alcohol
Columbus OH58 rapists content
President’s Com- 151 cases of 13 6 Police reports
mission on rape alcohol presence
Crime (1966) 200 offenders
Washington59 151 victims
Amir (1971)35 646 cases of 24 31 Police reports
Philadelphia rape alcohol presence
1292 offenders
646 victims
Tardif (1966) 112 cases of 31 16 Police reports
Montreal60 rape alcohol presence
67 offenders
112 victims
Johnson (1978) 217 “founded” 37 36 Police reports
Winnipeg36 cases of rape alcohol presence
offense from sex offenders in prison than the estimates of use based on police
reports.
Estimates of alcohol use prior to criminal events vary considerably among
studies apparently similar in design for several reasons. These include differ-
ences between studies in the number of cases (small numbers leading to
chance variation), quality of data, or ecological differences. Both the Washing-
ton, DC, and Philadelphia studies (Table II) use a study design modeled on
the 1958 Wolfgang26 research on homicide. A closer look at these studies can
illustrate the difficulty the analyst has in trying to reconcile disparate find-
ings. The difference in estimates of alcohol involvement is considerable, al-
though both use police reports (see Table II). Both studies were carried out in
large metropolitan areas with populations comparable on most major demo-
graphic characteristics except ethnicity. In the years in which these studies
were carried out, 61% of the population of Washington, DC, was nonwhite
(largely black), while blacks made up only 18% of the population of Phila-
delphia. There are known differences in alcohol use by ethnic group. Amir35
reported that 42% of white rape arrestees had been drinking prior to the
alleged crime, contrasted with 24% of black rape arrestees, an ethnic differ-
ence supported by other research. This ethnic difference in reported drinking
prior to the crime could in part explain the difference in measured alcohol
presence between these two studies. However, the data from Washington
and Philadelphia show similar ethnic distributions of arrestees, although
there are different ethnic distributions in the population. Thus, this substan-
tial difference in ethnic distributions in the two communities in this case does
not explain the difference in alcohol presence. However, differences in demo-
graphic characteristics of samples are potentially important to explanations of
differences between studies in reported alcohol involvement; these are rarely
fully analyzed in relation to the alcohol variables.
Other possible explanations for the variation in alcohol presence in these
studies include differences in the level of attention paid to drinking that
occurs prior to criminal events in the different cities, in the availability of
alcohol in neighborhoods where crimes are likely to occur, or, as Johnson et
al.36 argued, they may be the result of a “real difference” in the use of alcohol
in different geographic areas. Whatever the explanation, these two studies
underscore the difficulty in obtaining consistent estimates of alcohol involve-
ment in criminal events even when research designs are similar and studies
are restricted to one type of criminal event.
The Amir35 study has gathered the most complete data on alcohol pres-
ence in rape events, although the study is not primarily focused on alcohol
use and some of the quantitative analysis is relatively poor. At the time
the Amir research was carried out, its value lay in the fact that it expanded the
focus of the investigation of criminal behavior beyond the offender to the
event and its situational and social context. This detailed analysis of 646 rape
events shows, for example, that over 40% of rapes involve multiple offenders;
in half of the rapes the victim and offender were acquainted and in 20% they
were neighbors. Half of the offenders had a criminal record, but few had
previous records of sexual offenses. The place of initial “meeting” of offender
and victim is frequently (41%) and somewhat surprisingly in one of their
homes. However, 42% occur “on the street.” Only 11% of rapes occur near
a bar.
These data also show a strong association between alcohol use and type
of interpersonal victim–offender relationship. Alcohol use was twice as likely
to be found in rapes involving strangers (in 44% of the rape events alcohol
had been used) compared with rapes involving primary relations (21% of
cases involved alcohol). It is particularly noteworthy that when only the vic-
tim had been drinking, the victim and offender were strangers in 77% of the
cases. Thus, drinking in rape, as in other crimes, may play any one of a
number of different roles: It may be present but have no effect; it may enhance
chances of victimization when the parties are strangers; it can be present in
the offender alone and exert an effect only on the offender, such as misread-
ing social cues in relation to prevailing norms; or it may begin an evening
gathering of a group of men that ends in drunkenness and rape.
Several other alcohol-specific findings are noteworthy from this study.
When rape involved a pair of men compared with a single man or a group of
men, the offenders were considerably more likely to have been drinking. A
number of studies of drinking and crime show excess force in alcohol-present
situations. Although the number of cases in which alcohol is present in the
offender only is small, all of them involved excess force against the victim.
Sexual humiliation was also more likely when alcohol was present. Alcohol
was present in 40% of the rapes committed on the weekend and 28% of those
committed during the week. Of those cases where alcohol was present in the
victim only, 40% occurred on a single day of the week, Saturday.
The Amir research shows that two thirds of the alcohol-present rapes
involved drinking by both victim and offender. For some investigators this
raises the question of whether or not the behavior of the victim may contrib-
ute to her victimization. “Victim precipitation,” or the victim’s own role in
influencing the course of the rape, is a socially sensitive issue. Progress has
been made in relation to the problem of blaming the victim by police, the
courts, and the general public in the two decades since Amir’s work. Amir’s
analysis is not sensitive to these issues. However, keeping this in mind,
Amir’s work contains some alcohol relationships that deserve further investi-
gation. Amir35 defined victim precipitation as
rape in a particular situation [in which] the behavior of the victim is inter-
preted by the offender either as a direct invitation for sexual relations or as
a sign that she will be available for sexual contact if he will persist in
demanding it. Excluded are the situations where no interaction was estab-
lished between the offender and the victim, and when the offense was a
sudden event which befell the victim.
Approximately one in five rapes was considered to be victim precipi-
tated. Victim-precipitated rape was more likely than other rape to involve a
white victim and/or a white victim and white offender pair. In the majority of
cases the offender and victim were at least acquaintances. Fifty-three percent
of victim-precipitated rapes involved alcohol compared with 29% of nonvic-
tim-precipitated rapes. In 35% of victim-precipitated rapes both the victim
and the offender had been drinking; in 18% only the victim had been drink-
ing. The proportion of victims-only drinking in victim-precipitated rape was
more than twice that in nonvictim-precipitated rape. However, the degree to
which a victim’s drinking may evoke a presumption, on the part of the police
or others, of blame for her involvement in the rape event has been the subject
of relatively little research (see, however, Richardson and Campbell37).
The finding that 60% of the victim-precipitated rapes involved sexual
humiliation, in contrast to 18% of other rapes, is a startling one. Amir95
argued that this is very likely due to misread signals on the part of the
offender:
[S]ubjecting the victim to forced sexual intercourse means that the imputa-
tion of sexual availability was a false interpretation on the offender’s part.
He may still hold to his views and try to prove them by subjecting her to
sexual humiliation, other than forced intercourse, or he may humiliate her
as a revenge just because of the failure of his imputation.
Drinking may contribute to the misreading of signals on the part of both the
victim and the offender.
Although “victim precipitated” is the wrong term for describing these
rapes, they are rapes in which the victim may have increased her vulnerability
by her own behavior. Drinking or some types of pub or bar behavior may be
factors that increase a woman’s vulnerability. Deming et al.,38 in their study of
fatal sexual assaults, report a positive blood alcohol content (BAC) for 40% of
the victims; of these, half were intoxicated. These investigators suggested that
the victims may have contributed to their deaths by their behavior and judg-
ment, including the inability to escape.
The research of Johnson et al.36 on alcohol and rape in Winnipeg shows a
much higher proportion of alcohol-present cases in their series, although the
study design is similar. In their series, 74% of victims or offenders were
drinking prior to the event. This difference may be geographic or more likely
the result of increased attention to reporting alcohol use since the Amir re-
search. Again, in the majority of alcohol-present cases, both the offender and
the victim had been drinking. This study shows a significant difference in the
use of physical force in alcohol-present compared with alcohol-absent rapes.
Rapes in which both the victim and offender had been drinking involved
substantial force in 37% of the cases; this is contrasted with 18% of the cases in
which no alcohol had been used. Looking at all alcohol-present cases, 85%
involved the use of some force contrasted with 68% of cases in which no
alcohol had been used. However, the highest level of force as measured in
their index of force was rarely (in 5% of the cases) but equally used in both
alcohol-present and alcohol-absent cases.
Few sexual assaults end in homicide. Those that do end in death fre-
quently show injuries and perversion. Deming et al.38 reported on 41 female
cases of proven fatal sexual assault over a 10-year period in Dade County,
Florida, nearly half of whom were physically traumatized and injured. Thirty
percent of the victims were black, in a county in which nonwhite residents
averaged 16% of the population over the period covered. Of the 37 victims
tested, 40% tested positive for alcohol use. More than half of those tested had
a BAC of 0.10 or higher. Only two of the victims were known to be prostitutes.
The role of alcohol in sexual assault with serious injury or resulting in homi-
cide is one that needs further investigation, especially in light of new evi-
dence that a substantial proportion (estimated to be between a quarter and a
third) of sexual offenders are reconvicted of a sexual or violent offense,39 and
the fact that these events are impulsive/explosive events that may involve a
drinking victim.
3.2. Studies of Prison Offenders

Research based on prison offenders offers a second window on the rela-
tionship between alcohol and violence. Estimates of alcohol involvement in
criminal events based on the self-reports of convicted offenders show a differ-
ent pattern of relationships between criminal behavior and alcohol use than
that based on samples of arrestees. While the prison data support the view
that a substantial proportion of violent offenders were drinking or drunk at
the time of the crime, these data show considerable alcohol presence in other
crimes as well. A detailed reanalysis of data from an early national survey of
prison offenders22 showed that although drinking at the time of the crime
varied by type of crime and was greater for violent interpersonal crime than
for property crime, these differences were not large. Among those who had
been drinking, “drunkenness” at the time of the crime was no less common
for property than for crimes against the person, despite the greater skill
assumed to be required for property crimes.
This pattern of relationships of drinking and type of crime from prison
studies is in marked contrast to the pattern found in arrested populations.
The arrest data show a strong relationship between seriousness of the crime
and alcohol presence in the offender and similarly significant differences in
alcohol presence in personal violent crime compared with property crime.
Research of similar design based on arrest record data shows 7% of rob-
beries,40 34% of rapes,41 24% of assaults,42 and 55% of homicides26 involved a
drinking offender. Comparable proportions based on the US Department of
Justice43 prison offender sample are 39%, 57%, 61%, and 53%, respectively.
The prison data also reveal that a large proportion of burglaries (47%) and car
thefts (46%) are committed after drinking. The national survey of prison
inmates carried out in 1979 largely supports the data from the earlier national
survey. However, the 1979 survey,43 based on personal interviews with 12,000
inmates, including women, gives a more detailed picture of the drinking
habits of prisoners than the earlier national survey. Violent offenders and
property offenders were about equally likely to have been drinking prior to
their current offense (50% and 46%, respectively). Of those who were drink-
ing, 60% of violent offenders and 68% of property offenders reported drink-
ing very heavily. As well, the proportions who reported being very heavy
drinkers in the year prior to the offense for which they were incarcerated were
also approximately equal. Thirty-five percent of violent offender and 40% of
property offenders reported being very heavy drinkers.
Ladouceur and Temple,18 using these data, compared the drinking be-
havior of rapists and other prison offenders. Their analysis shows that rapists
are no more likely to drink heavily before the offense for which they are
incarcerated than are those convicted of assault or burglary, and that they
are about as likely to report feeling drunk as those committing burglary.
Ladouceur and Temple18 noted, “This study finds no differences for heavy
alcohol use or for level of drunkenness between offenders who committed
violent and nonviolent, or sexual or nonsexual crimes.” Further, their results
show that both rapists and other offenders are likely to drink less heavily at
the time of the offense than on a typical drinking occasion in the past year.
While almost 90% of rapists drank moderately to heavily in the year prior to
incarceration, only 60% drank prior to the offense. The fact that there are no
significant differences in drinking behavior by offense group suggests that
criminal behavior may not be seriously influenced by drinking in the event, but
rather that criminal offenders generally are very heavy drinkers and if there is
any contribution made by alcohol, it is in this way. Ladouceur and Temple18
concluded,
Because drinking during the past year is not typically associated with the
commission of a crime, we conclude that drinking at the time of offense is
likely to reflect a typical drinking pattern, or in some other way is unre-
lated to the commission of the crime. If there was a causal link between
alcohol use and crime, such that heavy drinking increased the probability
of committing the crime, then we would expect offenders to drink more
heavily at the time of offense than on typical drinking occasions.
The work of Barnard and colleagues23 suggests that future research on
alcohol and rape, based on samples of prison offenders, should differentiate
offenders with a long history of drinking problems from others. These investi-
gators came to conclusions that are similar to those of Ladouceur and Temple
in relation to the failure of acute alcohol effects to explain rape or other
criminal behavior. Although it has a small number of cases, the Barnard et
al.23 study is important for its attention to the multiple social and psychologi-
cal problems most offenders have. These investigators reviewed the psychi-
atric evaluations prepared for the Florida courts of 88 offenders charged with
rape. Of the 88, 60 were classified as nonalcoholic, although others met some
of the investigators' criteria for alcoholism. Both groups of offenders had
experienced problems in their parental families either through divorce or
death. Nearly half of the offenders had a parent die or the parents divorce by
the time the offender reached age 18. Both groups had school problems and
low levels of educational attainment. The alcoholic group began drinking
considerably earlier than the nonalcoholic group: at about 14 for alcoholics
and over 16 for nonalcoholics. Of those called for military service, 69% of
alcoholics and 44% of nonalcoholics were either rejected at entrance or re-
ceived a dishonorable discharge. Work histories show frequent impulsive
changes or firings. While 82% of the alcoholics had been married at some
time, only 27% were married at the time of the offense. Comparable percent-
ages for nonalcoholics are 53% and 25%, respectively. The groups differ sig-
nificantly in criminal histories. While 36% and 45% of the alcoholics had been
convicted of assault or other violent charges, respectively, this was the case
for only 18% and 13% of nonalcoholics. About half of both groups had previ-
ously used drugs. The two groups differ significantly in their relationship to
their victims. Thirty-two percent of the alcoholics raped a relative and 41% an
acquaintance. This was the case for 11% and 28% of the nonalcoholics. In both
groups, substantial proportions of offenders had medical and psychiatric
problems.
In relation to the alleged offense, nearly 60% of the alcoholics reported
drinking heavily at the time of the incident compared with 30% of the non-
alcoholics. Seventeen offenders reported blackouts due to alcohol and could
not describe the context of the offense at all. Barnard et al.23 concluded that
For both the alcoholic and non-alcoholic prisoners, long-standing and mul-
tifaceted histories of disturbed behavior were recorded. It appears there-
fore that alcohol abuse is but one part of the picture, with sociopathy and
other forms of interpersonal disturbance contributing to the criminal
act. . . . The alcoholics stand out as more severely disturbed than the non-
alcoholics in the amount and pattern of deviant behavior. . . . [T]he data
suggest that such immediate effects of alcohol [as are seen] are not suffi-
cient to account for the observed cases of rape which arise out of long-
standing patterns of deviance.
Collins and Schlenger44 carried out a multivariate analysis of the relation-
ship of acute and chronic alcohol effects (i.e., the effects of long-term alcohol
use rather than the immediate effects, whether pharmacologically or cultur-
ally defined) in a sample of those recently admitted to North Carolina prisons.
They found that chronic effects were not significantly associated with either
incarceration for a violent offense or with committing a violent offense in the
year prior to incarceration. Age, race, marital status, education, and criminal
career variables were included in the logistic regression models. These inves-
tigators concluded that “it is the proximal effect of alcohol use, rather than
characteristics associated with being alcoholic, that is associated with in-
creased likelihood of violence.”
Can the conclusion from these different studies be reconciled? Does alco-
hol contribute to violent criminal behavior? Is the evidence in? The answer is
that it is not. What is clear is that broad categories of offense do not ade-
quately distinguish the actual behavior involved. Even specific event types
(e.g., “rape” as compared to “violent crime”) may mask significant variation

in alcohol use in different types of rape events. That is, sadistic rape or date
rape or incest (as compared to other types of rape or sexual offense) may well
be caused by different alcohol effects and characterized by different levels of
drinking, insofar as alcohol is a determinant of rape at all. Research on crimi-
nal behavior and alcohol and drug effects, therefore, must be more theoreti-
cally driven and these theoretical investigations must control for the other
social, economic, mental health, and other health problems of the offender.
The theory that alcohol use is only a marker for an intercorrelated set of other
problems must be considered in any investigation.
Groth and Birnbaum’s45 extensive empirical work on rape suggests direc-
tions for further theoretically based empirical research on drinking and rape.
Based on interviews with a sample of 500 sexual offenders, Groth and Birn-
baum outline three patterns:
1. Anger rape: “Sexuality becomes a means of expressing and discharg-
ing feelings of pent-up anger and rage. The assault is characterized by
physical brutality. ”
2. Power rape: “In these assaults, it is not the offender’s desire to harm
his victim but to possess her sexually. Sexuality becomes a means of
compensating for underlying feelings of inadequacy and serves to
express issues of master, strength, control . . .”
3. Sadistic rape: “Both sexuality and aggression become fused. . . .
There is a sexual transformation of anger and power so that aggres-
sion itself becomes eroticized.”
We would expect alcohol to play a different role in these types of rape.
For example, in anger rape, alcohol may enhance assaultive feelings. In power
rape, alcohol may be used for “Dutch courage” or, as some cases suggest, as a
way of trying to suppress sexual responses. Sadistic rape fits a pattern of
alcohol-related violence that involves sexual humiliation and excess violence.
Rada,46 for example, has suggested that in some offenders alcohol has a
direct, triggering effect on both violent sexual fantasies and behavior. The fact
that many rapists report that they cannot have intercourse in the rape situa-
tion may also be an alcohol effect, one that leads to angry and sadistic re-
sponses.
Unfortunately, Groth and Birnbaum45 paid little attention to alcohol in
their work, arguing that
The use of alcohol, in and of itself, is insufficient to account for the offense.
Although some offenders were to some extent intoxicated at the time they
committed their assaults, these same men were more often not sexually
assaultive when intoxicated. Our data suggest that alcohol may at most
serve as a releasor only when an individual has already reached a frame of
mind in which he is prone to rape.
However, they also argued
that alcohol may contribute to the releasing of rape impulses or assaultive
tendencies in some offenders . . . may impair such cognitive functions as

reasoning and judgment . . . may be a necessary component in a process
that evolves into an assault. . . . in other cases, alcohol abuse and sexual
abuse may constitute two parallel but independent symptoms of person-
ality dysfunction.
3.3. Rape in the General Population

A third window on the relationship between alcohol and rape comes
from general population victimization surveys. Official surveys such as the
national crime survey30 and parallel surveys in other countries estimate the
overall level of victimization and the degree of underreporting of crimes such
as rape. They give little or no attention, however, to risk factors such as
alcohol. Pernanen’s important recent work on alcohol and violence in a gener-
al population sample does not treat sexual offenses separately. The best
source of data on alcohol and rape based on a sample of the general popula-
tion is the work of Koss and her colleagues,47,48 although this work is limited
to college students, The most recent research is based on a national sample of
college and university undergraduates and includes 6159 men and women in
32 higher education institutions. Twenty-seven percent of women reported a
sexually coercive experience since the age of 14 that met the legal definition of
rape, including rape attempts. Fifteen percent of women reported having
been raped and 12% reported attempts. Eight percent of men reported per-
petrating an act that met the legal definition of rape. Five percent admitted
rape and 3% admitted attempts. The difference in these percentages between
men and women suggest either that women’s sexually coercive experiences
were with men outside the higher education system, for example, with a
family member, or that there are considerable differences in women’s and
men’s perceptions of how coercive these sexual events were. There is, of
course, no reason to believe that all men will admit in a questionnaire to
having committed a violent act such as rape, even if they believe in the
anonymity of their responses.
Eight percent of women reported having had unwanted sexual inter-
course because “a man had given you alcohol or drugs.” (Unwanted sex as a
result of the woman’s own drinking and perceived loss of control was not
included.) A man’s giving unwanted intoxicants was considerably less impor-
tant, however, than being “overwhelmed by a man’s continual arguments
and pressure,” which 25% of women reported.
Although the fact that women’s drinking patterns are found to be a risk
factor for rape and alcohol use had predictive power in the discriminant
analyses used, the relationship between alcohol and rape is not a particularly
strong one.13 Using measures of typical drug use (i.e., frequency of drinking,
frequency of drunkenness, and usual numbers of drinks per drinking occa-
sion), the raw means of the drinking index, which is unreported but has a
range of 3–15, were as follows: nonvictimized, 6.89; sexual contact, 7.38;
sexual coercion, 7.98; attempted rape, 7.82; and rape, 8.01. Four categories of
sexual coercion are used in this analysis. Sexual contact includes kissing and
fondling under pressure; sexual coercion includes sexual intercourse under
pressure but not by use of force.
As these data show, the differences in these scores on the alcohol use
index cover a narrow range of drinking behaviors given the scope of the
index, with its potential range of scores from 3–15. The investigators noted,
An inspection of the means on alcohol used indicated that women who
had been raped on average received a score that reflected a usual drinking
pattern of (a) 1-3 times a month; (b) usually no more than 4 cans of beer
(or equivalent in wine or spirits); and (c) getting drunk less than once a
month but at least once per year. The score for the group of women who
had not been victimized represented the next lower usage level in any one
of these three categories.
Since the great majority of college women drink and as many as 12% may
be considered heavy drinkers, the level of drinking represented by those
women who have been raped is by no means rare (Johnson et al.,49 Engs and
Hanson,50 Gleason51). Drinking patterns vary by area of the country and type
of higher education institution, as no doubt do sexual norms and behaviors.
These factors need further analysis before drinking can be seen as a risk factor
for the sexual victimization of college women.
Women’s and men’s alcohol use in the event is analyzed in Koss et al.52
Comparing stranger (n = 52) and acquaintance rape (n = 416), based on the
survey described above, shows substantial alcohol and drug presence in both
types of rape. Women had been drinking and/or taking drugs in 68% of the
stranger rapes and 55% of the acquaintance rapes. Comparable numbers for
the men involved were 76% and 67%, respectively. About 45% of both men
and women in both types of rape had used alcohol only; the remaining cases
had used alcohol and drugs or drugs only. The use of alcohol and drugs
varied by type of acquaintance rape. The proportions of women and men
(respectively) using alcohol and/or drugs in the different types of rape events
were 65 and 75% in “nonromantic” rapes, 78 and 84% for rapes occurring on
casual dates, 45 and 55% on steady dates, and 13 and 42% in rapes involving a
spouse or family member. (Men’s use of intoxicants is as perceived by the
women involved.)
The level of force used by the offender varied by type of rape. Greatest
force was used in stranger rapes and those involving family members. The
least force was used on casual dates. However, alcohol use was greatest on
casual dates for both women and men. Eighty-one percent of the men in-
volved in rape on a casual date had used alcohol, as had 70% of the women.
While the work of Koss and her colleagues suggests that alcohol use might be
a risk factor for rape, there is no simple positive association between force and
alcohol use. Family and spouse rape involved the least alcohol and drug use
on the part of the offender, while alcohol and drugs were used by three
quarters of stranger rapists. In both types of rape the use of offender force is
considerable. Thirty one percent of spouse/family rapes involved choking,

beating, or using a weapon (11%). Comparable proportions in stranger rapes
were 32% (16% of offenders used a weapon). Unfortunately, the Koss survey
does not report the amount of alcohol and drug use nor other characteristics
of the rape events, information that would help establish the role of alcohol
and drugs, if any, in these rape events. Furthermore, as with all violent acts,
there is a great range in the severity of the threat and the outcome. Although
the rapes and attempts found in the Koss sample meet the legal definition of
rape, they no doubt differ in many characteristics from the rapes found in
samples of arrested and convicted rape offenders. Only 23% of the women to
whom acquaintance rape happened described themselves as victims of rape;
44% of the victims reported having sex with the offender again.
Muehlenhard and Linton,14 in a much smaller study of college students
at a single university, show a significant relationship between alcohol and
drug use and sexual aggression. Comparisons of most recent dates with dates
in which unwanted sexual activity occurred showed that significantly more
dates in which sexual aggression occurred involved acting or feeling moder-
ately or extremely intoxicated (as a result of alcohol and/or drugs). This was
true for both women and men based on the responses of women and men
reported separately. The difference in reported intoxication between the two
types of dates is considerably greater from women’s reports than men’s.
Women reported heavy use of intoxicants by both themselves and the man
involved four times as frequently on dates involving sexual aggression con-
trasted with the most recent date. Men reported heavy use about twice as
frequently. However, this study’s definition of sexual aggression is very broad
(i.e., including anything from kissing and touching to forced oral sex and
sexual intercourse) and occurred to 78% of the women and was perpetrated
by 57% of the men.
These studies raise important questions about alcohol and drug use and
sexual activity. The degree to which men excuse their own sexual aggression
and women explain their sexual activity to themselves and others by using
drinking explanations needs further investigation. But considerably more re-
finement of the alcohol measures, description of the context of the event, and
controls for usual drinking and drug taking are needed.
In this review of alcohol and rape we have seen the complexity in assess-
ing the contribution of alcohol to this type of violent behavior. A number of
contextual factors are shown to be related to alcohol in the rape event. How-
ever, many rapists have multiple social and personal problems that may
themselves explain this deviant sexual behavior. Rape offenders, like other
violent offenders, are typically heavy drinkers and drug users. Alcohol use in
the event may represent no more than everyday use. Extending the study of
rape into the student population as Koss and her colleagues have done sug-
gests, however, a rather different set of correlates of rape than we find in the
prison offender population.
In the next section we turn to a single study that looks in detail at alco-
hol’s role in violence and is the most important contribution to the epidemio-
logical literature in this area of research in many years.
4. The Recent Work of Pernanen

The recently published work of Kai Pernanen,2 AIcohol in Human Violence,
deserves a special place in this chapter for a number of reasons. First, the
work is wholly devoted to the problem of alcohol and violence, whereas
much of the other work reviewed here has many competing agendas, often
raising more questions than giving answers to the question of the relationship
between alcohol and violence. Second, and of very considerable importance,
is the fact that Pernanen’s work is cumulative in relation to the study of
alcohol and violence. Unlike many of those who carry out research on alcohol
and violence, he is not making an occasional foray into the field. His work is
based on his own considerable work in this area of research and a close
reading of that of others, including the very large related experimental litera-
ture. This sort of cumulative research is rare in contemporary social science
where analysts often move from problem to problem as funding or interest
compels. Third, the work is of a very high standard. The survey is a classic
piece of survey research in an area of research that is extremely patchy with
respect to quality.
As Pernanen2 wrote, “The main strength of these data is that they repre-
sent ‘real’ naturally occurring events of aggression and violence,” which can
provide much needed descriptive analyses of aggressive episodes and their
incidence and prevalence and can serve as models for controlled studies of
aggression. As he argued, “Both middle range theories and middle range data
have been missing from the study of human aggression.” While underscoring
the importance of description in the study of violence and the paucity of good
data, despite the many studies of alcohol and violence, it is description in the
service of providing an explanatory framework for alcohol-related violence
that is the strength of this research.
Inevitably, in a tightly argued book-length manuscript, the reviewer must
select among the many findings a few that give the flavor of the work and that
epitomize its essential contribution. The summarized findings below include
some of the important descriptive findings from the survey as well as several
that will contribute to explanation and theory in this area of research.
The survey is based on a probability sample (Thunder Bay, Ontario) of
933 men and women aged 20 and over representing a city of 112,500. Of these
933 respondents, 492 had been victims of violence at some time since they
were 15 years of age. The most recent incident of violence is the subject of
most of the analyses. Violent incidents in the 12 months prior to the survey
are also analyzed, but these numbers are smaller. About 10% of the 495 men
in the survey had been victims of violence, 10% had been threatened with
violence in the previous year, and 39% had witnessed violence. Comparable
figures for women are 10%, 6%, and 28%, respectively.
This is a victimization study in the sense that violent incidents are de-
scribed from the perspective of the victim. The focus of the study, then, is the
role of alcohol in violent victimizations, not the role of alcohol in the aggres-
sive and violent behavior of the respondents. A comparison study of violent
crimes (n = 781) based on police records was carried out at roughly the same
time. Only 4% of the violent episodes from the interview survey were re-
corded by the police in the year of the study, although the police were made
aware of 15% of the episodes. This demonstrates the fact that the analysis of
cases from police records involves a small and selective subset of all cases of
violence, although these probably consist predominantly of the most serious
cases.
Although the risk of violent victimization in the 12 months preceding the
survey was about equal for men and women, 60% of male and 44% of female
respondents reported having been victimized since age 15. There is the prob-
lem of the adequacy of recall for the violent incidents that make up the main
analysis: 40% of the index incidents, that is, the 492 incidents, occurred dur-
ing the 3 to 4 years prior to the survey; however, another 40% occurred more
than 8 years prior to the survey. Of these incidents, men were dispropor-
tionately likely to have had their last victimization in their youth, while wom-
en reported more recent incidents. Some of the major findings of this work
are outlined below.
4.1. Pervasiveness of Alcohol

In more than half of the index incidents of violence in the community
sample and 42% of the violent crimes reported in the police sample, either the
victim, the assailant, or both were drinking. In the interview study, 51% of the
assailants (note: as perceived by the victims) and 30% of the victims had been
drinking; in the violent crime study the comparable percentages were 31%
and 26%, respectively. Pernanen2 concluded:
We now have some evidence that, at least in a cultural sphere where
alcohol is implicated in criminal violence, it is also abundantly present in
day-to-day violent confrontations. The relationship between alcohol use
and severe aggression, as reflected in studies of police and court records
and in emergency room samples of injured persons, does not seem to be
mainly an artifact created by biasing selection processes.
Nor, as he rightly concluded, can this relationship be seen as pertaining only
to a small group or particular subcultures in the population. The question of
the representativeness of event samples is often raised, and this work of
Pernanen’s gives us an answer based on a general population survey in one
community.
4.2. Differential Risk of Alcohol-Involved Violence

Many studies show that both heavy drinking and drinking problems are
related to gender and age. The data from Thunder Bay, perhaps not sur-
prisingly, also demonstrate that particular demographic groups in the popula-
tion have higher risks than others of alcohol-involved violence and that this is
in excess of what would be expected merely by the frequency of their drink-
ing. Young men are most at risk, although all young adults are at greater risk
than others.
The risk of injury from the index violent incidents (i.e., not those in the
year prior to the survey) was surprisingly high. Twenty-six percent of the
incidents resulted in a physical injury; 11% involved seeking medical atten-
tion. It is an important finding of this work that alcohol-present episodes did
not result in any greater rate of injury than those that did not involve alcohol.
However, the risk of injury increased with the amount of alcohol consumed
by the victim.
4.3. Selected Findings on Alcohol and Violence from Pernanen’s Work

The findings reported here are important in their own right in the devel-
opment of both empirical research and theory in this field; they are also some
of the findings that refer to themes from other studies reviewed in this chap-
ter and in the longer paper from which it is drawn. Included is a comparison
of drinking during violent episodes contrasted with usual drinking patterns
and an examination of differential alcohol involvement in violent episodes
involving acquaintances versus strangers, with different gender mixes of vic-
tim and assailant, and in different locations.
1. The amount of alcohol consumed by both men and women in their
index (i.e., most recent) victimization was considerably higher than the mean
levels of consumption during their most recent drinking episodes. This sug-
gests the need for further work on victim precipitation or vulnerability to
violence.
2. Alcohol involvement differed according to the gender of the victim and
assailant. Total alcohol involvement in episodes of a male victim and assailant
was 62%, a female victim and male assailant was 53%, and of a female as-
sailant was 27%. Violent episodes between men not only had higher levels of
alcohol involvement but were also more likely to lead to injury.
3. Alcohol involvement differed according to the relationship between
the victim and the offender. Total alcohol involvement was greatest in epi-
sodes between strangers. Seventy-eight percent of these incidents involved
either a drinking victim or assailant. In 36%, both were drinking. More needs
to be known about these “stranger” episodes that make up nearly a quarter of
violent episodes.
Over half of the violent incidents reported by women involved conflicts
with their spouses, Only 12% of the incidents reported by men were reported
as family violence. This difference is difficult to explain without more data. It

may be the case that men “forget” their incidents of family violence or that
men do not see them to be as serious as women do. (As we have seen in the
previous section, some men have perceptions of sexual coercion that are quite
different from those of women.) Nearly half of the violent episodes between
spouses involved drinking by the victim or the assailant. The victim (in most
cases the wife) was drinking in only a third of these episodes. Pernanen2
noted, “The serious nature of alcohol use in some marital violence is probably
reflected in the finding that divorced or separated respondents had an alcohol
involvement of 69 percent in their most recent subjection to violent acts”. He
notes that the sample is small (n = 37). One fifth of episodes of family violence
resulted in an injury.
4. Based on the episodes of violence in the year prior to the survey,
Pernanen2 found no
clear-cut relationships between the typical drinking frequency of the indi-
vidual and the three types of experiences of aggression during the preced-
ing year: men who were more frequent drinkers were not more likely to
experience acts of violence, threats and witness violence than were other
men.
The same relationship was not true for women. Among both men and wom-
en, those who drank once or twice a week were considerably more likely to
witness violence (and presumably to increase their chances of being partici-
pants) than more frequent drinkers. He observed,
The point that this discontinuous finding should make clear is that, even
though a statistical connection between alcohol use and aggressive en-
counters seems very likely in many jurisdictions and cultural spheres, we
should not expect a linear relationship between frequency of drinking and
these experiences.
This is an important point, one that has consequences both for choice of
analytic methods and choice of alcohol variables used in research on alcohol
and violence. There is growing evidence that heavy infrequent or binge drink-
ers may be disproportionately involved in violent behavior (see, for example,
Kantor and Straus53). This needs further exploration.
5. The findings on violence that occurred in a tavern are noteworthy. All
except one of the assailants had been drinking. The victim had been drinking
in about 80% of the cases. The proportion of injuries resulting from these
violent encounters was almost twice as great as from incidents that occurred
in the respondents’ own home. This may reflect the fact that tavern violence
reported in the interviews occurred in large part among strangers.
4.4. Contributions of Pernanen’s Recent Work to Theoretical Debate

Pernanen2 considers three “clusters of hypotheses” that are relevant to
determining the role of alcohol in violence: these include severity and per-
sistence hypotheses, indiscrimination hypotheses, and elicitation hypothe-

ses. The latter is not dealt with in Pernanen’s book but will be in later work; it
suggests that when alcohol is added to any situation, the risk of eliciting an
aggressive response is greater. The other two hypotheses are briefly reviewed
below.
4.5. Alcohol and Severity of Choice of Acts and Outcomes

Tests of seriousness of the choice of violent acts and their consequences in
relation to alcohol-involved violence are important in the development of a
coherent theory of alcohol-related aggression. Severity hypotheses are rele-
vant both to disinhibition-type theories and to establishing whether a dose-
response relationship exists in relation to alcohol and untoward outcomes.
The persistence hypothesis is related: that an intoxicated aggressor will per-
sist in violence beyond what would occur in “normal” violence. Wolfgang26
and his students have called this “excess violence .”
Pernanen2 concludes that “no support has been found for a general se-
verity hypothesis in these data.” This is largely based on the failure to find a
difference in injury outcome between drinking and nondrinking episodes and
the failure to find a difference in rate of injury related to the assailant's drink-
ing. There is limited conditional support for finding a difference in rate of
injury when the assailant was judged to be drunk, but the difference is not a
large one.
In my view, the evidence is not in on this question. These data are not
sufficiently finely drawn to support such a conclusion. The fact that there
were both (1) a clear relationship between very heavy drinking on the part of
the victim and the risk of injury and (2) an elevated risk when the assailant
was judged to be drunk suggest that there may indeed be a relationship
between level of drinking and severity. Furthermore, it is in the nature of the
sample that it may not capture many very heavy and frequent drinkers. Thus,
if there were a relationship between amount of alcohol consumed and the
severity of the outcome, this relationship would be attenuated. The two
weakest aspects of this research program as a whole are the alcohol variable
for assailants' drinking (i.e., the respondent's memory of what the assailant
had been drinking) and the length of time between an index incident and the
survey.
4.6. Indiscrimination in Acts of Violence and Alcohol

Pernanen defines these hypotheses as follows:
The "indiscrimination" hypotheses state that acts of aggression after
drinking will not be as well attuned as acts of sober aggression to the
requirements of the situations and the social norms applying to it, such
as the restraints (or "inhibitions") related to the location, the types of
acts performed, the characteristics of the target of aggression, and so
forth. . . . [A]cts performed would be as serious as in other social contexts,

regardless of normally attenuating factors.
There is some evidence in this work of “less discrimination” in the use of
violence in relation to how well the assailant and the victim know one anoth-
er. For example, in relation to the gender of the victim, Pernanen concludes
that his data contain “rather clear evidence of the continued importance of
conditional social-contextual cues and normative factors in the determination
of types of aggression and physical violence after drinking.” In the alcohol-
present episodes, more violent acts such as punching and kicking were used
against both male and female victims, but the difference in types of acts
between alcohol-present and alcohol-absent is small. Less severe and less
indiscriminate violence is generally used against female victims, and this does
not change substantially even when alcohol is involved in the incident and
when the assailant is drinking.
Pernanen2 concludes, “It can be said that once aggression occurs in con-
nection with drinking, it has the same general character of a ‘guided doing’
[using Goffman’s term] as in sober conflict.’’ Even in incidents involving both
violence and drinking, normative constraints are still operative. This is, of
course, consistent with the theories of drunken comportment of MacAndrew
and Edgerton54 and others. However, Pernanen is not yet prepared to declare
this debate over. He further argues,
Nonspecific “indiscrimination” and “excessiveness” may be more charac-
teristic of determination in the initial stage of a conflict, in the processes
involving instigating cues and cognitive issues in angry arousal, and in the
process by which these instigations produce open conflict.
These are important findings. However, we must question the extent to
which such findings are generalizable. Certainly they have relevance to every-
day violence, but as Pernanen himself points out, “Samples of violence that
occur in specific subcultures with more extreme drinking habits, such as ‘skid
row’ . . . could yield different results altogether.” Larger, more urban com-
munities with a greater representation of those who use excessive violence,
alcohol, and drugs may yield different results.
Although Pernanen is concerned with motivations and meanings, social
surveys such as his, however well carried out, do not allow us to pull out the
important scenarios that may give greater insights into alcohol involvement in
violence. Looking at “victim–offender” relationship, location, and so forth
separately is no substitute for getting into the context of violent events and
the “minds” of those involved. The fine work of Tony Parker,55,56 who has
sensitively interviewed a sample of men and women who are or have been
imprisoned for murder, is an important contribution. Although these are a
small number of “ideal typical” cases and are not meant to be a random
sample of such offenders, they give no support to theories that suggest that
substance abuse is a major determinant of homicidal behavior. Interviews of
this type done with violent offenders who were drinking or drinking heavily
at the time of the event would be an important contribution to work on

alcohol and violence.
5. The Future of Epidemiological Research

on Alcohol and Violence
One section of this chapter concentrated on a single type of violent be-
havior and another on a single study of violent behavior generally. The longer
review from which this chapter is drawn looks in similar detail at other types
of violent criminal behavior and at the research on domestic violence using
the same “windows” as in this chapter. What can we conclude from this
research about the relationship between alcohol and violence?
First, although there is a considerable alcohol presence in both offenders
and victims involved in violent events, there is evidence that they have many
other social, economic, and mental health problems.57 Additionally, alcohol
use is related to a number of situational variables that describe violent events.
These different types of variables are rarely included in the same piece of
research. The strength of the alcohol explanation is therefore not tested. In
addition, there is some evidence that the co-occurrence of multiple social and
health problems may preclude a clear explanation of alcohol’s relation to
many violent behaviors. This is, in part, a consequence of the multivariate
explanations of social behavior, and it is a problem that research and policy-
making have not adequately confronted.
Second, typologies of violent events that are theoretically driven are rare
in this research. Global divisions of behaviors into such categories as “violent”
versus “nonviolent,” or even groups of behaviors such as “homicide” or
“domestic violence,” do not offer enough specificity to establish clearly alco-
hol’s relationships with the behavior in question, although there is often
considerable alcohol presence in samples of these behaviors.
Third, empirical studies of alcohol and violence are typically unclear
about precisely what effects of alcohol are under investigation. Thus, in the
same piece of research the sociobehavioral effects of alcohol as an excuse for
untoward behavior are not distinguished from the pharmacological or other
effects. There is growing evidence that violence is a rational choice of particu-
lar actors. Yet alcohol-involved violence is often viewed as irrational, uncon-
trollable behavior. But these effects are often not clearly explicated. Often
researchers do not even address the question of why alcohol is included in
their research. Why, for example, do Koss and her colleagues include alcohol
and drugs as risk factors for sexual aggression? What theories of alcohol’s
effects lie behind the inclusion?
Fourth, the methodology of studying untoward events such as violence is
underdeveloped. This is particularly evident in the paucity of work on appro-
priate comparison groups for violent events. Finding appropriate “controls” is
essential if the use of alcohol by persons who are similarly situated in relation
to variables of theoretical importance in a given study is to be evaluated.
The past decade has seen changing perceptions of the victim–aggressor
relationship. This has been the case especially in relation to spousal violence
in which the woman as aggressor has been perceived by some courts, by
special interest groups, and by others as having committed justifiable violence
because of past victimization. This has also characterized some abusive par-
ent-child relationships that have involved violence, including homicide. The
possible contribution of alcohol and/or drugs to these violent outcomes must
be assessed not only in relation to the final violent event but to what might
have been an aggressive interaction over a long period of time, with the
aggression in some cases coming from each party at different times. The
victim-offender relationship in these cases is a mutable one. It is essential
then that the analyst is clear in denoting who is the victim and who is the
aggressor in relation to a specific act, and whether the analysis of the ”final”
act is the proper unit of analysis. Ideally, a time scale for the interaction
should be specified. The Pernanen-type questions (i.e., those used in Per-
nanen) do not clearly specify each of these. Indeed, the respondent is labeled
the “victim,” because the study is a “victimization” survey that takes the
perspective of the respondent, despite the fact that in an unknown proportion
of cases the “victim” may have “started” the conflict. An analysis of the time
scale of violent interactions and the use of alcohol over time in these interac-
tions will depend on more qualitative evidence than is typically collected.
This need for both quantitative and qualitative evidence analyzed within the
same study is a significant challenge for future work.
If epidemiological research on alcohol and violence is to contribute to our
understanding of the role of alcohol in violent events and violent lives, each of
these four factors needs considerably greater attention in future research.
Ironically, this progress may depend on the development of qualitative re-
search on the natural history of violent events rather than epidemiological
research itself.
ACKNOWLEDGMENT. An earlier version of this chapter appeared in “Alcohol

and Interpersonal Violence: Fostering Multidisciplinary Perspectives,” pub-
lished in 1993 by the National Institutes of Health as NIAAA Research Mono-
graph 24. Preparation of the chapter was supported by the Alcohol Research
Group, Berkeley, California, from its NIAAA Center Grant.
References
1. Roizen J: Alcohol and violence: An evaluation of the evidence from epidemiological research.
Prepared for the Working Group on Alcohol-related Violence: Fostering Interdisciplinary
Perspectives, convened by the National Institute on Alcohol Abuse and Alcoholism, Wash-
ington, DC, May 14-15, 1992.
2. Pernanen K: Alcohol in Human Violence. New York, Guilford Press, 1991.
3. Leonard KE, Jacob T: Alcohol, alcoholism and family violence, in Van Hasselt V, et al (eds):
Handbook of Family Violence. New York, Plenum Press, 1990, pp 383–406.
4. Kuhn T: Structure of Scientific Revolutions. Chicago: University Press, 1970.
5. Collins JJ Jr: Alcohol use and criminal behavior: An empirical, theoretical, and methodologi-
cal overview, in Collins JJ Jr (ed): Drinking and Crime: Perspectives on the Relationships between
Alcohol Consumption and Criminal Behavior. New York, Guilford Press, 1981, pp 288- 316.
6. Stinchcombe A: The origins of sociology as a discipline. Acta Sociol 27:51–61, 1984.
7. Collins JJ Jr: Alcohol and interpersonal violence: Less than meets the eye, in Weiner N,
Wolfgang M (eds): Pathways to Criminal Violence. Newbury Park, CA, Sage Publications, 1989,
pp 49–67.
8. Roizen J: Alcohol and trauma, in Giesbrecht N, Gonzalez R, Grant M, et al (eds): Drinking and
Casualties. London, Routledge, 1989, pp 21- 66.
9. Murphy JM, Jellinek M, Quinn D, et al: Substance abuse and serious child mistreatment:
Prevalence, risk, and outcome in a court sample. Child Abuse Neglect 15:197–211, 1991.
10. Orme TC, Rimmer J: Alcoholism and child abuse: A review. J Stud Alcohol 42:273–287, 1981.
11. Parker RN: Bringing " Booze" Back In: The Relationship between Alcohol and Homicide. Berkeley,
CA, Prevention Research Center, 1992.
12. Abbey A: Acquaintance rape and alcohol consumption on college campuses: How are they
linked? J Am College Health 39:165–169, 1991.
13. Koss MP, Dinero TE: Discriminant analysis of risk factors for sexual victimization among a
national sample of college women. J Consult Clin Psychol 57:242–250, 1989.
14. Muehlenhard CL, Linton MA: Date rape and sexual aggression in dating situations: Inci-
dence and risk factors. J Counsel Psychol 34:186–196, 1987.
15. Pernanen K: Alcohol and crimes of violence, in Kissin B, Begleiter H (eds): The Biology of
Alcoholism: Social Aspects of Alcoholism. New York, Plenum Press, 1976, pp, 351–444.
16. Cordilia A: Alcohol and property crime: Exploring the causal nexus. J Stud Alcohol 46:161–171,
1985.
17. Petersilia J, Greenwood PW, Lavin M: Criminal Careers of Habitual Felons. Washington, DC, US
Government Printing Office, 1978.
18. Ladouceur P, Temple M: Substance abuse among rapists: A comparison with other serious
felons. Crime Delinquency 31:269–294, 1985.
19. Aarens M, Cameron T, Roizen J, et al (eds): Alcohol, Casualties and Crime. Alcohol, Casualties
and Crime Project Final Report, Report No. C-18, Social Research Group, University of
California, Berkeley, 1977.
20. Greenberg SW: Alcohol and crime: A methodological critique of the literature, in Collins JJ Jr
(ed): Drinking and Crime: Perspectives on the Relationships between Alcohol Consumption and Crimi-
nal Behavior. New York, Guilford Press, 1981, pp 70–109.
21. Pernanen K: Theoretical aspects of the relationship between alcohol use and crime, in Collins
JJ Jr (ed): Drinking and Crime: Perspectives on the Relationships between Alcohol Consumption and
Criminal Behavior. New York, Guilford Press, 1981, pp 1–69.
22. Roizen J, Schneberk D: Alcohol and crime, in Aarens M, et al (eds): Alcohol, Casualties and
Crime. Alcohol, Casualties and Crime Project Final Report No. C-18, Social Research Group,
University of California, Berkeley, 1977.
23. Barnard GW, Holzer C, Vera H: A comparison of alcoholics and non-alcoholics charged with
rape. Bull Am Acad Psychiaty Law 7:432–440, 1979.
24. Straus MA, Gelles RJ: Societal change and change in family violence from 1975–1985. J
Marriage Family 48:465–479, 1986.
25. Greenfield TK, Weisner C: Drinking problems and self-reported criminal behavior, arrests
and convictions: 1990 national alcohol and 1989 county surveys, presented at the American
Public Health Association 120th Annual Meeting, Washington, DC, November 8- 12, 1992.
26. Wolfgang ME: Patterns in Criminal Homicide. Philadelphia, University of Pennsylvania Press,
1958.
27. Roizen J: Estimating alcohol involvement in serious events, in Alcohol Consumption and Related
Problems. Alcohol and Health Monograph No. 1. DHSS Publication No. (ADM 82-1190),
Washington, DC, US Government Printing Office, 1982, pp 179–219.
28. US Department of Justice, Federal Bureau of Investigation: Crime in the United States: Uniform
Crime Reports. Washington, DC, US Government Printing Office, 1988.
29. US Department of Justice, National Criminal Justice Information and Statistics Service: Census
of State Correctional Facilities 1974 Advance Report. NPS Special Report, July. Washington, DC,
US Government Printing Office, 1975.
30. US Department of Justice, Bureau of Justice Statistics: Criminal Victimization in the United
States, 1982 (publication No. NCJ-92820). Washington, DC, US Government Printing Office,
1984.
31. Dietz PE: Social factors in rapist behavior, in Rada RT (ed): Clinical Aspects of the Rapist. New
York, Grune & Stratton, 1978.
32. Clark L, Lewis D: Rape: The Price of Coercive Sexuality. Toronto, The Women’s Press, 1977.
33. Selling LS: The role of alcohol in the commission of sex offenses. Med Rec 151:289–291, 1940.
34. Visher C: Reported use of alcohol by DU arrestees. Drugs and Crime 1990 Annual Report:
Drug Use Forecasting. Washington, DC, National Institute of Justice, 1990.
35. Amir M: Patterns in Forcible Rape. Chicago: University of Chicago Press, 1971.
36. Johnson SD, Gibson L, Linden R: Alcohol and rape in Winnipeg, 1966–1975. J Stud Alcohol
39:1887–1894, 1978.
37. Richardson D, Campbell JL: The effect of alcohol on attributions of blame for rape. Pers Soc
Psychol Bull 8:468–476, 1982.
38. Deming JE, Mittleman RE, Wetli CV: Forensic science aspects of fatal sexual assaults on
women. J Forensic Sci 28:572–576, 1983.
39. Gibbens T, Soothill K, Way C: Sex offenses against young girls: A long-term record study.
Psychol Med 11:351–357, 1981.
40. Normandeau A: Trends and Patterns in Crimes of Robbery. Ann Arbor, University Microfilms
International, 1968.
41. Amir M: Alcohol and forcible rape. Br J Addict 62:219–232, 1967.
42. Pittman DJ, Handy W: Patterns in criminal aggravated assault. J Criminal Law, Criminol Police
Sci 55:462–470, 1964.
43. US Department of Justice, Bureau of Justice Statistics: Survey of Inmates of State Correctional
Facilities, 1979. (ICSPR 7856) 2nd ICSPR ed. Ann Arbor, International Consortium for Social
and Political Research, 1981.
44. Collins JJ Jr, Schlenger WE: Acute and chronic effects of alcohol use and violence. J Stud
Alcohol 49:516–521, 1989.
45. Groth AM, Birnbaum HJ: Men Who Rape: The Psychology of the Offender. New York, Plenum
Press, 1979.
46. Rada RT: Alcoholism and forcible rape. Am J Psychiatry 132:444–446, 1975.
47. Koss MP, Gidycz CA, Wisniewski N: The scope of rape: Incidence and prevalence of sexual
aggression and victimization in a national sample of higher education students. J Consult Clin
Psychol 55: 162–170, 1987.
48. Koss MP, Dinero TE: Predictors of sexual aggression among a national sample of male college
students. Ann NY Acad Sci 528:133–146, 1988.
49. Johnson LD, O’Malley PM, Bachman JG: Drug Use, Drinking and Smoking: National Survey
Results from High School, College and Young Adult Populations, 1975–1988. National Institute on
Drug Abuse, DHHS Pub. No. (ADM) 86-1638, 1989.
50. Engs R, Hanson D: The drinking patterns and problems of college students. J Alcohol Drug
Educ 31:65–83, 1983.
51. Gleason N: Toward a Model for Preventing Alcohol Abuse by College Women: A Relational Perspec-
tive. The Stone Center for Developmental Services and Studies, Wellesley College, 1992.
52. Koss MP, Dinero TE, Seibel C, et al: Stranger, acquaintance and date rape: Is there a difference
in the victim’s experience? Psychol Women Q 121–24, 1988.
53. Kantor GK, Straus MA: The drunken bum theory of wife beating. Social Problems 34:213–231,
1987.
54. MacAndrew C, Edgerton R: Drunken Deportment: A Social Explanation. Chicago, Aldine, 1969.
55. Parker T: Life After Life. London, Martin Secker and Warburg, 1990.
56. Parker T: The Violence of Our Lives. London: HarperCollins, 1995.
57. Abram KM, Teplin LA, McClelland GM: The effects of co-occurring disorders on the relation-
ship between alcoholism and violent crime, in Martin SE (ed): Alcohol and Interpersonal Vio-
lence: Fostering Multidisciplinary Perspectives. NIH Publication No. 93-3496, Rockville, MD,
National Institutes of Health, 1993.
58. Shupe LM: Alcohol and crime: A study of the urine alcohol concentration found in 882
persons arrested during or immediately after the commission of a felony. J Criminal Law,
Criminol Police Sci 44:661–664, 1954.
59. President’s Commission on Crime: Report on the President’s Commission on Crime in the District
of Columbia. Washington, DC, US Government Printing Office, 1966.
60. Tardif G: La Criminalite de Violence. MA Thesis, University of Montreal, 1966.
2
The Relationship of Alcohol

to Injury in Assault Cases
Susan Ehrlich Martin and Ronet Bachman
Abstract. Little is known about the precise role of alcohol in the escalation of interactions from
threats into physical violence or its contribution to the risk of injury. Experimental studies
indicate that intoxicated subjects (allegedly) give markedly higher electric shocks than sober
subjects and are less sensitive to their cries of pain. However, few studies in a naturalistic setting
have examined whether aggressive acts become more serious and result in higher injury rates
when the assailants have been drinking than when they are sober. This chapter reviews the two
bodies of research on the effects of alcohol on interpersonal aggression and violence; presents
new data on the escalation of threatening interactions to assaults and the likelihood of victim
injury given an assault, using data from the National Crime Victimization Survey for the years
1992 and 1993; and suggests future directions for research based on our findings that alcohol's
impact on both escalation and injury differed according to the victim–assailant relationship.
In the past three decades, more than 100 studies have confirmed the wide-
spread belief that alcohol frequently has been consumed by offenders and
victims prior to violent incidents. These findings have been summarized in
several reviews of the literature on alcohol consumption and violence1–7 and
indicate that a majority of homicide and assault cases involve alcohol use.
Another body of research, focused on the contribution of alcohol to aggres-
sion in a laboratory setting, also has found that alcohol is a potent antecedent
of aggressive behavior. Nevertheless, there is surprisingly little empirical data
concerning alcohol's precise role in the escalation of hostile interactions or
threatening situations into physical violence, particularly assaults that do not
Susan Ehrlich Martin • Prevention Research Branch, Division of Clinical and Prevention Re-
search, National Institute on Alcohol Abuse and Alcoholism, Rockville, Maryland 20892-7003.
Ronet Bachman • Department of Sociology, University of Delaware, Newark, Delaware 19716.
41
result in homicide. Similarly, knowledge is limited regarding the contribution

of alcohol to the risk of injury or injury severity in incidents in which violence
occurs. Nor are the mechanisms or mediating factors by which alcohol con-
sumption affects aggression well understood.
This chapter reviews the research literature on the contribution of alcohol
to: (1) the escalation from threats to violence in naturalistic settings; (2) the
risk of victim injury and injury severity given an assault; and (3) acts of
aggression in experimental laboratory settings. It next tests two hypotheses
regarding the contribution of alcohol to incident severity and to the perpetra-
tor’s lack of discrimination in the use of violence, using a large nationally
representative sample of victims of violence from the National Crime Victim-
ization Survey. It concludes with a discussion of explanatory frameworks and
future research directions for studying alcohol’s contribution to assault.
1. Research on Alcohol in Human Violence

Evidence regarding the role of alcohol in human violence comes primari-
ly from two types of research: correlational studies of naturally occurring
alcohol use and violent behavior and laboratory experiments.
1.1. Correlational Studies

Surveys have typically focused on real-life incidents of violence reported
by victims or perpetrators and have frequently found associations between
violence and drinking or drunkenness. For example, in an update of an earlier
study, Roizen8 reported that in 14 studies of homicide, alcohol was present in
between 28 and 86% of the offenders; in eight studies of assaults, it was
present in 24 to 72% of the assault offenders. Alcohol also was present in 21 to
50% of the victims or assailants in seven studies of marital violence. Another
review of studies of alcohol and violent crime that included five studies with
data on assaults, using data primarily from police records, found that be-
tween 24 and 82% of assault offenders and between 24 and 40% of assault
victims had been drinking prior to the assault.3
In addition to data on alcohol prevalence, findings from several studies
suggest that alcohol is positively associated with the severity of injuries re-
sulting from violence. One study examined the contribution of alcohol to
injury by comparing the frequency of its presence in incidents of specific
types of aggression of differing degrees of severity. In all four comparisons,
alcohol was more likely to be involved in the act producing the greatest
injury.9 For example, 28% of marital cases of threatening behavior but 49% of
marital assaults and 34% of common assaults but 46% of assaults causing
injuries involved alcohol.
Surveys of incarcerated offenders also find they frequently used alcohol
2 • Relationship of Alcohol to Injury 43
alone or in combination with other drugs prior to the assault for which they
were locked up, including 34% of youth in custody10 and 54% of jail in-
mates.11 While these data are suggestive of an alcohol–violence association,
samples of arrested or incarcerated offenders are neither representative of all
violent offenders nor of their patterns of alcohol consumption.
Research focused on victims of violence, such as emergency room (ER)
studies of injuries using a case-control methodology, also has found alcohol
consumption to be associated with violence-related fatal and nonfatal inju-
ries8,12,13 (for a review of injury studies). For example, one study found that
29% of persons with severe violence-related injuries reported drinking more
than 10 drinks in the 12 hr preceding their injury compared with 18% with
only minor injuries, suggesting a positive association between the degree of
victim’s intoxication and the severity of the injury.14 A Finnish study found
that alcohol intoxication increased the likelihood of hospitalization for male
assault victims treated in an ER (odds ratio 1.5; P < 0.06).15 A third study
reports a higher proportion of coroner cases (47%) that were alcohol-related
compared with violence-related ER nonfatal injury cases (19%).16
Emergency room studies primarily focus on the presence of alcohol in the
blood, breath, or tissues of a victim of violence who receives medical treat-
ment. They rarely ascertain data on the presence of alcohol in the perpetrator
of the violence or document the processes of violence escalation. Thus, the
contribution of ER studies to understanding of assault severity is limited.
In the past two decades, dozens of new studies have focused specifically
on the role of alcohol in violence among intimates. Several reviews have
pointed out the methodological shortcomings of such studies and their con-
sistent finding that drinking is associated with intimate-perpetrated domestic
violence.17–20 Often the research has focused on distal influences such as
husband’s drinking pattern and socioeconomic status. While such an ap-
proach can identify high-risk groups, it sheds little light on the underlying
processes that lead to marital violence and on the proximal influence of alco-
hol consumption on the event.19
Evidence regarding acute effects of drinking on the escalation of conflicts
to violence among intimates and the interactions of alcohol effects with con-
textual factors in the escalatory processes is quite limited. Estimates of the
proportion of violent events among intimates that are associated with alcohol
use range from 2220 to 60%.21 It is noteworthy, however, that in many of these
violent incidents only the assailant (usually the man) is drinking, whereas in
violent incidents involving other victim–assailant relationships it is more like-
ly that both the offender and victim have been drinking. For example, police
records from one community over a 1-year period found that in reports in-
volving marital abuse, 44% of the assailants but only 14% of the victims had
been drinking (compared with 31 and 26% of all of the violent crime inci-
dents).6 The 1985 National Family Violence Resurvey (NFVR), based on a
nationally representative sample of 5159 couples,20 found that 22% of the
husbands and 10% of the wives had been drinking prior to incidents of
spousal violence (while overall, 24% of the incidents involved drinking).
Studies of the effect of alcohol on the severity of the spousal violence,
usually measured as the extent of injuries, provide inconsistent findings.
Fagan and colleagues22 reported that the severity of spouse abuse was pos-
itively associated with alcohol use by the assailant (while there was a weak,
negative association with use of other substances). Conversely, an analysis of
262 domestic disturbance calls to which the police responded found that the
man’s drinking pattern was related to the severity of the victim’s injuries but
drinking by the offender and victim were unrelated to the injury.23 Kantor
and Straus’24 reanalysis of NFVR findings found that the husbands drunken-
ness was associated with higher rates of minor violence but not severe vio-
lence against their wives.
This finding, as well as those of more than 100 studies of spouse abuse in
the past 15 years, is based on the Conflict Tactics Scale (CTS), which classifies
cases of spousal assault as minor or severe in a way that corresponds to the
legal distinctions between simple and aggravated assault.25,26 Thus, the CTS
has a scale for physical aggression that is subdivided into minor and severe
violence. Minor violence includes pushing, grabbing, and slapping; severe
violence includes kicking, punching, hitting with an object, choking, beating
up, and using a knife or gun. It also includes threatening with a knife or gun,
since such threats meet the legal definition of assault. But threats do not
involve actual physical contact or violence, although they constitute nearly a
third of all assaultive incidents between intimates and more than half of all
assaults in a nationally representative sample of victimization.27 Kantor and
Straus’24 inclusion of threats in their measure of severe violence, therefore,
may be partly responsible for their failure to find drunkenness to be a predic-
tor of serious violence among intimates.
In a uniquely comprehensive study of the role of alcohol in human vio-
lence that used multiple methods to study one community, Kai Pernanen5
conducted interviews with 933 community residents aged 20 and over drawn
from a probability sample of the general population in one medium-sized
Canadian community (Thunder Bay, Ontario). Of those interviewed, 435 re-
ported experiencing violent victimization and provided details on their most
recent experience involving an actual physical assault, including the drinking
behavior of both the assailant and victim prior to the assault.
Pernanen explored two hypotheses regarding the role of alcohol in natu-
rally occurring violent events. The “severity” hypothesis argues that the rate
and extent of injuries will be more serious when assailants have been drink-
ing than under similar conditions when the assailants are sober. The “indis-
crimination” hypothesis states that acts of aggression after drinking are less
likely to be tailored to the situation and social norms applying to it than acts of
sober aggression. For example, drinking would be expected to loosen re-
straints on serious violence directed at “restricted” targets such as women
and children; if such acts occur after drinking, they are expected to be as
serious as sober acts.
Although alcohol was involved in more than half of the 435 incidents
examined in detail, Pernanen’s data provided no support for a general severi-
ty hypothesis. Injury rates were closely related to the age, gender, and rela-
tionship of victim and perpetrator, but were not consistently associated with
whether the assailant was drinking. For example, men were more likely to be
injured when their assailants had been drinking and women were at higher
risk when their assailants were sober. Nevertheless, the finding that the
injury rate was significantly higher when the assailant was judged to be
drunk provides conditional support for the severity hypothesis.
Support for the indiscrimination hypothesis was similarly inconsistent.
There was a tendency to greater rather than diminished discrimination by
gender of victim when alcohol had been used. Drinking, even drunkenness,
did not lead men to abandon the normative restrictions regarding the types of
violence deemed permissible when a man is fighting a woman. Drinking,
however, did increase likelihood that victimizations by strangers would result
in injury.
1.2. Experimental Studies

Experimental research provides the second source of evidence on the role
of alcohol in human violence and served as the primary basis for severity and
indiscrimination hypotheses. Reviews of the many studies of alcohol-induced
aggression under controlled conditions have concluded that alcohol is a po-
tent determinant of aggression.28–30 Pernanen’s severity hypothesis is predi-
cated on studies using variations of the Buss31 “aggression machine,” which
measures aggression in the form of the intensity and/or duration of electric
shocks (allegedly) administered by subjects to other subjects.32–36 In general,
these studies have found that intoxicated subjects give markedly higher
shocks than sober subjects, particularly under conditions of frustration, prov-
ocation,37 or threat.38
Similarly, there is experimental support for the indiscrimination hypoth-
esis. For example, following administration of controlled doses of alcohol,
researchers have found that subjects showed less sensitivity to cries of pain
from a decoy victim39 and were less sensitive to empirical contingencies in the
evaluation of feedback and in the aggression committed.40 However, experi-
mental findings also have shown that intoxicated subjects are influenced by
social pressure from a “bystander” and by social norms.41–43
The prevailing explanation for these findings is that alcohol has an im-
pairing effect on information processing. This view was expanded by Steele
and Joseph44 into a broader theory of the affective and interpersonal conse-
quences of “alcohol myopia.” However, in the past several years there has
been little follow-up research testing measures of information processing or
mediating factors such as emotions that may be affected by attentional impair-

ment.45
1.3. Limitations of Existing Studies

The limitations of both correlational and experimental studies suggest the
need for additional research on the effects of alcohol on the escalation and
severity of injury in assaults. The principal problems of laboratory studies are
the limitations related to the artificiality of the drinking situation and the
difficulty of successfully carrying off variations of the balanced placebo design
that include the “expectancy effect” that has been found for direct physical
aggression. Other problems include the extent to which pain thresholds vary
across trials, the putative analgesic effect of alcohol, and the failure of the
commonly used aggression paradigms to study the interactions between
people who know each other.45
Correlational studies in general and Pernanen’s analyses in particular
also have a number of limitations. First, most studies do not include incidents
of threat only and therefore cannot examine the contribution of alcohol con-
sumption to the risk of an incident escalating from threat to physical attack.
Second, small sample size limited Pernanen‘s ability simultaneously to con-
trol for other important contextual characteristics of the incident (e.g., both
gender and the victim–assailant relationship) in examining the contribution
of alcohol to the risk of injury.
2. The NCVS Study

To ameliorate the shortcomings and to test more refined versions of the
severity and indiscrimination hypotheses, we used data from a much larger,
nationally representative sample of the US population aged 12 years and
over. We focused on the effects of alcohol on assault severity and on discrimi-
nation in the use of violence for five types of assault determined by the victim’s
and offender’s gender and the victim–offender relationship. Specifically, an-
alyses were performed separately for men’s assaults on male friends and
acquaintances and on male strangers and for men’s assaults against women
who were intimates (i.e., husbands, boyfriends, and former husbands and
boyfriends), acquaintances, and strangers. The initial descriptive analysis re-
vealed too few cases of female-perpetrated assaults involving alcohol to in-
clude in multivariate analysis.
2.1. Sample
The data used in our analysis come from the Bureau of Justice Statistics-
sponsored National Crime Victimization Survey (NCVS) for 1992 and 1993.
The NCVS obtains information about crimes, including incidents not reported
to the police, from a continuous, nationally representative sample of house-

holds in the United States, which is more fully detailed elsewhere.27 The
redesign has added questions that directly ask respondents about attacks that
were perpetrated by relatives or other persons known to them, permitting
collection of fuller information about intimate violence.46 Prior to 1992, NCVS
estimates of intimate-perpetrated violence against women were much lower
than estimates obtained from other national surveys,47 probably because the
NCVS did not directly ask respondents about attacks that were perpetrated
by relatives or other offenders known to them. Prompted by criticism of this
earlier methodology, a new NCVS screening instrument gradually was imple-
mented into the sample beginning in 1989. By 1992, the entire NCVS sample
was using the redesigned survey and the number of victimizations reported
to interviewers, particularly assaults perpetrated by intimates, had greatly
increased.48
In this chapter we focus exclusively on single-offender assaults. The
NCVS defines assault as an unlawful physical attack, including attempted or
threatened attacks upon a person. The offenses categorized as assaults in our
analyses range in severity from minor incidents such as attempted and verbal
threats of assault to more severe incidents such as those involving use of
weapons and resulting in injury. However, because NCVS classifies victim-
izations according to a seriousness hierarchy, a number of the more serious
incidents may have been excluded from the analysis. For example, an assault
also involving theft is classified as a robbery; similarly, rapes and attempted
rapes are categorized separately.
2.2. Measures
2.2.1. Dependent Variables: Assault Severity. The severity of assault was
operationalized in two ways. The first dependent variable we examine indi-
cates whether or not the assault actually culminated in a physical attack by the
offender. To operationalize this variable victims were asked, "Did the of-
fender hit you, knock you down, or actually attack you in any way?" In the
multivariate analyses, the variable was coded 1 if the answer was yes and 0 if
only a threat of assault was involved.
The second dependent variable indicated whether victims sustained any
injuries. In the multivariate analysis, this variable was coded 1 if the victim
was injured and 0 if no injuries had been sustained. In the bivariate analysis
displayed in Tables II–IV, however, we treated these variables as three values
of a single assault variable, namely, threat only, assault without injury, and
assault with injury. We would have liked to have determined the extent to
which alcohol increased the severity of injuries sustained, adding a "severe
injury" category. However, because the distribution of assaults resulting in
injury from this sample primarily consisted of minor injuries such as bruises
and scratches, there was not enough variation in injury severity to investigate
the extent to which alcohol increased the seriousness of an injury.
2.2.2. Independent Variables. A number of independent variables were

used in the analyses, including alcohol use, location of occurrence, victim–
assailant relationship, and demographic controls including victim’s gender,
age, and marital status. The determination of assailant’s alcohol use was
based solely on victim perceptions. Victims were asked, “Was the offender
drinking or on drugs?” If the reply was affirmative, they were next asked,
“Which was it, drinking, drugs, or both?” If victims said the offender had
been drinking, this variable was coded 1. If the assailant was not perceived to
be under the influence of any substance, this variable was coded 0. To make
the comparison categories as pure as possible, all assailants perceived to be
under the influence of drugs other than alcohol or perceived to be under the
influence of both substances were excluded from the analysis.
The location of the occurrence was dichotomized as occurring in a public
place (coded 1) or at or near a private residence (including the home of the
victim, a friend, or neighbor; coded 0). Age of victim was a continuous vari-
able ranging from 12 to 84 years. Marital status was coded 1 if married, 0 for
single, separated, divorced, or widowed.
2.3. Analytic Procedures

The effects of alcohol on assault severity were examined for five different
types of assault, determined by the victim’s and assailant’s gender and the
victim–assailant relationship. Specifically, analyses were performed separate-
ly for male-on-male assaults involving friends and acquaintances and those
involving strangers. Male assaults against females were examined separately
for intimates, acquaintances, and strangers. Intimate was defined as all hus-
bands or ex-husbands, and boyfriends or ex-boyfriends. Nonintimate rela-
tives were excluded from the analyses.
We first conducted bivariate analyses predicting the assault severity mea-
sure using assailant’s alcohol use only. For relationships found to be signifi-
cant at the bivariate level, we performed multivariate logistic regression anal-
ysis to control for other factors that may affect assault severity.
2.4. Findings
The descriptive characteristics of the victim sample and the victimization
category are displayed in Table I. This table indicates that most “assault”
incidents (58%) involved threats rather than actual physical attacks. Men
were more likely than women to be victims of assaultive incidents, but female
victims were more likely than male victims actually to be both physically
attacked and injured.
Escalation from threat to attack also was closely related to the victim’s
age; the younger the victim, the greater the likelihood of an actual physical
assault. Most attacks occurred in public places, but those that occurred in
private places were more likely to result in an attack and injury to the victim.
Table I. Characteristics of Victim Sample by Percent in Victimization Category

Assault Assault
Threat without injury with injury
Total (N = 5592) 58% 19% 23%

Gender of victim
Male (N = 3234) 60 20 20
Female (N = 2358) 55 18 27
Age of victim
12–18 (N = 1587) 46 25 29
19–29 (N = 1679) 56 19 25
30–45 (N = 1732) 65 16 19
46 and up (N = 594) 69 14 17
Location of incident
Public (N = 3973) 61 19 20
Private (N = 1619) 49 20 31
Victim–assailant relationship
Intimate (N = 613) 32 22 46
Acquaintance (N = 1650) 55 22 23
Stranger (N = 1321) 68 19 13
Assailant’s perceived alcohol use
Not drinking (N = 2064) 58 20 22
Had been drinking (N = 941) 50 23 27
Most assaults involved people who know each other, and the closer the rela-
tionship between victim and assailant, the greater the likelihood of escalation
to attack.
Table I also indicates that most assailants were not perceived to have been
drinking prior to the assault incident. Nevertheless, the consumption of alco-
hol increased the chance that an incident would result in physical violence
and victim injury, providing preliminary support for the severity hypothesis.
In Table II, we explore the joint percentage distributions of the assault
severity measures by assailant’s alcohol use and present the results of the Chi
square analysis. The upper portion of the table presents male-on-male as-
saults by victim–assailant relationship. For assaults involving friends or ac-
quaintances, both measures of assault severity appear to be unaffected by the
perpetrator’s alcohol use. Similar percentages of assaults culminate in actual
attacks in the presence and absence of alcohol; likewise, the proportion of
male victims that sustained injuries was unchanged by whether the assailant
had been drinking.
In contrast to the findings for acquaintance assaults, alcohol did affect the
severity of male-on-male assaults involving strangers. Men who had been
drinking were significantly more likely to physically attack another man dur-
ing an assault incident (39%) than men who had not been drinking (26%). The
assailant’s use alcohol did not, however, increase the likelihood that the as-
sault would result in injury to the victim.
Assaults involving male perpetrators and female victims are presented in

the lower portion of Table II. For two of the three relationships, assault severi-
ty escalated if the man was perceived to have been drinking. Both the percent-
age of women actually assaulted and the percentage of women who were
injured increased if a male intimate or stranger had been drinking; a male
acquaintance's drinking did not affect the incident. The only relationship to
attain statistical significance, however, was the increased likelihood of injury
for women assaulted by intimates who had been drinking (54%) in contrast to
nondrinking intimates (43%).
We next examine whether the use of alcohol by the assailant increases the
probability of an assault escalating to a physical attack even after controlling
for other important contextual characteristics of the assault. Table III displays
the results for the logistic regression analysis predicting a physical attack
(coded 1) versus a threat only (coded 0) for male-on-male assaults involving
strangers. This table indicates that men's encounters with male strangers
were significantly more likely to escalate from threat to physical attack when
the assailant had been drinking compared to incidents in which he had not. In
Table II. Contribution of Alcohol to Assault Victimizations by Victim and Assailant

Gender and Relationship
Assailant perceived Assailant not perceived

as using alcohol as using alcohol
Male assailant/male victim

Acquaintance assaults
Threat 50% 50%
Assault without injury 26 27
Assault with injury 24 23
Stranger assaults
Threat 61a 74a
Male assailant/female victim
Intimate assaults
Threat 26 34
Assault with injury 54 a 43a
Acquaintance assaults
Threat 64 64
Stranger assaults
Threat 64 74
Indicates Chi-square between assailants perceived to be using and not to be using alcohol was significant at
a
P < 0.05 level.

Table III. Logistic Regression of Male Physical Attack on Male Strangers

Controlling for Other Contextual Characteristics
B S.E. Sig. Exp (B)
Assailant drinking .6300 .1538 .0001 1.877

Incident in public place – .2221 .2591 .3914 .800
Victim’s age – .0300 .0069 .0001 .970
Model Chi-square = 38.55; P < 0.0001
fact, the odds of an assault resulting in a physical attack were almost twice as
great for assailants who have been drinking (Exp(B) = 1.877) compared to
those not under the influence of alcohol.
In addition to the assailant’s use of alcohol, the victim’s age was also
significant in predicting assault severity in male-on-male assaults involving
strangers. Assaults involving younger victims were significantly more likely
to result in physical attacks than assaults involving older victims. Whether the
assault occurred in a public place, however, did not affect the likelihood of a
male-on-male stranger incident escalating into a physical assault.
Table IV presents the results of the logistic regression model predicting
the probability that female victims sustained injuries from the assault (coded
1) versus assaults that did not result in injury (coded 0) by male assailants
who were intimates, while simultaneously controlling for location and vic-
tim’s age and marital status. The results of this logistic regression model
indicate that alcohol use by the perpetrator was significant in predicting as-
sault severity (i.e., victim injury), even after controlling for location, marital
status, and victim’s age. The odds of a woman sustaining an injury in an
assault by an intimate partner who had been drinking increased by a factor of
1.49 compared to women attacked by a nondrinking partner. In addition,
injury was significantly more likely when the assault occurred in a private
home. Age and marital status did not affect assault severity for male-on-
female assaults involving intimates.
Table IV. Logistic Regression of Male Assailants’ Injury of Intimate Female Victims,
Controlling for Other Contextual Characteristics
B S.E. Sig. Exp (B)
Assailant drinking .4023 .1820 .0271 1.495

Incident in public place – .4725 .2421 .0509 .623
Victim’s age – .0068 .0083 .4095 .993
Married .2162 .2437 .3750 1.241
Model Chi-square = 11.42; P < 0.0222

2.5. Discussion
Based on data from the NCVS, we have found that the effects of drinking
by the assailant on the escalation and outcome of assaultive behavior varies
according to the assault context. Specifically, alcohol's contribution to the
escalation of an incident from a threat to a physical attack and to injuries
sustained by the assault victim varies with the gender of the victim and the
victim–assailant relationship.
In male-on-male acquaintance assaults, alcohol has no effect either on the
likelihood of a threat escalating to an attack or on it resulting in injury. Sim-
ilarly, alcohol has no effect on either outcome in male-on-female acquaintance
attacks, although a smaller proportion of these incidents (36%) than male-on-
male incidents (50%) result in actual attacks regardless of alcohol involve-
ment.
The pattern of escalation from threat to assault with and without victim
injury is similar in male-on-male and male-on-female stranger assaults, al-
though alcohol's effect on the escalation only achieves statistical significance
in the male-on-male incidents. Multivariate analysis indicates that even after
controlling for other variables, alcohol increases the likelihood that a threat
will escalate to an attack in male-on-male incidents but does not increase the
probability of injury to the victim.
Assailant alcohol use also affected the outcome of male-on-female as-
saults between intimates. While alcohol use by the assailant did not increase
the likelihood that an intimate partner would escalate his threat to physical
violence, drinking by the perpetrator did increase the likelihood that the
female victim would sustain an injury even after controlling for her marital
status and age and the location of the incident. It should also be noted that
male assaults on intimate female partners were more likely to result in actual
attacks and in victim injury compared to any other victim–offender relation-
ship regardless of alcohol use by the offender.
Our findings differ somewhat from those obtained by Pernanen. After
controlling simultaneously for victim–perpetrator relationship and gender,
we found that men's drinking increases the already-elevated probability of in-
jury to an intimate female partner. This suggests that violent acts are less lim-
ited and tailored to the situation and relationship in the presence of alcohol.
The extremely high rate of injury to female intimates from male partner
violence (46% of all such assaults irrespective of alcohol involvement and 54%
of the assaults when the man had been drinking) in contrast to the injury rates
in all other victim-perpetrator categories needs some explanation. Our find-
ings are consistent with other studies that have found that alcohol abuse is a
relatively strong and consistent correlate of partner aggression.18–20,49 Going
beyond measures of the prevalence of alcohol in intimate violence, however,
we have found that alcohol also significantly increases the likelihood of injury
in such incidents. It is unclear, however, the extent to which psychological,
situational, and pharmacological factors contribute to women’s alcohol-re-

lated injury by an intimate.
An alternative interpretation of this finding, however, may be related to
the NCVS data. Although the redesign has increased the rate at which victim-
izations by intimates are reported, women may still use a different threshold
for reporting violence by an intimate than they use in similar incidents involv-
ing acquaintances and strangers. Threats by nonintimate men may be viewed
as more unusual, more memorable, and thus more likely to be reported in a
survey, whereas spousal threats may rarely be regarded as “criminal” or
worth reporting in a survey except when they culminate in an attack or injury.
Several other limitations of the NCVS data supported here should be
noted. The alcohol-related information contains only the victim’s report of the
perpetrator’s drinking and indicates only the apparent presence of alcohol.
There is no information on the victim’s use of alcohol, which is a serious
limitation given the frequency with which the victim has been found to have
been drinking in assault and spousal assault cases.3,6,7,20 Furthermore, data
are lacking on the amount consumed by the perpetrator. Moreover, the pres-
ence of alcohol does not necessarily indicate that it had a casual role. The
NCVS data also lack information on many cognitive influences (e.g., provoca-
tion) and cultural expectancies (e.g., the belief that violence toward a spouse
is acceptable under certain circumstances) that appear to mediate alcohol-
related behavior.
Despite these limitations, our findings provide the first examination of
the contribution of alcohol to the escalation of violence and its contribution to
victim injury based on a nationally representative sample of naturally occur-
ring episodes of violence. They suggest that drinking by a male assailant does
increase the likelihood that he will actually physically attack the victim in an
interpersonal confrontation, particularly when the other is a female intimate.
3. Future Research Directions

The limited prior research coupled with the findings from the NCVS
presented here suggest the need for expanded research focusing on the con-
tribution of alcohol consumption to the escalation of interactions from threats
to actual attacks and the severity of injury given a physical attack. Such
research should bridge the theoretical gap between the naturalistic perspec-
tive that rests on the view that alcohol’s pharmacological effects disinhibit the
drinker and the sociocultural perspective that treats drunken behavior as
purely determined by social norms.6,44 This requires developing and testing
multivariate models that explore the linkages among interpersonal relation-
ships, situational factors, and social processes as well as blood alcohol concen-
tration levels, individual motivation, and orientation to the situation.
Because violence typically represents a low base-rate phenomenon,
promising directions for such research include an expanded community-wide

approach using a larger sample than Pernanen’s, drawn from a US city with a
heterogeneous population and high violent crime rate, to explore the wide
range of violent incidents. Alternative strategies for exploring incident escala-
tion and assault severity include secondary analyses of existing data, direct
observation of natural episodes of intoxicated behavior, and experimental
research using pairs of friends and intimates to explore the social, situational,
and individual factors that affect behavior while intoxicated.
Additional secondary analyses of the NCVS might focus on exploring the
relationship of the presence of alcohol to the nature of various threats, factors
that affect the seriousness of an injury, including the presence of a weapon
and of other persons, and the specific types of assaultive acts directed at
different types of victims. The size and complexity of the NCVS data set allow
for development of more complex models and multivariate analyses to exam-
ining the conditional and interactive factors that, together with alcohol inges-
tion, explain the escalation from threat to violence and victim injury in some
assaultive incidents. The addition of alcohol-related questions to the NCVS
would greatly enhance its value for exploring the escalation of violent events.
Other data sets, particularly the 1985 National Family Violence Resurvey,20
also may lend themselves to further analyses of the role of alcohol in incidents
that do and do not actually result in violence of various types.
Specific questions related to alcohol’s role in assaults include the condi-
tions under which social rules related to gender apply. Does alcohol affect the
types of violent acts men use against women and men? The application of
general norms that restrict violence against women in dating situations and
interactions with intimates and spouses? Why was there no alcohol effect on
men’s violent encounters with friends and acquaintances? Under what condi-
tions does inhibition of conflict increase or decrease? In summary, many
questions remain regarding the impact of alcohol consumption on the pro-
cesses of escalation and discrimination in violent interactions.
ACKNOWLEDGMENT. An earlier version of this chapter was presented at the
annual meeting of the American Sociological Association, Washington, DC,
August 19, 1995.
References
1. Collins JJ: Alcohol and interpersonal violence: Less than meets the eye, in NA Weiner and ME
Wolfgang (eds): Pathways to Criminal Violence. Newbury Park, CA, Sage Publications, 1989, pp
49-67.
2. Collins JJ, Messerschmidt P: Epidemiology of alcohol-related violence. Alcohol Health Res
World 17:93–101, 1993.
3. Murdoch DD, Pihl RO, Ross D: Alcohol and crimes of violence: Present issues. Int J Addict
25:1065–1081, 1990.
Alcoholism: Social Aspects of Alcoholism. New York, Plenum Press, 1976, pp 351–444.
JJ (ed): Drinking and Crime: Perspectives on the Relationships between Alcohol Consumption and
6. Pernanen K: Alcohol in Human Violence. Guilford, New York, 1991.
7. Roizen J: Issues in the epidemiology of alcohol and violence, in SE Martin (ed): Alcohol and
Interpersonal Violence: Fostering Multidisciplinary Perspectives. (NIAAA Research Monograph
No. 24). NIH Pub. No. 93-3496. Rockville, MD, DHHS, 1993, pp 3–36.
8. Roizen J: Estimating alcohol involvement in serious events, in alcohol consumption and
related problems. Alcohol and Health (Monograph No. 1) DHHS Publication No. 1 ADM
82-1190. Washington, DC, US Government Printing Office, 1982, pp 179–219.
9. Gerson LW, Preston D: Alcohol consumption and the incidence of violent crime. J Stud
Alcohol 40:307–312, 1979.
10. Beck AJ, Kline SA, Greenfeld LA: Survey of Youth in Custody, 2987. Washington, DC, Bureau of
Justice Statistics, 1988.
11. Beck, AJ: Profile of Jail Inmates, 2989. Washington, DC, Bureau Justice Statistics, 1991.
12. Cherpitel C: What emergency room studies reveal about alcohol involvement in violence-
related injuries. Alcohol Health Res World 17:162–166, 1993.
13. Roizen J: Alcohol and trauma, in Giesbrecht N, Gonzales R, Grant M, et al (eds): Drinking and
Casualties. London, Routledge, 1989, pp 21–66.
14. Shepherd J, Irish M, Scully C, Leslie I: Alcohol intoxication and severity of injury in victims of
assault. Br Med J 296:1299, 1988.
15. Honkanen R, Smith GS: Impact of acute alcohol intoxication on the severity of injury: A
cause-specific analysis of non-fatal trauma. Injury 21:353–357, 1990.
16. Cherpitel C: Alcohol and casualties: A comparison of emergency room and coroner data.
Paper presented at the Research Society on Alcoholism annual meeting, June 18-24, 1993,
San Antonio, TX.
17. Hamilton CJ, Collins JJ: The role of alcohol in wife beating and child abuse: A review of the
literature, in Collins JJ (ed): Drinking and Crime. New York, Guilford Press, 1981, pp 253–287.
18. Hotaling GT, Sugarman DB: An analysis of risk markers in husband to wife violence: The
current state of knowledge. Violence Victims 1:101–124, 1986.
19. Leonard KE, Jacob T: Alcohol, alcoholism, and family violence, in Van Hasselt VB, Morrison
FL, Bellack AS, Herson M (eds): Handbook of Family Violence. New York, Plenum Press, 1988,
pp 383–406.
20. Kantor GK, Straus MA: The “drunken bum” theory of wife beating. Social Problems 34:213–
231, 1987.
21. Roberts AR: Psychosocial characteristics of batterers: A study of 234 men charged with do-
mestic violence offenses. J Family Violence 2:31–93, 1987.
22. Fagan JA, Stewart DK, Hansen KV: Violent men or violent husbands? Background factors
and situational correlates, in Finkelhor D, Gelles RJ, Hotaling GT, Straus MA (eds): The Dark
Side of Families: Current Family Violence Research. Beverly Hills, CA, Sage Publications, 1983, pp
49–67.
23. Berk RA, Berk SF, Loseke DR, Rauma D: Mutual combat and other family violence myths, in
Finkelhor D, Gelles RJ, Hotaling GT, Straus MA (eds): The Dark Side of Families: Current Family
Violence Research. Beverly Hills, CA, Sage Publications, 1983, pp 197–212.
24. Kantor GK, Straus MA: Substance abuse as a precipitant of wife abuse victimizations. Am J
Drug Alcohol Abuse 15:173–189, 1989.
25. Straus MA: Measuring intrafamily conflict and violence: The conflict tactics scale. Durham,
University of New Hampshire, Family Research Laboratory, 1979.
26. Straus MA: The conflict tactics scale and its critics: An evaluation and new data on validity
and reliability, in Straus MA, Gelles RJ (eds): Physical Violence in American Families: Risk Factors
and Adaptations to Violence in 8,145 Families. New Brunswick, NJ, Transaction Books, 1990, pp
49–73.
27. Bureau of Justice Statistics: Criminal Victimization in the United States, 2992. Washington, DC,
Bureau of Justice Statistics, 1993.
28. Bushman BJ, Cooper HM: The effects of alcohol on human aggression: An integrative review.
Psychol Bull 107:341–354, 1990.
29. Gustafson R: What do experimental paradigms tell us about alcohol-related aggressive re-
sponding? J Stud Alcohol (Suppl 11):20–29, 1993.
30. Steele CM, Southwick L: Alcohol and social behavior I: The psychology of drunken excess. J
Personality Social Psychol 48: 18–34, 1985.
31. Buss A: The Psychology of Aggression. New York, Wiley, 1961.
32. Gustafson R Threat as a determinant of alcohol-related aggression. Psychol Rep 58:287–297,
1986.
33. Gustafson R: Alcohol, aggression and the validity of experimental paradigms in women.
Psychol Rep 59:51–56, 1986.
34. Lang AR, Goeckner DJ, Adesso VJ, Marlatt GA: Effects of alcohol on aggression in male social
drinkers. J Abnormal Psychol 84:508–518, 1975.
35. Pihl RO, Smith M, Farrell B: Alcohol and aggression in men: A comparison of brewed and
distilled beverages. J Stud Alcohol 45:278–282, 1984.
36. Taylor SP, Schmutte GT, Leonard KE: Physical aggression as a function of alcohol and frustra-
tion. Bull Psychonom Soc 9:217–218, 1977.
37. Taylor SP, Leonard KE: Alcohol and human physical aggression, in Green RG, Donnerstein
EI (eds): Aggression: Theoretical and Empirical Reviews, vol 2. San Diego, CA, Academic Press,
1983, pp 77–101.
38. Gantner A, Taylor SP: Human physical aggression as a function of alcohol and threat of
harm. Aggressive Behavior 18:29–36, 1992.
39. Schmutte GT, Taylor SP: Physical aggression as a function of alcohol and pain feedback. J
Social Psychol 10:235–244, 1980.
40. Zeichner A, Pihl RO: The effects of alcohol and instigator intent on human aggression. J Stud
Alcohol 41:265–276, 1980.
41. Jeavons CM, Taylor SP: The control of alcohol-related aggression: Redirecting the inebriate’s
attention to socially appropriate conduct. Aggressive Behavior 11:93–101, 1985.
42. Taylor SP, Gammon CB: Aggressive behavior of intoxicated subjects: The effect of third-party
intervention. J Stud Alcohol 37:917–930, 1976.
43. Taylor SP, Sears JD: The effects of alcohol and persuasive social pressure on human physical
aggression. Aggressive Behavior 14:237–243, 1988.
44. Steele CM, Josephs RA: Alcohol myopia: Its prized and dangerous effects. Am Psychol 45:921–
933, 1990.
45. Lang AR: Alcohol-related violence: Psychological perspectives, in SE Martin (ed): Alcohol and
Interpersonal Violence: Fostering Multidisciplinary Perspectives. (NIAAA Research Monograph
No. 24). Rockville, MD, US Department of Health and Human Services, 1993, pp 121–147.
46. Bachman R, Taylor BM: The measurement of family violence and rape by the redesigned
national crime victimization survey. Justice Q 11:499–512, 1994.
47. Straus MA, Gelles RJ (ed): Physical Violence in American Families: Risk Factors and Adaptations to
Violence in 8,145 Families. New Brunswick, NJ, Transaction Books, 1990.
48. Bachman R, Saltzman L: Violence against Women: Estimates from the Redesigned Survey (NCJ
#154848). Washington, DC, Bureau of Justice Statistics, 1995.
49. Leonard KE: Drinking patterns and intoxication in marital violence: Review, critique and
future directions for research, in S Martin (ed): Alcohol and Interpersonal Violence: Fostering
Multidisciplinary Perspectives (NIAAA Research Monograph No. 24). NIH Pub. No. 93-3496.
Rockville, MD, DHHS, 1993, pp 253–280.
3
Alcohol and Spouse Abuse

Ethnic Differences
Glenda Kaufman Kantor
Abstract. This chapter examines theoretical and empirical evidence on the interplay between
ethnicity, structural and cultural factors, and alcohol-related assaults against wives and considers
whether there is a differential vulnerability to such assaults among varying ethnic groups. Our
review demonstrated that structural factors emerged as dominant in their influence on alcohol-
related wife assaults in varying ethnic groups. The empirical evidence, though limited, showed
that the linkages between drinking and wife beating are not just a problem of poor ethnic
minorities. Heavy drinking per se is associated similarly in Hispanic-American and Anglo-Ameri-
can families. However, we also identified differences among Hispanic subgroups, as well as
cultural variations in drinking patterns that differentially affected wife assaults. Although data on
alcohol–wife assault relationships among African Americans are extremely limited, the available
evidence indicates little or no effect of drinking by African-American men on wife assaults, after
taking other socioeconomic variables into account. Empirical evidence did not support the salien-
cy of particular cultural beliefs favoring violence toward women as intrinsic to any one ethnic
group. The major cultural differences in alcohol-related cognitions are consistent with the greater
legitimation of alcohol-related misbehavior and the acceptance of "machismo" drinking by His-
panic Americans compared to Anglo-Americans.
1. Introduction
There is strong empirical grounding for both the existence of linkages be-
tween poverty and violent intrafamily crimes, such as wife abuse,1–3 and the
greater prevalence of alcohol-related problems among the impoverished.4
There is also a considerable body of research, often divorced from socio-
economic considerations, that establishes a correlation between excessive
Glenda Kaufman Kantor Family Research Laboratory, University of New Hampshire, Dur-
ham, New Hampshire 03824.
57
58 I Epidemiology
drinking and violent interpersonal crimes.5 In particular, alcohol has emerged

as a consistent predictor of wife assaults in a wide range of studies. A previ-
ous study,6 which did consider socioeconomic factors along with alcohol
consumption patterns, found evidence that binge-drinking patterns by blue-
collar men, in combination with the approval of violence, were associated
with higher wife abuse rates. The study findings supported the need for a
comprehensive theoretical framework integrating factors at the individual,
structural, and cultural levels. However, research on alcohol-related wife as-
saults is rarely situated in the context of ethnicity or culture. This chapter
examines evidence on the interplay between ethnicity, structural and cultural
factors, and alcohol-related assaults against wives and considers whether
there is a differential vulnerability to such assaults among varying ethnic
groups.
Violence is often an adaptation to stress produced by structural inequal-
ities. If adequate resources are not available, violence may be used by men
against wives to maintain their dominant status in the family.2 Several studies
have shown that the stress associated with unemployment or income dis-
parity between husband and wife leads to an increase in marital violence.2,7
Furthermore, the relationship between socioeconomic stress and family as-
saults may be mediated by substance abuse. Drug and/or alcohol use may
become a method for reducing the stress caused by changes in family interac-
tion.
The susceptibility of minority families to alcohol-related assaults is a
plausible assumption because poverty, unemployment, substandard hous-
ing, and segregation have characterized the minority experience in the United
States.1 For example, in 1991, the median income for Hispanic families was
only 65% of that earned by non-Hispanic families. Over a quarter of Hispanic
families lived below the poverty line in 1991 compared to 10% of non-Hispan-
ic families. African Americans also tend to be considerably more impov-
erished than white Americans.8
Despite the fact that Hispanic and African-American families constitute a
disproportionate percentage of impoverished families, and that lower socio-
economic status is associated with higher rates of family violence,2 evidence
for the cultural patterning of family violence is limited.9 There are few studies
of family violence using ethnicity rather than class as their central focus.
Research examining the linkages between alcohol and wife assaults within
differing ethnic groups is even more uncommon. The intent of this current
review is not to perpetuate stereotypes that dysfunctional family patterns and
family crimes are lower-class phenomena but to consider the available evi-
dence on structural and cultural areas that may have major implications for
those concerned with alcohol–aggression relationships and those concerned
with preventing violence against women.
Specifically, the questions considered in this chapter are: (1) Do Hispanic
and African-American husbands who drink heavily have a higher probability
3 Alcohol and Spouse Abuse 59
of wife beating than Anglo-American husbands who drink heavily? (2) Does
taking into account the effects of other variables such as poverty, accultura-
tion, and gender role attitudes alter the relationship between ethnicity, drink-
ing, and wife beating? (3) Are such linkages between drinking and wife beat-
ing found primarily among poor ethnic minorities?
2. Drinking and Violence in Ethnic Groups

Alcohol has been associated with a number of social ills, and over half of
all violent crimes are believed to be alcohol related.10 A substantial literature
establishes distinctive cultural drinking patterns and the greater prevalence of
alcohol problems, including alcohol-related violence, in certain subgroups of
the population.11–7 Problem drinking has been identified more often among
ethnic groups such as the Irish (both in America and in Ireland)18 and among
African-American and Hispanic-American men. While studies specifically de-
signed to examine the presence of alcohol-related aggression in different cul-
tures are uncommon, studies that do so find that aggressive behavior is not a
constant in heavy drinking cultures19 and that alcohol-related aggression is
more common in the Southern United States, among youth, and also in some
Hispanic subgroups.20 Additionally, Levinson’s9 cross-cultural analysis of
family violence in 90 small-scale and peasant societies found that wife beating
occurred in conjunction with the husband’s drunkenness in only 9% of soci-
eties. In approximately 6% of societies, the husband’s intoxication was incon-
sistently linked to wife beating.
2.1. Problem Drinking in African-American Men

A review of the literature on the extent of drinking among African-Amer-
ican men21 suggests that the differences between African Americans’ and
white Americans’ drinking are actually small or that African Americans drink
less than whites. At the same time, analyses of American drinking patterns
have identified greater alcohol problems for African Americans15,17,22,23 com-
pared to white Americans. The seeming inconsistencies of these findings
suggest that while abstention and moderate drinking values prevail among
African Americans, there may also be a segment of African-American society
where drinking is associated with negative consequences. According to
Herd’s17 historical analysis of African-American drinking patterns, the effects
of racial oppression and subsequent African-American migration to urban
areas, along with the targeting of African Americans first by the illicit alcohol
market and then by the legal alcohol industry, sowed the seeds of a heavy-
drinking culture.
Excessive drinking and other forms of substance abuse in largely African-
American and poor urban areas are often linked to violent crimes.24 The
mechanisms postulated for this association suggest that subcultural values

promoting heavy drinking, the accessibility of liquor stores and taverns in
poor African-American communities, and the social stresses of poverty lead-
ing to escapist drinking are major influences. Anger and frustration at racist
forces, unemployment, and feelings of powerlessness may lead to drinking,
family disorganization, and violence among some African Americans.24,25
2.2. Wife Abuse in African-American Families

Wife-abuse studies of shelter populations most often examine race as a
background variable and reach inconsistent conclusions about the relevance
of race to wife-abuse situations.26,27 Studies specifically designed to examine
wife abuse in African-American families28,29 conclude that race effects are not
significant when class factors are controlled. However, Lockhart28 also finds
that violence rates are higher for African-American middle-class women com-
pared to white middle-class women. Lockhart30 explains violent conflict reso-
lution strategies among the African-American middle class as a function of
lower-socioeconomic developmental experiences and the stresses resulting
from the tenuousness of newly acquired middle-class roles.
One reason for the inconsistent findings of the studies cited above may
be the methodological limitations in using clinical or convenience samples to
establish ethnic variations or prevalence rates. Prevalence rates of wife abuse
are best estimated in large national probability samples.31 Results from two
national surveys (the 1975 and 1985 National Family Violence Surveys)2,32,33
indicate that wife abuse rates are significantly greater in African-American
families compared to white American families even when occupational class
factors are controlled. However, few studies using large probability samples
have examined the possibility that excessive drinking or variables other than
socioeconomic factors contribute to the higher rates of wife assaults by Afri-
can-American men.
2.3. Moderators of Alcohol-Related Wife Assaults

in African-American Families
Failure to specify the complex array of causal and moderating factors
implicated in wife assaults by varying ethnic groups is undoubtedly related to
the gaps in research. Indeed, Asbury’s34 discussion of African-American
women in violent relationships takes wife-abuse researchers to task for failing
to mention the race of abused women. Her review of the wife-abuse literature
adopts an “afrocentric perspective” (i.e., a worldview that is framed by Afri-
can-American historical traditions and experiences of slavery, oppression,
poverty, and discrimination) in identifying themes relevant to black women’s
victimizations by spouses. Her themes, derived largely from Roy’s35 list of
triggers to wife abuse, emphasize economic difficulties, alcohol use, tradition-
al sex-role socialization, and African-American women’s family allegiances as
3 • Alcohol and Spouse Abuse 61
contributors to assaults on black American women. Using a sociological per-

spective, Asbury34 also suggests that greater anomie* among black American
men unable to achieve societal goals may result in violence toward women
and others, Although this is a promising theoretical framework for examining
wife assaults among African-American families, no empirical data are pro-
vided.
Even research on anomie and violence is limited. In fact, researchers
have failed to identify significant linkages between anomie and violence, as
well as for anomie and excessive drinking.37 However, the lack of statistically
significant associations may be more a function of the level of abstractness
present in the concept of “anomie” and difficulties in operationalization rath-
er than the absence of any relationship between alcohol, anomie, and vio-
lence. Such data also do not address the effects of anomie on the occurrence
of violence in intimate partnerships.
Most experts on wife assaults concur that these are complex phenomena
rooted in both social structural and family processes. There is, for example,
speculation that traditional sex roles contribute to assaults on African-Ameri-
can women. Historically, African-American families have been stereotyped as
matriarchal and wife dominated. However, other data on marital satisfaction
and family role structure suggest that wife-dominant African-American
couples are as satisfied with their marriages as their egalitarian counter-
parts.38 Additionally, research comparing the gender roles of African-Ameri-
can and white American men has yielded inconsistent results. Generaliza-
tions from the body of research on African-American family structure and
family violence regarding predictors or moderators of spousal assaults are not
easily made because so few studies have been done. There is clearly a need to
empirically examine the complex interrelationships between alcohol, socio-
economic status, family structure, marital quality, and marital violence.
2.4. Problem Drinking in Hispanic-American Men

There is also evidence that high rates of alcohol consumption and prob-
lem drinking exist among Hispanic men to a greater extent than for other
ethnic groups.13,15 However, with the exception of Caetano’s39 survey, which
did examine a representative sample of Americans, most studies of Hispanic-
American drinking patterns have had limited generalizability because of the
low numbers of Hispanics sampled or sampling limited to specific subgroups
or geographical areas.40–42
There are several reasons why we would expect to find greater alcohol
problems in Hispanic Americans compared to other groups such as their lack
of economic and educational resources, lesser access to external social sup-
* According to Merton’s36 theory of deviance, anomie leads to deviance when there are dis-
parities between culturally prescribed goals and socially structured avenues to achieve those
goals.
ports such as human services,43,44 and potentially greater social isolation.

Using the suppositions of a social stress perspective,45 the numerous daily life
stresses associated with poverty or isolation could increase the likelihood of
using alcohol as a coping mechanism. Another explanation is that cultural
influences increase the risk of problem drinking.
2.5. “Machismo” Drinking and Wife Assaults

An ethic supporting heavy drinking by Hispanic men, i.e., “machismo
drinking,” as an indicator of their masculinity, is often assumed to be at the
heart of Hispanic drinking problems, but it is also rejected as an inaccurate
and pejorative stereotype.40,46–48 In fact, Kaufman Kantor, Aldarondo, and
Jasinski49 argue that the machismo ethic is actually one where “real” men
consume substantial quantities of alcohol without demonstrating ill effects and
without interference with their work and family roles. In other words, there
are both positive and negative aspects of the machismo concept, such that the
ethic of machismo drinking supports alcohol tolerance but purportedly with-
out negative consequences to others.
Few studies provide information on how wife assaults and intoxication
might be associated in Hispanic families. Despite our assertions about the
complexity of the machismo concept, the broader stereotype of alcohol ma-
chismo, e.g., the conjunction of heavy drinking and assertion of manliness
through physical force, does suggest a possible mechanism. Some support for
the stereotype is provided by ethnographic accounts of native Hispanic cul-
tures. One such account50 describes a practice among the Tzeltal of Mexico
where wife beating following the husband’s drinking has been formalized to
such an extent that it is incorporated into the marriage ceremony. Levinson50
reports that “the beatings are routinely expected and accepted by Tzeltal
wives who excuse their husbands as crazy” (p. 50). Another ethnographic
study in Puerto Rico51 identifies a practice known as “Social Fridays,” where
men stay away from home all weekend and drink to intoxication. However,
no research has examined this pattern in the mainland. If this were so, it
suggests a potential source of marital conflict and, at worst, a greater risk of
marital assaults, since binge drinking patterns have been linked to wife
abuse.6,52 However, drinking patterns and attitudes approving of drunken-
ness or violence may also differ depending on the degree to which accultura-
tion has occurred53–55 or social status has changed.
2.6. Ethnic Comparisons of Wife Abuse

Until recently, few studies examined the occurrence of marital violence in
Hispanic-American families. Studies examining differences by race in shelter
populations of battered women find no differences between white, African-
American, and Hispanic women in the frequency and severity of abuse, but
Hispanic women appear to have endured abuse longer before entering shel-
ters.56–58 Hispanic-American women are said to have a higher “tolerance”

compared to Anglo-American women for minor physical abuse such as slap-
ping and for verbal abuse.58 Hispanic women are also more likely to have
larger families and might remain in relationships longer because of their
children.56,58 Although these findings are consistent with some discussions of
Hispanic family structure,59 there are also methodological limitations in gen-
eralizing conclusions from studies using clinical samples to the Hispanic pop-
ulation at large.
2.7. Wife Assault Patterns among Hispanic Americans

Straus and Smith’s3 analysis of national data concludes that Hispanic-
American husbands assault their wives at a rate more than double that of
white American husbands. Although family assaults seem inconsistent with
descriptions of Hispanic families as cohesive and close-knit,60 characteriza-
tions of Hispanic families as male dominant,59,61 along with high levels of
poverty and unemployment, are consonant with typical predictors of wife
abuse. Authoritarian role structures assigning dominance to the husband and
dictating obedience by the wife are said to persist at least as an ideology
of Mexican-American, Puerto Rican, and Cuban cultures, if not in prac-
tice.59,62–65 Despite theoretical evidence to support contentions of wife abuse
among Hispanics, there is, in fact, little empirical evidence on wife beating in
Hispanic families. However, high rates of intrafamily homicides in Hispanic
families suggest the possibility of severe spousal violence,66 and this is sup-
ported by reports based on the 1985 National Family Violence Survey3,67 and
the 1992 National Alcohol and Family Violence Survey.68
2.8. Drinking and Wife Assaults in Multiethnic Groups

A previous analysis of ethnicity and drinking,67 using the 1985 National
Family Violence Survey, examined the incidence of drinking problems and
wife abuse by Hispanic, African-American, and white American husbands.
The findings reveled greater prevalence of wife assaults and binge-drinking
problems among ethnic minorities (i.e., African-American and Hispanic-
American men). Hispanic-American men were approximately three times as
likely as white American men and almost four times as likely as African-
American men to engage in high-volume binge drinking. This study also
found that Hispanic-American women with binge-drinking husbands were
more than ten times as likely to be assaulted than were Hispanic-American
wives of low-moderate drinkers. The findings also revealed that African-
American women were at high risk of assault if their husbands were excessive
daily drinkers. Additionally, a multivariate analysis was conducted to exam-
ine the joint effects of poverty, drinking, and ethnicity on wife abuse proba-
bilities. The results of the multivariate analysis comparing white and African-
American husbands showed that although overall violence rates by minority
husbands were generally higher than those of white Americans, the effects of
race among African-American respondents dominated those of alcohol in
predicting wife assaults even when socioeconomic status was controlled.
In contrast to the findings for African-American husbands, the highest
rates of violence by Hispanic-American husbands occurred for those at pover-
ty level with high-volume binge-drinking patterns. The multivariate analysis
comparing Hispanic and white American husbands also showed that drink-
ing was the only significant predictor of wife abuse when socioeconomic
status and ethnicity were controlled. However, there are also important limits
to generalizing the findings of this study regarding Hispanic Americans be-
cause subgroup differences between Hispanics and acculturation levels of
Hispanic Americans were not measured in this earlier (1985) national study.
2.9. Moderators of Alcohol-Related Wife Assaults in Hispanic-American

Families:
Acculturation
Acculturation needs to be considered as a potential moderator of differ-
ential risks for wife assaults between Hispanic-American subgroups. Accul-
turation is a process by which an immigrant group takes on the norms and
behavior patterns of the host society and modifies customs, habits, language,
lifestyle, and values over time.69,70
2.9.1. Acculturation and Mental Health. Some have suggested that a num-
ber of psychological disorders (e.g., alcoholism, depression, psychosomatic
symptoms) are more prevalent among highly acculturated Mexican Ameri-
cans compared to native born (Mexican born) Mexican Americans.71–73 This is
believed to be caused by prolonged stresses associated with adaptation, such
as alienation from Mexican culture, or the experience of discrimination or
deprivation relative to Anglo-American society. However, the evidence for
detrimental effects of acculturation is not unequivocal. For example, Griffith
and Villavicencio74 found that high acculturation is correlated with social
support, extended kin network, and lower incidence of psychological distress
in Mexican American families. They also attribute this relationship to the
higher socioeconomic status of more acculturated Mexican Americans.
2.9.2. Acculturation and Alcohol. Studies that have examined acculturation-
al influences on alcohol consumption yield mixed results on the direction of
the effects, though the majority suggest that drinking increases with accultur-
ation.72,75,76 For example, Caetano and Medina Mora’s75 study comparing
men of Mexican descent in their native country with Mexican Americans in
the United States suggests that Mexican-American men drink more fre-
quently in the United States. There is also evidence that escapist drinking
patterns characterize lesser acculturated Mexican Americans.77 Gordon’s55
study of Dominican men demonstrates a shift from macho drinking, i.e.,
heavy drinking and drunken fighting, to a more conservative lifestyle in the
3 Alcohol and Spouse Abuse 65
United States. Yet another study by Gordon78 finds that Guatemalan men
maintain their practices of macho drinking. A more careful examination of
these studies suggests that factors other than acculturation, such as altered
family structure, work arrangements, socioeconomic status, and drinking op-
portunities, are responsible for either the maintenance or the severance of
drinking behavior patterns. The divergent views on acculturation effects are
most likely due to the lack of uniform measures that tap different dimensions
of the construct or the failure to control for the confounding effects of socio-
economic status.
2.9.3. Acculturation and Wife Abuse. One would expect that factors likely to
modify the linkages between alcohol and wife assaults such as family roles,
norms approving violence, and attitudes about drinking and drinking behav-
iors may be influenced by the level of individual acculturation. However, few
studies of Hispanic family violence have specifically measured acculturation.
Torres,79 for example, describes the degree of acculturation for the subject
population of battered women in shelters but does not use acculturation as a
variable in her empirical analysis. Additionally, a major limitation of previous
studies examining the linkages between Hispanic alcohol abuse and wife
abuse67 or studies examining the incidence of intrafamily violence among
Hispanic and white families3 is their inability to adequately measure the con-
cept of Hispanic ethnicity. The validity of categorizing diverse members of
society under one classification is questionable. This approach fails to detect
the actual subgroup members at risk. One could plausibly assume consider-
able heterogeneity within and between Hispanic subgroups regarding family
roles, attitudes toward alcohol use, and beliefs about the legitimacy of vio-
lence. It is not at all clear that Hispanicity is a coherent construct or even that
certain subgroups, (e.g., Puerto Ricans, Mexicans, Dominicans) share a
unique cultural experience or consistent set of beliefs about issues that may
affect the risks of family violence.
Studies that have empirically examined the role of acculturation in wife
assaults68,80 indicate that wife abuse rates are higher among US-born Hispan-
ics. Sorenson and Telles’s80 analysis of Epidemiological Catchment Area data
on the prevalence of wife abuse in Los Angeles indicates that Mexicans born
in the United States have the highest life-time prevalence of wife assaults
compared to Anglos and Mexican-born Hispanics. Thus, using country of
birth as an alternative measure for acculturation indicates that acculturation
may actually increase the likelihood of wife abuse.
The 1992 National Alcohol and Family Violence Survey (NAFVS)68 pro-
vides a primary source of data on the prevalence and incidence of Hispanic
and Anglo-American spousal violence. The research design of this study
included face-to-face, bilingual (Spanish–English) interviews with a national
probability sample of 1970 persons, including an oversample of approx-
imately 800 Hispanic persons. The study included specific measures of accul-
turation and perceived ethnicity, among others. The population examined
included sufficient numbers of Hispanics representing the three major sub-

groups in the United States (Mexicans, Puerto Ricans, and Cubans) to allow
for subgroup analyses.
2.10. The 1992 National Alcohol and Family Violence Survey

Using the NAFVS, we examined the relationship between sociocultural
status and the incidence of marital violence in Hispanic-American and Anglo-
American families.68 Findings from this study revealed that the highest rates
of wife assaults were found among more acculturated Mexican Americans
and Puerto Ricans. We also replicated the Sorenson and Telles findings that
country of birth (US born) is significantly associated with wife assaults. At the
same time, we did not find significant effects of our acculturation measure
(reflecting language preference and Spanish–English utilization) on wife as-
saults when other cultural and socioeconomic indicators were held constant.
This suggests that acculturation is confounded with these other measures.
Another plausible explanation is that the acculturation measure used in this
study, though widely validated, was conceptually limited in its linguistic
focus.
The major findings of this study did support the importance of a socio-
cultural approach to understanding wife abuse. Similar to the findings of the
1985 National Family Violence Survey,3,67 this study found that Hispanic
Americans, taken as a single ethnic group, do not differ significantly from
Anglo Americans in their odds of wife assaults when cultural norms regard-
ing violence approval, age, and economic stressors are held constant. Second,
our findings provided evidence for not regarding Hispanic Americans as a
homogeneous group. Rather, considerable heterogeneity was apparent be-
tween ethnic subgroups relative to their extent of acculturation, country of
birth, impoverishment, and their acceptance of wife abuse as normative. We
found that cultural norms sanctioning wife assaults, while significantly asso-
ciated with actual husband-to-wife violence, are not uniquely Hispanic cultur-
al values. Our results show that the presence of these norms within any
ethnic subculture, regardless of socioeconomic status, is a risk factor for wife
abuse. A third important finding of this study, amplifying the discovery of
Sorenson and Telles,80 was that being born in the United States increases the
risk of wife assaults by Mexican- and Puerto Rican-American husbands.
One possible and important inference from our study of sociocultural
status and our finding of a significant association between country of birth
and wife assaults is that family strengths derived from intact cultural values
may provide buffers against stresses that might otherwise lead to conflict and
violence. This suggests the importance of building and preserving family
strengths and kin networks in the prevention of intrafamily violence among
Hispanic families. At the same time, nonegalitarian family power structures,
which exist in some Hispanic-American families, and problem drinking,
which is also higher in Hispanic-American families relative to other groups,

can each affect levels of verbal and physical conflict between husbands.
One important limitation to both studies emphasizing acculturation ef-
fects on wife assault, discussed above, is that neither of these studies provide
information on how acculturation or being born in the United States might
have such deleterious effects. There is a need to empirically examine the
relationship of other dimensions of acculturation to wife assaults such as
feelings of cultural marginality77 and perceptions of social distance and dis-
crimination from the dominant Anglo majority.81 This is particularly impor-
tant because studies arguing for a culture of violence interpretation, i.e., more
normative acceptance of violence by certain subgroups, rarely specifically test
alternative explanations such as acculturation stress or discriminatory effects.
Additionally, neither of these studies took alcohol's role in wife beating into
account.
2.10.1. Alcohol-Related Wife Assaults in Minority Families. A second anal-

ysis of the 1992 National Alcohol Family Violence Survey data49 evaluated the
association between alcohol consumption and wife assaults among the three
largest Hispanic subgroups in the United States and among Anglo-American
families.
2.10.2. Violence-Related Behaviors. Because variations in violence rates may

also be influenced by ethnicity, we first considered the overall incidence rates
by ethnic groupings. The results of this second analysis support the findings
discussed above. As shown in Fig. 1, the highest rates of assaults on wives
occur in Puerto Rican-American families. Rates of wife beating in Puerto
Rican-American families are more than double those reported in Anglo fami-
lies. Assault rates in Mexican-American families are almost equally as high. It
can also be seen that rates of wife assaults by less acculturated Mexican
husbands living in the United States are virtually the same as those of Anglo-
American husbands. However, as shown in Fig. 1, husband-to-wife assaults
are rarely reported by Cuban-American families.
2.10.3. Average Drinking Patterns and Marital Assaults. Figure 2 examines

the association between husband's average alcohol consumption patterns
dichotomized into nonheavy (abstinent through low-volume binge) and
heavy (high daily and high-volume binge) drinking patterns and wife as-
saults. The results based on an analysis of variance show significant main
effects of ethnicity and drinking. Consistent with the results of our previous
research,6,82 it is evident that heavy-drinking husbands engage in considera-
bly more wife abuse. The highest proportion of alcohol-associated wife as-
saults are again found in Puerto Rican families. Heavy-drinking Puerto Rican
husbands are five times more likely to hit their wives than their nondrinking
counterparts. Virtually the same proportions of wife assaults are reported by
Figure 1. Husband-to-wife violence by ethnic group.
Anglo- and Mexican-American heavy-drinking husbands. Although Cuban

husbands appear to be rarely violent, the proportion of husbands that hit
their wives more than triples when heavy drinking is taken into account.
2.10.4. Alcohol as a Temporal Antecedent of Wife Assault. Unlike most of

the research discussed thus far, we also examined the extent to which intox-
ication was a component of the aggressive episode against a wife. Figure 3
shows the proportion of family members who report actual drinking or intox-
ication of self or partner at the time of a spousal assault. The findings are
generally consistent with the relationships shown in the previous figures; that
is, the likelihood of intoxication or drinking by one or both partners at the
time of a wife assault is greatest among mainland Puerto Rican family mem-
bers. This was the case in over half of the latter families where a wife assault
had occurred. However, they also illustrate that drinking by wives contrib-
utes to alcohol-related assaults only in Puerto Rican and Anglo-American
families. It should also be noted that the analyses presented in Fig. 3 were
conducted only on families where wife assaults were reported. The small
sample sizes, particularly for Puerto Rican and Cuban men, indicate that
caution is needed in generalizing from these results. Nevertheless, with the

exception of Cuban families (where both heavy drinking and violence are
rare) and some variations in the role of drinking by wives, the results show
similar patterns of alcohol-related marital assaults regardless of ethnicity.
2.10.5. Alcohol Cognitions. In additional analyses, the role of alcohol expect-

ancies and norms legitimizing alcohol-related violence among Hispanic-
American and Anglo-American men were explored. The findings showed
that the main way in which Hispanic subgroups differed from Anglo sub-
groups is in the greater legitimation of alcohol-related behavior and in the
acceptance of machismo drinking (“A real man can hold his liquor”) by His-
panic Americans. Cubans differ significantly from all other groups with low
expectancies about alcohol’s ability to transform behavior. This is also consis-
tent with Cuban’s overall modest drinking habits. However, we concluded
that the overall homogeneity of Hispanics, including Cubans, regarding
acceptance of alcohol-related misbehavior is rather striking. We speculated
that this may reflect a more forgiving and empathic standard among His-
Figure 2. Percent husband-to-wife violence by ethnic group and husband’s average

drinking.
Figure 3. Drinking at the time of husband-to-wife violence by ethnic group.
panics compared with the sterner, more reproachful morality of Anglo-Amer-

ican men.
For Anglo-American men, further analyses using logistic regression indi-
cated that aggressive expectancies in combination with youth significantly
increased the likelihood of aggression toward wives, net of the effects of
alcohol consumption patterns and socioeconomic factors. For Hispanic men,
both disinhibition and aggressive expectancies affected the likelihood of wife
assaults. Additionally, youth and unemployment added to the probability
that Hispanic-American men would engage in assaults on wives. It should
also be noted that both aggressive alcohol expectancies and unemployment
each more than doubled the odds of wife assault among Hispanic-American
husbands.
Anglo-American men in this study were considerably more economically
advantaged than the Hispanic men and families studied. Yet their rate of high
daily drinking was almost twice that of Cuban and less acculturated Mexican
men. Despite the latter pattern for Anglo men, alcohol-related problems were
less frequent. A significant linear association between alcohol consumption
and husband-to-wife violence is evident only for Anglo husbands. The lin-
earity of the alcohol–wife assault relationship for Anglos but not for Hispan-
ics may be because binge-drinking patterns (at both lower- and higher-vol-
ume alcohol consumption) are more common to Hispanics, and the latter
alcohol consumption pattern is linked to wife assaults among some Hispanic-
American subgroups.
2.11. Evidence from Community and Clinical Samples

of Multiethnic Groups
Unlike the analyses presented on data from the two national probability
surveys, most research on alcohol and wife assaults in ethnic populations,
albeit limited, has been conducted in particular geographical areas or among
clinical samples of battered women. A recent study by Neff, Holamon, and
Schluter83 investigated wife assaults among Anglos, African Americans, and
Mexican Americans residing in San Antonio, Texas. The study respondents
differed from the currently partnered sample described in the 1992 national
study in that respondents included previously partnered individuals. The
study by Neff and associates also used a more limited measure of violence
(one item) in contrast to the multiple-item measure used in the national sur-
veys. This study considered the moderating effects of financial stress rather
than acculturation or specific socioeconomic indicators such as unemploy-
ment. Additionally, the moderating effects of sex roles, partner drinking, and
social desirability effects were taken into account by Neff and associates.
Neff and associates83 concluded that females, the formerly married, and
African-American females were most likely to report being beaten by and
beating a partner. High rates of partner assaults persisted among married
African-American females, even after taking into account the potential mod-
erating variables discussed above and the drinking quantity of the partner. In
fact, even while violence rates were highest among African Americans, drink-
ing was least important.
No significant differences in violence were found in comparisons of Mex-
ican-American and Anglo-American respondents. Sex role traditionalism, the
principal sociocultural indicator, did not emerge as a significant predictor of
spousal assaults. However, the authors note the limitation of the absence of
data on the partner’s sex role orientation. The authors also suggest that a
curvilinear relationship exists between alcohol and violence such that the
highest-quantity drinkers are less violent than lighter drinkers. Furthermore,
among Mexican-American females, quantity of alcohol consumption by the
wife or husband was not directly related to spousal assaults. The authors find
little support for minority stress explanations of wife assault. They conclude
that future research must pay more attention to relationship dynamics as a
means of explaining alcohol-related wife assaults in differing ethnic groups.
A second study examining cultural explanations of wife assault among
abused Latinas also considered the contributions of alcohol and sex role orien-
tation, along with acculturation and family communication patterns. Perilla,

Bakeman, and Norris84 interviewed a sample of 60 women in a Southeastern
city, including 30 battered women seeking help for abuse and 30 other wom-
en who sought other services from a Catholic hospital. Multivariate analyses
found no significant effects of acculturation or sex role orientation. The au-
thors did find that frequent drunkenness of the partner was associated with
higher levels of abuse. However, they also concluded that the relationship
was largely mediated by their measure of mutuality (items measuring ”empa-
thy, engagement, authenticity, empowerment, zest, and diversity”). Indirect
evidence for the importance of culture related to male machismo was inferred
by the finding that women’s contributions to family income were positively
associated with abuse. The authors argue that this is due to the husband’s
perceived loss of competency as a provider.
3. Summary and Conclusions

3.1. Examining the Theoretical Framework
The evidence examined in this chapter provided some support for utiliz-
ing an integrated theoretical framework, including individual, structural, and
cultural constructs, in examining alcohol-related wife assaults. In particular,
structural factors emerged as dominant in their influence on wife assaults.
However, support for all of the constructs identified here cannot be unqual-
ified because the available empirical evidence on ethnic differences in alcohol-
related consequences largely omits personality correlates and family history
factors. The only intraindividual characteristic examined in the studies re-
viewed here was the husband’s alcohol use. Although alcohol use is central to
our thesis of wife assaults because of its psychopharmacological effects and
individual expectancies about the effects of alcohol, other factors such as
aggressive personality style,85 communication patterns, or family history of
violence86 may influence the likelihood that any one individual will engage in
wife beating. Additionally, it is important to emphasize the need for caution
in drawing firm conclusions about all of the questions considered here. This is
necessary because research regarding the structural-cultural patterning of
alcohol-related wife assaults in different ethnic groups is in its infancy.
3.2. Does Alcohol Influence Wife Beating Equivalently

across Ethnic Groups?
The empirical evidence shows that the linkages between drinking and
wife beating are not just a problem of poor, ethnic minorities. Research based
on two large national probability samples of American families with Hispanic
oversamples support the contention that heavy drinking per se is associated
similarly in Hispanic-American and Anglo- American families. However, we
also identified differences among Hispanic subgroups. Additionally, the

study by Perilla and associates84 found that husbands’ frequent drunkenness
increases assaults against Latina wives. Evidence from the two national stud-
ies also demonstrates that cultural variations in drinking patterns differen-
tially affect wife assaults. For example, among Anglo men, alcohol-related
problems were less frequent, but such problems were linearly and signifi-
cantly associated with husband-to-wife violence. The linearity of the alcohol-
wife assault relationships for Anglos and not Hispanics may occur because
binge-drinking patterns (at both lower- and higher-volume alcohol consump-
tion) are more common to Hispanics, and the latter alcohol consumption
pattern is linked to wife assaults among some Hispanic-American subgroups.
This phenomenon may be a function of where drinking takes place. That is, if
men drink away from home, this may in turn create conflict between spouses
that is fueled by the husbands intoxication and then escalates into violence.
There also may be physiological differences in tolerance or metabolism be-
tween high daily drinkers and episodic drinkers that increase the likelihood of
aggressive responding under the influence. Although Neff and associates83
failed to find significant effects of alcohol quantity on wife assaults in Mexi-
can-American families, this may be related to their use of alcohol quantity per
se rather than volume-variability-based patterns such as binge drinking.
Experimental studies87 do find greater negativity of alcoholic couples
(more disagreements by wives) on “drink night” compared with normal
couples. But there may also be differences in communication patterns de-
pending on the drinking pattern of the husband. When binge and steady
drinkers were compared, binge drinkers were found to have more negative
communications when drinking, while steady drinkers had more positive
communications when drinking during the experimental manipulation. Such
effects may occur because partners can better cope with the predictability of
steady-drinking behaviors, or possibly because binge drinkers may them-
selves have other characteristics such as greater aggressiveness that are con-
founded with drinking.
3.3. The Importance of Race and Social Structure

We found very limited empirical data on the relationship between alcohol
and wife assaults in African-American families. However, the two studies
that examined this relationship67.83 found little or no net effect of drinking by
African-American men after taking other variables into account. Rather, the
effects of race dominated those of other variables examined. We believe that
the effects of race represent variables unmeasured by these studies such as
marginalization and the sequelae of years of segregation and discrimination.
Additionally, studies demonstrating significant alcohol effects also show that
these effects are moderated by relationship, structural, and cultural factors.
Structural factors were strongly predictive of wife assaults in minority
and Anglo families, net of drinking. The greater importance of structural
factors among minority families occurs because unemployment and poverty

are disproportionately experienced by minority individuals. Additionally, mi-
nority individuals often reside in urban communities characterized by high
rates of drug dependence, poverty, and substandard housing and plagued by
high levels of predatory and violent crimes. The social pathology of such
environments and the inequalities relative to the dominant culture can influ-
ence stress and, subsequently, assaultiveness in families. However, to a large
extent many of these indicators of urban poverty went unmeasured in the
studies examined. Although we have focused on net effects of structural
variables in the models examined, variables do not exist in isolation in reality,
and empirically the additive effects of alcohol, unemployment, and poverty
sum up to greatly increased risks of wife assaults.
3.4. The Importance of Gender Roles

None of the studies examining gender role effects found significant asso-
ciations with wife assaults. This may reflect the inability to measure central
aspects of gender role conflicts between spouses or male dominance, or it
may be that shifting gender roles with women’s necessary entry into the work
force may render many gender-role measures irrelevant in today’s society.
However, loss of the male provider role, traditionally defined as central to
concepts of masculinity, does appear to affect the likelihood of wife abuse.
Some argue that this is due to the husband’s perceived loss of competency as
a provider or loss of self-esteem when the wife must become the primary
breadwinner. Additionally, some assert that violence by African-American
men against women may occur because some aspect of the woman’s behavior
threatens his masculinity.87
3.5. The Importance of Culture and Acculturation

There was little evidence supporting a relationship between acculturation
and wife assaults when linguistic measures of acculturation were used. On
the other hand, acculturation as measured by country of birth did show a
positive and significant association with wife assaults, such that being US
(mainland) born increased the risk of wife assault in Puerto Rican- and Mexi-
can-American families. The reasons for the latter effects are unclear. How-
ever, we68 have previously hypothesized several explanations, such as aban-
donment of cultural norms of familialism and religiosity, or greater candor
about disclosing violent behavior, or even the adoption of violent norms from
exposure to the cultural or structural patterns of the United States.
There was no evidence that the saliency of particular cultural beliefs
favoring violence toward women was intrinsic to any one ethnic group. Addi-
tionally, aggressive expectancies about alcohol characterized both violent An-
glo and Hispanic men, suggesting a similar cultural framework for violence
against women across ethnic groups. The major evidence for cultural differ-
ences was in the greater legitimation of alcohol-related behavior and in the

acceptance of machismo drinking (“A real man can hold his liquor”) and
alcohol-related misbehavior by Hispanic Americans.
3.6. Implications
Our review suggests that more empirical testing is needed to assess the
possibilities of cultural patterning and the dynamics of alcohol-related vio-
lence among different social groups. Longitudinal examination of the pro-
cesses and empirical data based on both partners in an intimate relationship
would improve our understanding of causal relationships.
Minority women do experience a disproportionate and unacceptable lev-
el of violence at the hands of their partners. However, they may be less
willing to use formal social services and have fewer resources to escape from
abusive situations. Similarly, minority men may be uncomfortable seeking
help from formal treatment programs or may find treatment providers hostile
or insensitive to their needs.88 Prevention efforts and improved outreach to
heavier-drinking subgroups are needed to counter normative beliefs about
alcohol. Additional outreach and greater resources should be made available
to women and families at risk. Additionally, both alcohol treatment and bat-
terer treatment programs need to develop sensitivity and expertise in work-
ing with minority clients in order to improve treatment outcomes.
ACKNOWLEDGMENTS. Our research reported here was supported by Research

Grant No. RO1AA09070 from the National Institute on Alcohol Abuse and
Alcoholism. We are grateful to Mary Austin, Kelly Foster, and Jana Jasinski
for their assistance in manuscript preparation.
References
1. Hampton RL: Family violence and homicide in the black community: Are they linked? in
Hampton RL (ed): Violence in the Black Family: Correlates and Consequences. Lexington, MA,
Lexington Books, 1987, pp 135 –156.
2. Straus MA, Gelles RJ, Steinmetz SK: Behind Closed Doors: Violence in the American Family.
Newbury Park CA, Sage, 1980.
3. Straus MA, Smith C: Violence in Hispanic families in the United States: Incidence rates and
structural interpretations, in Straus MA, Gelles RJ (eds): Physical Violence in American Families:
Risk Factors and Adaptations to Violence in 8,145 Families. New Brunswick, NJ, Transaction
Publishers, 1980, pp 341–367.
4. Cahalan D: Problem Drinkers. San Francisco, CA, Jossey-Bass, 1970.
5. Collins JJ: Drinking and violence: An individual offender focus, in: Alcohol and Interpersonal
Violence: Fostering Multidisciplinary Perspectives,NIH Publication No. 93-3496. Rockville MD,
1993, pp 221–235.
6. Kaufman Kantor G, Straus MA: The “drunken bum” theory of wife beating. Social Problems
34(3):213–230, 1987.
7. Smith C: Status discrepancies and husband-to-wife violence. Paper presented at the Annual
Meetings of the Eastern Sociological Society, 1988.
8. US Bureau of the Census, Current Population Reports: Hispanic Americans Today. (P23-183)
Washington, DC, US Government Printing Office, 1993.
9. Levinson D: Family violence in cross-cultural perspective, in Van Hasselt VB, Morrison RL,
Bellack AS, Hersen M (eds): Handbook of Family Violence. New York, Plenum Press, 1988, pp
435–455.
10. Gottheil E, Druley KA, Skoloda TE, Waxman HM: Aggression and addiction: Summary and
overview, in Gottheil E, Druley KA, Skoloda TE, Waxman HM (eds): Alcohol, Drug Abuse and
Aggression. Springfield, IL, Charles C. Thomas, 1983, pp 333–356.
11. Bennett LA, Ames GM (eds): The American Experience with Alcohol: Contrasting Cultural Perspec-
tives. New York, Plenum Press, 1985.
12. Block CR: Lethal violence in the Chicago Latino community, 1965 to 1981. Paper presented at
Research Conference on Violence and Homicide in Hispanic Communities, Los Angeles, CA,
1987.
13. Caetano R: Drinking Patterns and Alcohol Problems in a National Sample of US Hispanics. Prepared
for presentation at the National Institute on Alcohol Abuse and Alcoholism conference on
Epidemiology of Alcohol Use and Abuse among US Minorities, Bethesda MD, 1985.
14. Cahalan D, Cisin IH, Crossley HM: American Drinking Practices: A National Study of Drinking
Behavior and Attitudes. New Brunswick, NJ, Rutgers Center for Alcohol Studies, 1969.
15. Cahalan D, Room R: Problem Drinking among American Men. New Brunswick, NJ, Rutgers
Center for Alcohol Studies, 1974.
16. Clark W, Midanik L: Alcohol use and alcohol problems among US adults: Results of a 1979
national survey, in: Alcohol and Health Monographs No. 1: Alcohol Consumption and Related
Problems, DHHS Publication No. ADM 82-1190. Washington, DC, US Department of Health
and Human Services, 1982.
17. Herd D: Ambiguity in black drinking norms: An ethnohistorical interpretation, in Bennett
LA, Ames GM (eds): The American Experience with Alcohol: Contrasting Cultural Perspectives.
New York, Plenum Press, 1985, pp. 149–169.
18. Greeley AM, McCready WC, Theisen G: Ethnic Drinking Subcultures. New York, JF Bergin
Publishers, 1980.
19. MacAndrew C, Edgerton RB: Drunken Comportment: A Sociological Explanation. Chicago, Al-
din, 1969.
20. Levinson D: Alcohol use and aggression in American subcultures, in Room R, Collins G
(eds): ADAMHA 1981 Alcohol and Disinhibition: Nature and Meaning of the Link, Monograph-12,
DHHS Publication No. (ADM) 83-1246. Washington, DC, DHHS, 1981, pp 306- 322.
21. Roizen J: Alcohol and criminal behavior among blacks: The case for research on special
populations, in Collins JJ Jr. (ed): Drinking and Crime: Perspectives on the Relationships between
Alcohol Consumption and Criminal Behavior. New York, Guilford Press, 1981, pp 207–251.
22. Barr KM, Farrell MP, Barnes GM, Welte JW: Race, class, and gender differences in substance
abuse: Evidence of middle-classiunderclass polarization among black males. Social Problems
40(3):314–327, 1993.
23. Stern MW: Drinking patterns and alcoholism among American Negroes, in Pittman DJ (ed):
Alcoholism. New York, Harper & Row, 1967.
24. Harper FD (ed): Alcohol Abuse and Black America. Alexandria, VA, Douglass Publishers,
1976.
25. Bourne P: Alcoholism in the urban black population, in Harper FD (ed): Alcohol Abuse and
Black America. Alexandria, VA, Douglass Publishers, 1976.
26. Fagan JA, Stewart DK, Hansen KV: Violent men or violent husbands? Background factors
and situational correlates, in Finkelhor D, Gelles RJ, Hotaling GT, Straus MA (eds): The Dark
Side of Families. Beverly Hills, CA, Sage Publications, 1983, pp 49- 68.
27. Lewis BY: Psychosocial factors related to wife abuse. J Fam Viol 2(1):1–10, 1987.
28. Lockhart LL: A reexamination of the effects of race and social class on the incidence of
violence: A search for reliable differences. J Marriage Fam 49:603–610, 1987.
29. Lockhart L, White B: Understanding marital violence in the black community. J lnterp Viol
4(4):421–436, 1989.
30. Lockhart L: Spousal violence: A cross-racial perspective, in Hampton RL (ed): Black Family
Violence: Current Research and Theory. Lexington, MA, Lexington Books, 1991, pp 85- 102.
31. Hampton RL, Gelles RJ, Harrop JW: Is violence in black families increasing? A comparison of
1975 and 1985 national survey rates. J Marriage Fam 51:969–980, 1989.
32. Cazenave NA, Straus MA: Race, class, network embeddedness and family violence: A search
for potent support systems. J Comp Fam Stud 10:280–299, 1979.
33. Hampton RL, Gelles RJ: Violence toward black women in a nationally representative sample
of black families. J Comp Fam Stud 25:105–120, 1994.
34. Asbury J: African-American women in violent relationships: An exploration of cultural differ-
ences, in RL Hampton (ed): Violence in the Black Family: Correlates and Consequences. Lexington,
MA, Lexington Books, 1987, pp 89–105.
35. Roy M (ed): The Abusive Partner. New York, Van Nostrand Reinhold, 1982.
36. Merton RK: Social Theory and Social Structure. New York, Free Press, 1968.
37. Johnson LV, LeMatre M: Anomie and alcohol use: Drinking patterns in Mexican American
and Anglo neighborhoods. J Stud Alcohol 39(5):894–902, 1978.
38. Gray-Little B: Marital quality and power processes among black couples. J Marriage Fam
44(3):633–646, 1982.
39. Caetano R: Self-reported intoxication among Hispanics in northern California. J Stud Alcohol
45(4):349–354, 1985b.
40. Gilbert MJ: Mexican-Americans in California: Intracultural variation in attitudes and behavior
related to alcohol, in Bennett LA, Ames GM (eds): The American Experience with Alcohol:
Contrasting Cultural Perspectives. New York, Plenum Press, 1985, pp 255–277.
41. Paine HJ: Attitudes and patterns of alcohol use among Mexican Americans. J Stud Alcohol
38(3):544–553, 1977.
42. Trotter RT: Mexican-American experience with alcohol: South Texas examples, in Bennett
LA, Ames GM (eds): The American Experience with Alcohol: Contrasting Cultural Perspectives.
New York, Plenum Press, 1986, pp 279–296.
43. Mirande A: The Chicano family: A reanalysis of conflicting views. J Marriage Fam 449:351–873,
1977.
44. Anderson RM, Giachello AL, Aday L: Access of Hispanics to health care and cuts in services:
A state-of-the-art overview. Publ Health Rep 101(3):238–265, 1986.
45. Linsky AS, Straus MA, Colby JJ Jr, Stressful events, stressful conditions and alcohol prob-
lems in the United States: A partial test of Bales’ theory of alcoholism. J Stud Alcohol 46:72–80,
1985.
46. Abad V, Suarez J: Machismo Alcoholism: Cross-Cultural Aspects of Alcoholism among Puerto Ricans.
Proceedings of the Fourth Annual Alcoholism Conference, NIAAA, 1975.
47. Caetano R: Alcohol use among Hispanic groups in the United States. Am J Drug Alcohol Abuse
14(3):293–308, 1988.
48. Trevino ME: Machismo Alcoholism: Mexican-American machismo drinking. Proceedings of the
Fourth Annual Alcoholism Conference NIAAA, 1975.
49. Kaufman Kantor G, Aldarondo E, Jasinski JL: Incidence of Hispanic drinking and intrafamily
violence. Paper presented at the Annual Meetings of the Research Society on Alcoholism,
San Antonio, TX, 1993.
50. Levinson D: Social setting, cultural factors, and alcohol-related aggression, in Gottheil E,
Druley KA, Skoloda TE, Waxman HM (eds): Alcohol, Drug Abuse and Aggression. Springfield,
IL, Charles C. Thomas, 1983, pp 41–58.
51. Aviles-Roig CA: Aspectos Socioculturales del Problema de Alcoholismo en Puerto Rico, in
Tongue E, Lambo RT, Blair B (eds): Report of the Proceedings of the International Conference on
Alcoholism and Drug Abuse. Puerto Rico, 1977, pp 78–85.
52. Jacob T, Leonard KE: Alcoholic- spouse interaction as a function of alcoholism subtype and
alcohol consumption interaction. J Abnormal Psychol 97(2):231–237, 1988.
53. Alcocer AM: Alcohol use and abuse among the Hispanic American population, in: Alcohol and
Health Monograph No. 4: Special Population Issues. (ADM) 82-1193, Rockville, MD, US Depart-
ment of Health and Human Services, 1982, pp 361–382.
54. Caetano R: Ethnicity and drinking in northern California: A comparison among whites,
blacks, and Hispanics. Alcohol Alcoholism 19(1):31–44, 1984.
55. Gordon AJ: Hispanic drinking after migration: The case of Dominicans. Med Anthropol
2(4):61–84, 1978.
56. Gondolf EW, Fisher E, McFerron JR: Racial differences among shelter residents: A compari-
son of Anglo, black, and Hispanic battered women. J Fam Violence 3(1):39–51, 1988.
57. Mirande A, Perez P: Ethnic and Cultural Differences in Domestic Violence: A Test of Conflicting
Models of the Chicano Family. Proceedings of the Research Conference on Violence and Homi-
cide in Hispanic Communities California, UCLA Publications Services, 1988.
58. Torres S: Hispanic American battered women: Why consider cultural differences? Response
10(3):20–21, 1987.
59. Carroll JC: A cultural consistency theory of family violence in Mexican-American and Jewish-
ethnic groups, in Hotaling GT, Straus MA (eds): The Social Causes of Husband–Wife Violence.
Minneapolis, University of Minnesota Press, 1980, chapter 5.
60. Eberstein I, Frisbie WM: Differences in marital instability among Mexican Americans, blacks,
and Anglos: 1960 and 1970. Social Problems 23(5):609–621, 1976.
61. Panitz DR, McConchie RD, Suaber SR, Fonseca JA: The role of machismo and the Hispanic
family in the etiology and treatment of alcoholism in Hispanic American men. J Family Ther
11:31–44, 1983.
62. Becerra R: The Mexican American family, in Mindel CH, Habenstein RW, Wright Jr R (eds):
Ethnic Families in America: Patterns and Variations. New York, Elsevier, Co., 1988, pp. 141–159.
63. Fernandez-Mariña R, Maldonado-Sierra ED, Trent RD: Three basic themes in Mexican and
Puerto Rican family values. J Social Psychol 48:167–181, 1958.
64. Peñalosa F: Mexican family roles. J Marriage Fam 30(4):680–689, 1968.
65. Szapocznik J, Hernandez R: The Cuban American family, in Mindel CH, Habenstein RW,
Wright R Jr. (eds): Ethnic Families in America: Patterns and Variations. New York, Elsevier, 1988,
pp. 160–172.
66. Loya R, Mercy JA, et al: The Epidemiology of Homicide in the City of Los Angeles, 1970–79.
University of California at Los Angeles and Centers for Disease Control, Department of
Health and Human Services, Public Health Service, Centers for Disease Control, 1985.
67. Kaufman Kantor G: Ethnicity, drinking and wife abuse: A structural and cultural interpreta-
tion. Paper presented at the 42nd Annual Meeting of the American Society of Criminology,
Baltimore, MD, 1990.
68. Kaufman Kantor G, Jasinski JL, Aldarondo E: Socioeconomic status and incidence of marital
violence in Hispanic families. Violence Victims 9(3):207–222, 1994,
69. Gordon M: Assimilation in American life: The Role of Race, Religion and National Origins. New
York, Oxford University Press, 1964.
70. Szapocznik J, Scopetta MA, Kurtines W, Aranalde M: Theory and measurement of accultura-
tion. Interam J Psychol 12:113–130, 1978.
71. Burnam AM, Hough RL, Karno M, et al: Acculturation and lifetime prevalence of psychiatric
disorders among Mexican Americans in Los Angeles, J Health Soc Behav 28:89–102, 1987.
72. Holck SE, Warren CW, Smith JC, Rochat RW: Alcohol consumption among Mexican Ameri-
can and Anglo women: Results of a survey along the US- Mexico border J Stud Alcohol
45(2):149–154, 1984.
73. Vega W, Warheit G, Buhi -Auth J, Meinhardt K: The prevalence of depressive symptoms
among Mexican Americans and Anglos. Am J Epidemiol 120:592–607, 1984.
74. Griffith J, Villavicencio S: Relationships among acculturation, sociodemographic characteris-
tics and social supports in Mexican American adults. Hispanic J Behav Sci 7(1):75–92, 1985.
75. Caetano R, Medina Mora ME: Acculturation and drinking among people of Mexican descent
in Mexico and the United States. J Stud Alcohol 49(5):462–471, 1988.
76. Galan RJ: Alcohol Use among Chicanos and Anglos: A Cross-Cultural Study. Unpublished doctoral
dissertation, Waltham, MA, Brandeis University, 1978.
77. Neff JA, Hoppe SK, Perea P: Acculturation and alcohol use: Drinking patterns and problems
among Anglo and Mexican American male drinkers. Hispanic J Behav Sci 9:151–181, 1987.
78. Gordon AJ: The cultural context of drinking and indigenous therapy for alcohol problems in
three migrant Hispanic cultures, in Gottheil E, Druley KA, Skoloda TE, Waxman HM (eds):
Alcohol, Drug Abuseand Aggression. Springfield, IL, Charles C. Thomas, 1983, pp 41–58.
79. Torres S: A Comparative Analysis of Wife Abuse among Anglo-American and Mexican-American
Battered Women: Attitudes, Nature, Severity, Frequency, and Response to the Abuse. Unpublished
doctoral dissertation, Austin, University of Texas, 1986.
80. Sorenson SB, Telles CA: Self-reports of spousal violence in a Mexican-American and non-
Hispanic white population. Violence Victims 6(1):3–15, 1991.
81. Portes A: The rise of ethnicity: Determinants of ethnic perceptions among Cuban exiles in
Miami. Am Soc Rev 49:383–397, 1984.
82. Kaufman Kantor G, Straus MA: Substance abuse as a precipitant of wife abuse victimizations.
Am J Drug Alcohol Abuse 15(2):173–189, 1989.
83. Neff JA, Holamon B, Schluter TD: Spousal violence among Anglos, blacks, Mexican Ameri-
cans: The role of demographic variables, psychosocial predictors, and alcohol consumption. J
Fam Violence 10(1):1–21, 1995.
84. Perilla JL, Bakeman R, Norris FH: Culture and domestic violence: The ecology of abused
Latinas. Violence Victims 9(4):325–339, 1994.
85. O’Leary KD: Physical aggression between spouses: A social learning perspective, in Van
Hasselt VB, Morrison RL, Bellack AS, Hersen M (eds): The Handbook of Family Violence. New
York, Plenum Press, 1988, pp 31–56.
86. Kaufman Kantor G, Asdigian N: Socialization to alcohol-related family violence: Disentan-
gling the effects of family history on current violence. Paper presented at the Annual Meet-
ings of the American Society of Criminology, Phoenix, AZ, 1993.
87. Campbell DW, Campbell J, King C, et al: The reliability and factor structure of the index of
spouse abuse with African-American women. Violence Victims 9(3):259–274, 1994.
88. Williams OJ, Becker RL: Domestic partner abuse treatment programs and cultural compe-
tence: The results of a national survey. Violence Victims 9(3):287–296, 1994.
4
Longitudinal Perspective
on Alcohol Use and Aggression
during Adolescence
Helene Raskin White
Abstract. While there is general agreement that alcohol use and aggression are related, few
studies have examined this relationship among youth. This chapter reviews the literature on
rates of alcohol use, aggression, and alcohol-related aggression among adolescents, as well as the
cross-sectional and longitudinal associations among these behaviors. In general, the literature
does not provide strong support for a unique association between alcohol use and aggressive
behavior during adolescence. The observed relationship between alcohol use and aggression
appears to be spurious because both behaviors are predicted by a similar set of individual, family,
and environmental factors. Prevention programs that reduce these common risk factors should
decrease both behaviors. Interventions with aggressive individuals, especially aggressive indi-
viduals who drink heavily, may be most indicated.
1. Introduction
The association between alcohol use and aggression is well documented.1
Numerous correlational and laboratory studies support the notion that indi-
viduals are more aggressive while under the influence of alcohol.1,2 Statistics
indicate that alcoholics are overrepresented among persons convicted of vio-
lent crimes, and in clinical and nonclinical populations alcohol consumption
is often reported immediately prior to violent offenses.3,4 There is a paucity of
similar statistics on adolescent samples, especially noninstitutionalized sam-
ples. The purpose of this chapter is to examine the relationship between
Helene Raskin White • Center for Alcohol Studies, Rutgers University, Piscataway, New Jersey
08855-0969.
81
alcohol use and aggression among adolescents. First, I briefly discuss models
advanced for explaining the nature and direction of the relationship. Then, I
examine the prevalence of alcohol use, aggression, and alcohol-related ag-
gression in community and institutionalized samples. Next, I present a broad
overview on the degree of association and developmental trends in the rela-
tionship between alcohol use and aggression. I go on to discuss the spurious
nature of the relationships among problem behaviors in adolescence. Finally,
I describe prevention approaches that address alcohol use and aggression in
youth.
Before I begin, let me clarify the scope of the chapter. This review is
limited to research on adolescents. Adolescence, for the purposes of this
review, encompasses approximately the ages of 12 to 18 years, unless other-
wise noted. Aggression is defined as the "intent to harm or create a noxious
condition for the target" (ref. 5, p. 313), whereas violence is defined as "be-
haviors by individuals that intentionally threaten, attempt or inflict physical
harm on others" (ref. 6, p. 2.). Violent behaviors are considered a subset of
aggressive behaviors; the use or threat of physical harm is a necessary ingre-
dient in the definition of violence, but not necessary for aggression. In this
chapter, I focus on the more general category of aggressive behavior and do
not include acts directed at oneself (i.e., suicide). (I also do not discuss labora-
tory studies because they do not include adolescents as subjects. Interested
readers are referred to refs. 2,7 for reviews.) I try to limit this review to alcohol
use. However, much of the research on adolescents focuses on drug use
including alcohol and usually excludes separate analysis on alcohol use alone.
Therefore, I will also discuss other psychoactive drugs when appropriate.
There are several methodological problems and issues that affect the inter-
pretation of the empirical studies on alcohol and aggression. These issues
relate to definition, operationalization, and measurement of variables, analy-
sis strategies, sample selection, and experimental design characteristics. Due
to space limitations these issues are not discussed here; the interested reader
is referred to refs. 1,4,8.
2. Explanatory Models
While it is agreed that aggressive behavior and alcohol use are related,
the extent of a direct causal relationship remains in question. Three primary
theoretical frameworks have been advanced to account for the nature and
direction of the relationship. The first model postulates that alcohol use
causes aggressive behavior due primarily to the psychopharmacological ef-
fects of the drug. This psychopharmacological model proposes that the effects
of intoxication (including disinhibition, cognitive–perceptual distortions, at-
tention deficits, bad judgment, neurochemical changes, etc.) cause aggressive
behavior .5,9,10 It also assumes that situational factors accompanying occasions
of intoxication, such as interpersonal interactions in certain bars, may con-
4 • Alcohol and Adolescent Aggression 83
tribute to aggression.11 In addition, chronic intoxication may contribute to

subsequent aggression due to factors such as withdrawal, sleep deprivation,
nutritional deficits, impairment of neuropsychological functioning, or en-
hancement of psychopathological personality disorders.12 (For greater detail
on the psychopharmacological model, see refs. 1,5,10.)
A second model postulates that aggressive behavior leads to heavy alco-
hol use. This model is based on the assumption that aggressive individuals
are more likely than nonaggressive individuals to select (or be pushed into)
social situations and subcultures in which heavy drinking is condoned or
encouraged. For example, several aspects of the professional crime lifestyle
are conducive to heavy drinking, such as working periodically, partying be-
tween jobs, being unmarried, and being geographically mobile.13 Aggression
can also lead to heavy drinking because aggressive individuals may drink
heavily in order to self-medicate.14 Alternatively, aggressive individuals may
drink to give themselves an excuse to act aggressively. That is, individuals
hold expectations that alcohol use causes aggression,15 and that if they act
aggressively when they are drunk, they will be held less accountable for their
behavior (deviance disavowal) and sanctions will be more lenient.3,11
It is also possible that both of the above models are correct so that the
relationship between alcohol use and aggression is reciprocal. Drinking prob-
lems may lead to more aggression and aggressive behavior may lead to more
drinking.5
The third model is the spurious model, which postulates that alcohol use
and aggression are related either coincidentally or because they share com-
mon risk factors rather than a direct causal link.16 For example, young males
account for a disproportionate share of violent incidents and are also the
heaviest drinkers. Also, common causes (such as genetic or temperamental
traits, antisocial personality disorder, parental modeling of heavy drinking
and aggression, and poor relations with parents) have been shown to predict
both aggression and heavy drinking.17 In addition, subcultural norms may
reinforce both aggression and substance use. Certain subcultures may pro-
mote both aggression and heavy drinking as proof of masculinity, which
would spuriously inflate the relationship.5 Thus, a spurious model argues
that alcohol and aggression are related because similar factors promote both
behaviors. (This model is discussed in greater detail below.)
A routine activities perspective can also explain the spuriousness of the
relationship between alcohol use and aggression. Aggressive behavior occurs
most often when and where people are drinking, such as at bars and sports
stadiums, at night, and on weekends.11 Proponents of this perspective argue
that bars are “ideal” places for violent crime because customers carry cash and
are often too intoxicated to defend themselves, there are weakened social
controls, and the bar atmosphere intensifies competition. Thus, situational
factors also contribute to a spurious relationship between alcohol use and
aggression.11,18
Clearly, no one model can account for all individuals or all types of
aggression. Each of these models may be applicable to different subgroups of

the population or to different incidents of alcohol-related aggression. Other
more complicated models such as interactive, conditional, and conjunctive
models have been described by Pernanen.19 These models suggest that alco-
hol will cause aggression only for certain individuals and only under some
conditions and situations. (For greater detail on these models, see refs.
1,5,19.) Before examining the empirical studies of youth that shed light on
these models, I first examine the prevalence of alcohol use and aggression in
adolescence.
3. The Extent of Alcohol Use and Aggression in Adolescence

In this section, I present data on the prevalence of alcohol use and ag-
gression from early adolescence into young adulthood. The data that are
presented in Fig. 1 come from the Rutgers Health and Human Development
Project (HHDP), a prospective longitudinal study of adolescent development.
A community sample of New Jersey adolescents was tested initially between
1979 and 1981 (time 1, T1) at the ages of 12, 15, and 18 (N = 1380). These
subjects returned 3 years later in 1982–1984 (time 2, T2), again in 1985–1987
(time 3, T3), and finally in 1992–1994 (time 4, T4). Over 90% of the original
sample returned at T4 (N = 1257). The data presented here are based on
subjects who were tested at all four points in time (N = 1201). The sample of
participants is most representative of white, working- and middle-class ado-
lescents living in a metropolitan environment. Results of analyses on validity
and reliability of the data indicate that the sample has a satisfactory degree of
representativeness and that empirical findings have an acceptable degree of
generalizability. (For more extensive details on the methodology of this study
as well as sample recruitment and description, see ref. 20.)
Figure la presents the last-year prevalence rates for alcohol use by age
and sex. There is a marked increase in alcohol use from age 12 to age 18 and
then rates remain fairly steady from age 18 to age 31. Prevalence rates for
males and females are remarkably similar, which is consistent with other
research on young populations.21
Figures 1b and 1c depict age and sex differences in typical quantity (10-
point scale from none to more than two six-packs of beer, a gallon of wine, or
a fifth of hard liquor) and last-year frequency (10-point scale from never to
daily) of drinking, respectively. Frequency increases throughout adolescence,
then peaks in the mid-20s (at a mean of once a week for males and two to
three times a month for females), and declines slightly thereafter (to a mean of
two to three times a month for men and once a month for women). Quantity
appears to peak somewhat earlier (at approximately five drinks per sitting for
males and three drinks for females between ages 18 and 21) and declines in
the late 20s (to an average of three to four drinks for men and two to three for
women. (For greater detail on prevalence of drinking and heavy drinking
over time and sociodemographic differences in a nationally representative

sample, see ref. 22.)
Age and sex differences in rates of (any) aggressive behavior (including
assault, vandalism, using a weapon in a fight, hitting parents, and robbery at
T1–T3 and assault, vandalism, attacking someone, gang fighting, and rob-
bery at T4) are shown in Fig. 1d. Males are more aggressive than females; at
all age levels, male 3-year prevalence rates are at least two to three times
greater than female rates. Aggressive behavior appears to peak at age 15 for
females and ages 15 or 18 for males and then there is a large drop after age 18.
These findings are consistent with other studies.23 Yet our data indicate a
slight increase in aggressive behavior in adulthood, which is inconsistent
with most self-report studies.24 This slight increase probably reflects the dif-
ference in the way the assault question was phrased at T1–T3 (hurt someone
badly) as compared to T4 (hit or threatened to hit someone).
As stated above, the HHDP sample is primarily a white, middle- and
working-class sample. While rates of alcohol use are comparable to those
found in national studies (e.g., ref. 22), rates of aggressive behavior are fairly
low and composed of primarily nonserious offenses. Data from the National
Youth Survey (NYS)25 indicate that self-reported rates of serious violent of-
fending (i.e., robbery, rape, and atrocious assault) peak at age 17 at which
time 36% of the African-American and 25% of the white males report one or
more offense. Peaks for females occur between ages 15 and 16, with approx-
imately 20% of the African-American and 10% of the white females reporting
at least one offense. Not only do females peak earlier, but their decline is
much steeper and the gender differential becomes greater over time. Inter-
estingly, arrest data indicate a later peak than self-report data, and gender
and ethnic differentials are substantially higher in official record studies.25
The NYS data indicate that after age 17 rates of serious violent offending
drop dramatically, and approximately 80% of those who were violent during
adolescence terminate by age 21.24 Nearly twice as many African Americans
as compared to whites continue their offending after age 21. In general, if a
person has not initiated violence by age 20, it is unlikely that she or he will
ever become a serious violent offender (ref. 24, pp. 1–2).
Arrest data indicate that about 6% of juvenile arrests are for violent
crimes.26 Between 1988 and 1992, the number of violent crime index arrests
for juveniles increased by 47%, more than twice the increase for individuals
age 18 and older. In this same time span, juvenile arrests for murder in-
creased by 51% as compared to 9% for adults (ref. 26, p. 1). (For greater detail
on the increases in juvenile violent crime by crime type, see ref. 27.) It should
be noted that juvenile rates for violent offenses are still much lower than adult
rates, so even a large percentage increase does not necessarily translate to a
large contribution to the overall rate of violence.27 Data from the National
Crime Victimization Survey indicate that in 1991, 19% of all violent crimes
were committed by juveniles.27 Nine out of 10 of these crimes were commit-
ted by males.
Age (years )
1a: Last year prevalence of alcohol use
Age (years)
1b: Typical quantity of alcohol use

Age (years)
1c: Alcohol use frequency in the last year
Age (years)
-
1d: Three-year prevalence of aggressive behavior
Figure 1. Longitudinal patterns in alcohol use and aggression by age and sex. (a)
Last year prevalence of alcohol use. (b) Typical quantity of alcohol use. (c) Alcohol use
frequency in the last year. (d) Three-year prevalence of aggressive behavior.
While this chapter focuses on involvement in aggressive behavior rather

than on victimization, it should be noted that teenagers and young adults
have the highest rates of victimization for crimes of violence.26,28 Half of all
rape victims are females under age 18, and homicides of teenagers have
doubled from 1984 to 1991, with rates especially high for African-American
males. Persons most likely to be victimized by juveniles are also juveniles.27
When one examines demographic variable differences in the patterns of
aggression and alcohol use, several discrepancies emerge. First, alcohol use
tends to peak later than violence and declines more steadily. Second, there
are large gender differences in violence but not in the use of alcohol. In
addition, African Americans as compared to whites have higher rates of vio-
lence yet lower rates of alcohol use. Thus, to some extent, these differences in
trends argue against a direct causal relationship between alcohol use and
violence.29 Data that specifically address the relationship between alcohol use
and aggression are discussed below.
4. The Association between Alcohol Use and Aggression in

Adolescence
4.1. Acute Incidents of Alcohol-Related Aggression
The HHDP study is one of the few that directly assesses acute incidents
of alcohol-related aggression. Subjects were asked to indicate the number of
times they engaged in aggressive behaviors (including fighting or acting
mean, vandalism, fire setting, forcing sex, and hurting someone badly) while
they were drinking or because of their drinking. Except for fighting (around
one fourth of females users), very few females have engaged in any type of
alcohol-related aggression (1 to 2% of the users for most categories). Rates for
fighting while intoxicated are very high for males and appear to increase with
age from about 30% of the 15-year-old users, to about one third to one half of
the 21-year-old users, and then drop off slightly at age 24 and continue to
drop off into the 30s (to about one fifth of the users). Rates for alcohol-related
vandalism among males are relatively high in late adolescence and early
adulthood (around 20% of the users) and then drop off considerably. Forced
sex (around 3% of the users or less), fire setting (2% or less), and hurting
someone badly (7% or less) while drinking are relatively rare (see ref. 20).
Mitic30 found that approximately 6% of junior high and 15% of high
school students reported engaging in vandalism or violence toward others
while drinking. Elliott and colleagues23 examined self-reported alcohol use
immediately prior to committing an offense. When subjects were 11–17 years
old, 23% of the aggravated assaults (including gang fights) and 10% of the
robberies were committed under the influence of alcohol only. In comparison,
20% of the motor vehicle thefts were also committed under the influence of
alcohol. As these subjects entered young adulthood, however, alcohol use
was more likely to precede an offense characterized by physical violence than
one characterized by a profit motive. The NYS data suggest that the nature of
the relationship between alcohol use and violent offenses may change over
the life course. Taken together, the above studies do not suggest that rates of
alcohol-related aggressive behaviors (except perhaps fighting) are especially
high during adolescence.
Similarly, research on adolescent offenders indicates relatively low rates
of alcohol use at the time of violent offenses.3 Only 8% of all youth in custody
in state institutions reported that they were under the influence of only alco-
hol when they committed a violent offense.31 On the other hand, 24% com-
mitted violent offenses while under the influence of both alcohol and other
drugs. Yet this percentage is the same as that for property crimes committed
under the influence of alcohol and drugs (23%). Thus, it appears that some
youth are committing crimes while under the influence of alcohol (just as
they are engaging in all sorts of other activities while intoxicated); however, a
unique effect of alcohol on violent crime is not borne out by these statistics. In
fact, studies of youth gangs suggest that alcohol plays a role in facilitating
intragroup cohesion as well as intergroup conflict (for greater detail on alco-
hol use among gang members, see ref. 11).
When adolescents are questioned about the contribution of alcohol intox-
ication to their aggression, mixed findings emerge. In a community interview
study, most adolescents held expectancies that alcohol/drug use causes crime
(due to disinhibition, economics, etc.) and attributed other people’s crime to
their drug use.32 These same adolescents did not attribute their own crime to
their drug use. The subjects said that they kept their drug use under control
and it did not interfere with or cause their participation in delinquency. The
researchers concluded that alcohol and drug use was incidental to and not a
contributing factor for crime. Although they found that heavy drug users
were most involved in violence, most of the violent incidents occurred while
subjects were not using alcohol or drugs. A recent study of inner-city adoles-
cents33 supports these findings. Very few adolescents admitted that they
used drugs while committing offenses or that they committed an offense in
order to obtain drugs or money for drugs. Unfortunately, alcohol use was not
included as a measure in this study.
In contrast, in a study of adolescents who were adjudicated for a violent
crime, over half of the youths said that taking alcohol (29%) or drugs (33%)
contributed to their acting violently, and almost half had used either alcohol
(17%) or drugs (34%) immediately prior to their adjudicated violent offense.34
Note, however, that the rates were higher for other drugs than for alcohol. In
a study of incarcerated adolescents, it was found that over two thirds of the
incidents of physically assaultive crime involved acute drug intoxication.35
Almost all of the cases of acute intoxication involved alcohol either alone or in
combination with another drug. Similarly, a large majority of the drug-related
sexually assaultive crimes involved alcohol use. On the other hand, mari-
juana use was underreported in offenses against persons. Secobarbital (a
sedative drug) was selected over alcohol as the drug most likely to lead to a
fight, while marijuana was selected as the drug most likely to decrease assaul-
tiveness. It is possible in both these studies that arrested juveniles over-
reported alcohol or drug use prior to their offense in order to justify their
behavior.3
The above studies on degree of association looked at acute incidents of
alcohol-related aggression. Many other studies have examined the associa-
tions between patterns of alcohol and aggression either through correlations
or typologies, and some of these will be reviewed below.
4.2. Associations between Patterns of Alcohol Use and Aggression

Using the HHDP data, we examined the correlations between alcohol use
(typical quantity times frequency) and aggression (sum of the five behaviors
included in the prevalence analysis above). The cross-sectional correlations at
each test occasion indicated that the two behaviors were significantly related
with an average correlation (r) of .23. These correlations were higher for males
(mean r = .22) than females (mean r = .13) and also higher for quantity (mean
r = .22) than frequency (mean r = .15). Finally, it should be noted that the
correlations between alcohol use and property crime (mean r = .30) were
higher than those for alcohol use and aggression. Thus, these data do not
support a unique relationship between alcohol as opposed to other drugs and
aggression.
Several other studies have looked at the correlations between alcohol use
and aggression. Most of these studies have been cross-sectional, although a
few have been longitudinal. Cohen et al.7 conducted a meta-analysis of the
cross-sectional studies. In these analyses they did not separate studies of
adolescents from studies of adults. They found that the average correlation
for criminal (as compared to domestic) violence and alcohol use at the time of
the offense was .12, while that for criminal violence and general pattern of
alcohol use (not necessarily at the time of the offense) was .16. In practical
terms, a mean correlation of .12 is equivalent to a contrast between a low-
alcohol-use group that is 10% violent and a high-use group that is 15% violent
(ref. 7, p. 27). While this difference may appear substantial, correlations of
less than .2 are relatively weak and indicate that about three fourths of the
variation in each behavior is not accounted for by the other. Cohen and
colleagues concluded that greater alcohol use is associated with higher levels
of violent behavior. They were careful to note that this association in no way
supports a causal connection between alcohol use and violence and may be
due to confounding factors. Further, while the correlations between criminal
violence and drug use patterns were lower than the alcohol–violence correla-
tions for most drugs, the correlation was higher for multiple drug use. Of
greater interest is that the mean correlation for alcohol–nonviolent crime was
higher than the mean alcohol-violence correlation, and this finding was ob-
tained across studies of primarily adolescent samples. All of the studies re-
viewed identified control variables that accounted for a significant portion of
the variance in violence independent of the effect of alcohol. In fact, in about

half of the studies, once these other factors were statistically controlled, the
relationship between alcohol use and criminal violence was no longer signifi-
cant. In examining differences in samples, Cohen et al. found that the criminal
violence–alcohol correlation was higher in adolescent (under age 21) than in
adult samples.
Similarly, Osgood29 examined correlations between self-reported sub-
stance use and various types of delinquency from the Monitoring the Future
Study. He found that the correlations between violence and other drugs were
generally the same as between violence and alcohol use (all coefficients
around .2), but were lower than for alcohol/other drugs and theft. In an
inner-city sample, Fagan and his colleagues36 reported higher correlations
between alcohol use and the incidence of violent crime than those reported
above (coefficients around .3). However they also found that the correlations
between alcohol and violent offenses were generally lower than those be-
tween alcohol and property crime.
In sum, the correlational studies reviewed above indicate that the asso-
ciation between alcohol use and aggression is similar to or weaker than the
association between alcohol use and other forms of delinquency. Thus, both
correlation studies, as well as self-reports of attribution, do not provide strong
support for a unique association between alcohol use and aggression in com-
munity samples. In fact, after a review of the literature on alcohol, drugs, and
violence among youth, Osgood29 (p. 33) concluded that there is little evidence
that substance use makes an independent contribution to adolescent vio-
lence.
While research on adolescents does not provide strong support for a
direct relationship between alcohol and aggression, the data suggest that
alcohol and drug users as opposed to nonusers are more likely to be delin-
quent and more likely to be involved in aggression.29 Carpenter and col-
leagues32 found that those youth involved in violence were also more in-
volved in delinquency and alcohol/drug use than their peers. Not only were
those heavily involved individuals more often perpetrators, but also they
were more often victims of violence. Major assaults were committed by the
heaviest drug users, while alcohol-only users had few felony violence of-
fenses. Those subjects most involved in gang fights were also the heaviest
drug users.
In studies of inner-city students, Fagan and his colleagues36,37 found that
substance use was prevalent regardless of level of delinquency. They found
that prevalence of alcohol use did not vary across delinquency types, al-
though frequency did. On the other hand, prevalence of marijuana and other
illicit drugs increased from nondelinquents to multiple-index offenders. In
the NYS data, Elliott et al.23 found that level of violence increased across
groups from nonusers to alcohol users to marijuana users to polydrug users.
Thus, there were greater levels of violence among polydrug users and mari-
juana users (who also used alcohol) than among those who only used alcohol.
Similarly, the level of alcohol use increased from nondelinquents to explora-

tory delinquents to nonserious delinquents to serious delinquents.
A US national study found that drug (including alcohol) users were more
likely than nonusers to fight, to take risks that predisposed them to assault,
and to be victimized both in school and outside school supervision.38 A study
of high school juniors and seniors found that binge drinking was a strong
predictor of fighting and that alcohol use was a strong predictor of carrying a
weapon.28 The authors noted that both behaviors may have been caused by a
common factor.
Watts and Wright39 studied white, African-American, and Hispanic adju-
dicated delinquents and high school students. They found that, across ethnic
groups, the best predictor of violent delinquency was use of illegal drugs. In
the multivariate analyses including all drug types, alcohol use was not a
significant predictor of violent delinquency for any of the three ethnic groups,
and, in fact, the beta weight was negative.
After reviewing the literature on the relationship between alcohol use
and crime during adolescence, Collins40 concluded that alcohol use by itself is
not important to the occurrence of serious criminal involvement. Although
those who drink and drink heavily are more likely to be involved in other
forms of deviance, the association is probably spurious due to common eti-
ologies. The spurious model is discussed in greater detail below. First, how-
ever, I review findings on the longitudinal associations between alcohol use
and aggression.
4.3. Developmental Trends

Research suggests that involvement in delinquency occurs prior to or
simultaneous with the first drink.23,25 Elliott and colleagues23 found that mi-
nor delinquency almost always came first in the sequence of delinquency and
drug use, and, in fact, no one initiated marijuana or polydrug use before
minor delinquency. Alcohol use came second; however, a substantial per-
centage of subjects initiated index offenses prior to alcohol use. In general,
however, after alcohol use came marijuana use, then index offending, and
finally polydrug use. Among subjects who initiated both marijuana use and
index offending, index offending was more likely to precede marijuana use
than vice versa. The researchers concluded that the onset of minor delinquen-
cy leads to the onset of drug use and not vice versa. Further, they stressed
that while this sequence represents the dominant pattern, many adolescents
do not adhere to it. More recent analyses of these data25 indicate that for a
substantial majority of adolescents, minor delinquency and alcohol use oc-
curred simultaneously. In general, marijuana use followed next, and most
adolescents who engaged in both index offending and multiple drug use
engaged in index offending first. In terms of serious violent offending, the
typical sequence was first aggravated assault, then robbery, and then rape.25
Although research indicates that conduct disorders and delinquency pre-
cede the initiation into drug and usually alcohol use, studies also show that
increases in property crime and acquisitive violence occur subsequent to regu-
lar use of hard drugs among adolescents.41 Elliott and colleagues23 found that
while delinquency is more likely to influence the onset of alcohol/drug use
than the reverse, serious drug use (repeated polydrug use) is more likely to
influence the maintenance of serious delinquency than the reverse. This re-
sult suggests the possibility that if alcohol/drug use does influence delinquen-
cy, it may be by reducing the probability of terminating rather than by in-
creasing the probability of initiating delinquent behavior. Collins40 also stated
that while problem drinking may not be important for the onset of a criminal
career, problem drinking may intensify or prolong serious involvement in
criminal behavior in young adulthood. In sum, the longitudinal research
indicates that initiation into aggressive behavior generally precedes alcohol/
drug use; however, changes in drug and maybe alcohol use affect changes in
aggression (see also ref. 42).
Longitudinal research has demonstrated that childhood antisocial behav-
ior (usually defined as conduct disorders, delinquency, or aggressive behav-
ior) is consistently related to the later development of alcohol problems in
adolescence and adulthood.43,44 However, Loeber’s41 review of the literature
on antisocial behavior indicated that nonaggressive rather than aggressive
antisocial behaviors are predictive of later substance use or abuse, while ag-
gression is more strongly predictive of other forms of delinquency such as
theft, burglary, and fraud. Similarly, White45 found that alcohol and drug
problems in late adolescence were predicted by nonaggressive delinquent
behavior but not by aggressive behavior in early adolescence.
McCord46 followed a cohort of 390 males over four decades and divided
the sample into four groups based on aggression in childhood (high or low)
and alcoholism at follow-up (present or absent). She found that both early
aggressiveness and later alcoholism contribute to the probability of later anti-
social behavior, but she was not able to ascertain the relationship between
early alcohol use and aggressive behavior.
Using the HHDP data, White and colleagues17,47 examined the longitudi-
nal associations among alcohol use, aggression, and incidents of acute alco-
hol-related aggression from early adolescence into adulthood. They found
that aggressive behavior was very stable for males from age 12 to age 24,
while alcohol use was not very stable from age 12 to age 15 but became highly
stable from age 15 on. Other longitudinal studies also attest to the stability of
aggressive behavior over the life course (e.g., refs. 48,49). Farrington50 found
that boys who were aggressive in childhood or adolescence tended to be more
violent in adulthood and to engage more often in violent and nonviolent
offenses. In addition, as adults these boys were more likely to be heavier
drinkers and smokers, drunk drivers, and drug takers. Farrington suggested
that this continuity is probably not specific to aggression, but rather part of a
general continuity in antisocial behavior from childhood to adulthood. On the
other hand, several longitudinal studies in the alcohol field have shown that
alcohol involvement or alcohol problems remain steady for some individuals,

whereas they are constantly shifting or changing for others (e.g., refs. 44,51).
White and colleagues17,47 found that early aggressive behavior predicted
later alcohol use, but alcohol use was not related to subsequent increases in
aggressive behavior. The findings also indicated that prior alcohol use was a
better predictor of alcohol-related aggression for females, while prior aggres-
sion was a better predictor for males.
Cohen and colleagues7 conducted a meta-analysis of the relationship
between alcohol use and criminal violence over time from the few longitudi-
nal studies that met their criteria. The meta-analysis indicated fairly weak
associations across time. The average correlation between time 1 alcohol use
and time 2 violence was .01 and between time 1 violence and time 2 alcohol
use was .09. The fact that the latter was stronger than the former is consistent
with the data presented above from the HHDP. Given the small number of
studies, the researchers were unable to draw any conclusions about temporal
sequencing. Overall, the results of the meta-analysis suggested that the evi-
dence was insufficient to establish a causal relationship between alcohol use
and violence.7
The studies reviewed above suggest that, in general, aggression occurs
prior to alcohol use in the life course, and aggression is a better predictor of
alcohol use than alcohol use is of aggression. In contrast, a study of juvenile
offenders in Finland found that those juveniles who had arrests for drunken-
ness were more likely to have arrests for both violent and property crimes 5 to
10 years later.52 Also, Dembo and colleagues53 found that early alcohol use
was an important predictor of later violent behavior among juvenile de-
tainees.
While victimization is not the focus of this chapter, from a developmental
perspective research on the consequences of childhood exposure to violence
provides some interesting results. For example, studies indicate that alcoholic
females have significantly higher rates of childhood sexual and physical abuse
than nonalcoholic females.54 (For greater detail on victimization and alcohol
use, see refs. 54,55), It has also been demonstrated that exposure to violence
in the family in childhood leads to greater participation in violence in adoles-
cence. In one study, both maltreatment in childhood (i.e., physical abuse,
sexual abuse, or neglect), as well as growing up in a home with violence (e.g.,
spouse abuse, family hostility, maltreatment of siblings), significantly in-
creased the chances of an individual engaging in violent behavior in adoles-
cence.56 In addition, children who grew up in families with multiple forms of
violence had twice the risk of engaging in violence in adolescence as com-
pared to children who grew up in nonviolent families (for a review, see ref.
57). Similarly, studies have found that children who grew up in alcoholic
homes were at a greater risk for exhibiting aggressive behavior in adoles-
cence.58
The above studies suggest that parental violence and alcoholism can
increase the risk for later violence and alcoholism. Therefore, a negative fami-
ly background is indicated as one common cause of both aggression and

heavy drinking. Other common causes are discussed in the next section.
5. A Common-Cause Model
The general consensus is that the relationship between alcohol use and
aggression in adolescence is spurious.16 As discussed above, developmen-
tally, aggressive behavior generally precedes alcohol use initiation and early
aggressive behavior predicts later alcohol use, but in most studies early alco-
hol use does not predict later aggression. Further, most acts of aggression by
adolescents occur in the absence of alcohol use. Thus, alcohol use does not
appear to cause aggressive behavior, but rather both are probably caused by a
similar set of factors.
Evidence supporting a common-cause model has been derived from ado-
lescent samples where relatively nonserious forms of substance use appeared
to occur simultaneously with relatively minor and infrequent forms of delin-
quent behavior. This association led Jessor and Jessor59 to identify a problem
behavior syndrome in which cigarette use, precocious sexual behavior, prob-
lem drinking, use of marijuana and other drugs, stealing, and aggression
clustered together. This cluster of behaviors was explained by the same set of
environmental and personality variables and was negatively related to con-
ventional behavior.
It appears, therefore, that alcohol use and aggressive behavior are related
to each other during adolescence because of a tendency for adolescents, who
share a similar set of risk factors, to experiment with a wide range of deviant
behaviors. Nevertheless, not all adolescents who engage in one behavior
engage in the others. Further, the risk factors may not have equal effects on
each behavior. Hence, the generality versus specificity of deviance has been
debated throughout the literature (e.g., refs. 45,60). According to one posi-
tion, alcohol use and aggression serve similar functions and are, therefore,
conceptualized as constituent behaviors of a more general problem behavior
syndrome59 or a general criminal propensity.61 The alternative position views
alcohol use and aggression as different and relatively independent manifesta-
tions of deviance.36 Although the notion of common causes and common
functions may apply to some individuals and help to account for the observed
relationships among deviant behaviors, specific forms of deviance are equally
likely to be shaped by causes and functions that are specific to each of the
various problem behaviors. In other words, it may be reasonable to assume
that the concept of a problem behavior syndrome applies only to a minority of
adolescents. Some adolescents may be undifferentiated "generalists," others
more differentiated "specialists"in deviance. That is, some adolescents may
engage in a wide variety of deviant behavior including alcohol use and ag-
gression (see ref. 61). At the same time, other groups of adolescents may
drink heavily without engaging in aggressive behavior and vice versa (see
refs. 36,62,63). In discussing this debate, Osgood64 noted that a single set of
explanatory factors might account for a large portion of the variance in each of
a strongly related subset of problem behaviors. However, there are also im-
portant unique features of each behavior that require separate explanations.
White and Labouvie’s60 findings support the notion that the expression
of deviance is not necessarily undifferentiated. Instead, their results indicated
that a considerable number of individuals seem to specialize in particular
problem behaviors. As compared to a group of generalists, both delinquent
specialists and drug use specialists displayed a more differentiated pattern of
coping reactions and all three groups differed in terms of psychological prob-
lems and personality traits. Subjects who were high in both drug use and
delinquency (i.e., generalists) had significantly higher levels of mental health
problems than all other subjects.
Most studies of young violent offenders indicate that they are generalists
and only a small proportion of their annual offenses are violent offenses (e.g.,
refs. 25,33,34). In the NYS data, serious violent offenders constituted less
than 5% of the entire sample, and yet they accounted for 83% of all the
reported index offenses and more than 50% of all offenses.25 In another
analysis of these data, it was found that less than 5% of all youth reported
serious crimes and used hard drugs. This small group accounted for approx-
imately 40% of all delinquencies, 60% of all index offenses, over 50% of all
felony assaults, over 60% of all felony thefts, 75% of all robberies, over 80% of
all drug sales, 30% of all marijuana use occasions, and 60% of all other drug
use occasions.65
The relationship between alcohol use and aggression in adolescence,
which we have seen is relatively weak, is best explained by a common-cause
model. Nevertheless, we cannot assume that both behaviors are caused by
the exact same set of predictors, or that individuals who have certain risk
factors will necessarily engage in both behaviors. In fact, having any specific
risk factor or set of factors is not a guarantee that an individual will engage in
either behavior. Rather, there are life experiences and opportunities as well as
protective factors that can mediate the relationship between risk factors and
problem outcomes (see ref. 66).
With these caveats in mind, an examination of the most often-cited pre-
dictors of violence and alcohol use reveals considerable overlap. For example,
the National Research Council Report6 identified numerous risk factors for
violence, including hyperactivity, impulsivity, attention deficit disorder, rest-
lessness, lack of concentration, risk taking, inability to delay gratification, low
empathy and low IQ, abnormal frequency of viewing television violence,
bullying in early years, harsh and erratic discipline, abuse or rejection in
family, lack of parental nurturance, low income, large family, familial criminal
behavior, early school failure, peer rejection, poor housing, and growing up
in a high-crime neighborhood. Interestingly, almost every one of these risk
factors (except, perhaps, violence on television, low income, low IQ, and
poor housing) have also been identified in the literature as risk factors for
teenage alcohol/drug use or for adult alcohol and drug problems.66,67 (For
other reviews of risk factors, see refs. 68–71.)
In addition to common risk factors, the same theories have been applied
successfully to explain delinquency/violence and alcohol/drug use. The most
often-tested theories are control theory,72 differential association theory,73
and integrations of the two (e.g., ref. 74). However, a social learning model
that takes into account both person and environmental factors may provide
the most comprehensive explanation for the specific relationship between
alcohol use and aggression. Such a model has been applied to each behavior
separately.48,75–77 According to Bandura’s78 social learning model, human
behavior is explained in terms of a continuous reciprocal interaction between
the person, the environment, and behavior. The person brings to the environ-
ment a unique endowment (including genetic makeup, temperament, per-
sonality, expectancies) that imposes certain limits on what and how things are
learned and influences the person’s selection of various environments. Ob-
servational learning (i.e., modeling) and reinforcement (i.e., reward and pun-
ishment) are the two primary social learning processes by which influences
from the environment are transmitted. Individuals in the proximal (e.g., fami-
ly and peers) and distal (e.g., social milieu, media) environments model and
reinforce behavior either directly or vicariously, and thus teach individuals
the appropriate manner in which to behave.76 Based on a social learning
model, one would hypothesize that alcohol use and aggression will be related
in certain individuals who have expectancies that alcohol use causes aggres-
sion, have genetic or temperamental traits that foster aggressive behavior
(e.g., impulsivity, hyperactivity, etc.), and have been exposed to significant
others who separately or simultaneously model and reinforce aggressive be-
havior and heavy drinking.17 Thus, research on risk factors will need to
address the interactions among individual and environmental variables.
In sum, this review of the literature suggests that the relationship be-
tween alcohol use and aggression among youth is spurious, and both aggres-
sive behavior and heavy drinking may be caused by a third factor or set of
factors. Individual, family, peer, and environmental factors account for a
clustering of risk factors for both behaviors. This common-cause model has
certain implications for prevention strategies, and in the following section
these strategies are discussed.
6. Preventing Alcohol Abuse and Aggression
Because the risk factors for violence overlap with those for substance
abuse, several of the prevention programs in each area have similar goals and
approaches. Youth prevention programs for both alcohol abuse (for reviews,
see refs. 79–81) and violence (for reviews, see refs. 6,68,70,71) target the
individual, the family, peers, and/or the community. Hawkins and col-
leagues71 have argued that efforts to prevent violence and substance abuse
should be combined because of the overlap of risk factors.
The general consensus is that prevention approaches need to be develop-
mental because risk factors differ over the life course.69 For example, in the
infant and preschool period, programs appear most effective that address
parenting skills and needs, as well as healthy physical and cognitive develop-
ment of the child.71 As children become school-age, in addition to parent
programs, selected school programs have also had positive results (e.g., re-
ducing kindergarten and first-grade class sizes, improving instructional strat-
egies, and monitoring behavior) (for a complete list, see ref. 70). Overall, most
of the reviews agree that early prevention is needed, that working with par-
ents or soon-to-be parents is advantageous, that programs should be aimed at
multiple risk factors, and that continuity of programs from birth into adoles-
cence is important.69–71 Although less evaluated, some societal-level ap-
proaches seem worthwhile, including reducing exposure to violence in the
media, reducing access to lethal firearms, and changing norms regarding
alcohol use.68,79 On the other hand, reviews consistently have found that
most peer programs (e.g., resistance skills, peer mediation, guided group
interaction, recruiting out of gangs) are ineffective for both violence68,70 and
substance abuse.80
The current trend in prevention is toward encouraging community-wide
approaches,71 although evaluations of community interventions for alcohol
abuse have not demonstrated positive results.81 Several prevention specialists
stress the need to increase protective factors (e.g., resilience and a positive
social orientation, social bonding to prosocial individuals and institutions,
and healthy beliefs and standards for behavior) in addition to decreasing risk
factors.70,71
Many of the current prevention programs have not been properly evalu-
ated, and when they have been, data have been weak.68 Thus, more sophisti-
cated evaluation designs need to be implemented (see ref. 68 for suggestions).
However, the existing data from well-evaluated programs show that some of
these programs can be effective. That is, early childhood prevention has had a
positive impact on reducing the risk factors for later violence and substance
abuse.69
7. Conclusions
The empirical evidence suggests that aggressive behavior is highly stable
during adolescence17,47 and into adulthood48,49 and that many adult violent
offenders were aggressive as youth.24 Thus, an aggressive behavior pattern is
formed at an early age and sets the stage for later violence as well as later
alcohol use and alcohol-related aggression.17 Further, it has been demon-
strated that intoxication facilitates aggression most in those individuals al-
ready inclined to aggression.1 Thus, early intervention with aggressive chil-
dren (especially males) would help prevent later violence as well as later
alcohol problems and alcohol-related aggression.
It should be kept in mind, however, that many adolescents mature out of
aggressive behavior and heavy drinking as they reach young adulthood.
Therefore, efforts should be made to identify those individuals who will
persist in deviant behaviors. Research identifying these individuals is cur-
rently underway and suggests that persistence may be related to early onset
as well as cognitive dysfunctions, impulsivity, and undercontrol.82
Obviously, the group of adolescents high in both alcohol use and aggres-
sive behavior clearly stands out as in need of attention. White and Labouvie60
found that cumulative involvement in a range of problem behaviors during
adolescence was coupled with the highest levels of distress and negative
affect as well as the highest levels of impulsivity, undercontrol, and emotional
instability. Thus, it is reasonable to expect that generalists (i.e., those adoles-
cents who engage in a wide range of problem behaviors) will be least likely
to make successful transitions into adulthood and most likely to develop
full-blown substance abuse and/or psychopathology in adulthood. For these
reasons, adolescent generalists in deviance would seem to constitute an im-
portant target for intervention programs. Such programs would require strat-
egies that address a multitude of problem behaviors and a multitude of un-
derlying psychological problems as well as other risk factors. At the same
time, heavy drinkers or aggressive behavior specialists may benefit more from
specific interventions designed to address their individual problems.
ACKNOWLEDGMENTS. Preparation of this chapter was supported, in part, by
grants from the National Institute on Drug Abuse (#DA/AA-03395) and the
Alcoholic Beverage Medical Research Foundation. The author thanks Patricio
Calderon for his assistance with preparation of the manuscript and Drs. Carl
Danziger and Allan Horwitz for their comments on an earlier draft of this
manuscript.
References
1. White HR: Alcohol, illicit drugs, and violence, in Stoff D, Brieling J, Maser JD (eds): Handbook
of Antisocial Behavior. New York, John Wiley & Sons, in press.
2. Bushman BJ, Cooper HM: Effects of alcohol on human aggression: An integrative research
review. Psychol Bull 107:341–354, 1990.
3. Collins JJ: Drinking and violence: An individual offender focus, in Martin SE (ed): Alcohol and
Interpersonal Violence: Fostering Multidisciplinary Perspectives. Rockville, MD, National Institute
of Health, NIAAA Research Monograph No. 24, 1993, pp 221–235.
Interpersonal Violence: Fostering Multidisciplinary Perspectives. Rockville, MD, National Institute
of Health, NIAAA Research Monograph No. 24, 1993, pp 3-36.
5. Fagan J: Intoxication and aggression, in Tonry M, Wilson JQ (eds): Drugs and Crime. Chicago,
University of Chicago Press, 1990, vol 13, pp 241–320.
6. Reiss AJ, Roth JA (eds): Understanding and Preventing Violence. Washington DC, National
Academy Press, 1993.
7. Cohen MA, Lipsey MW, Wilson DB, et al: The role of alcohol consumption in violent behav-
ior: Preliminary findings. Nashville, TN, Vanderbilt University, 1994.
8. Fagan J: Interactions among drugs, alcohol, and violence. Health Affairs 12;65–79, 1993.
9. Collins JJ (ed): Drinking and Crime. New York, Guilford, 1981.
10. Pihl R, Peterson J: Alcohol and aggression: Three potential mechanisms of the drug effect, in
Martin SE (ed): Alcohol and Interpersonal Violence: Fostering Multidisciplinary Perspectives. Rock-
ville, MD, National Institute of Health, NIAAA Research Monograph No. 24, 1993, pp 149–
159.
11. Fagan J: Set and setting revisited: Influences of alcohol and illicit drugs on the social context
of violent events, in Martin SE (ed): Alcohol and Interpersonal Violence: Fostering Multidisciplin-
ary Perspectives. Rockville, MD, National Institute of Health, NIAAA Research Monograph
NO. 24, 1993, pp 161–191.
12. Virkkunen M, Linnoila M: Brain serotonin, type II alcoholism and impulsive violence. J Stud
Alcohol (Suppl 11): 163–169, 1993.
13. Collins JJ, Messerschmidt PM: Epidemiology of alcohol-related violence. Alcohol Health Res
World 17:93–100, 1993.
14. Khantzian EJ: The self-medication hypothesis of addictive disorders: Focus on heroin and
cocaine dependence. Am J Psychiatry 142:1259–1264, 1985.
15. Lindman RE, Lang AR: The alcohol-aggression stereotype: A cross-cultural comparison of
beliefs. Int J Addict 29:1–13, 1994.
16. White HR: The drug use-delinquency connection in adolescence, in Weisheit R (ed): Drugs,
Crime and the Criminal Justice System. Cincinnati, OH, Anderson Publishing, 1990, pp 215–256.
17. White HR, Brick J, Hansell S: A longitudinal investigation of alcohol use and aggression in
adolescence. J Stud Alcohol (Suppl 11): 62-77, 1993.
18. Roncek DW, Maier PA: Bars, blocks and crimes revisited: Linking the theory of routine
activities to the empiricism of "hot spots." Criminology 29:725–753, 1989.
JJ (ed): Drinking and Crime. New York, Guilford Press, 1981, pp 1–69.
20. White HR, Hansell S, Brick J: Alcohol use and aggression among youth. Alcohol Health Res
World 17:144–150, 1993.
21. White HR, Huselid RF: Gender differences in alcohol use during adolescence, in Wilsnack R,
Wilsnack S (eds): Gender and Alcohol. New Brunswick, NJ, Rutgers Centers of Alcohol Stud-
ies, 1977, 176–198.
22. Johnston LD, O’Malley PM, Bachman JG: National Survey Results on Drug Use from the Monitor-
ing the Future Study, 2975–2993. Rockville, MD, National Institute on Drug Abuse, 1994.
23. Elliott DS, Huizinga D, Menard S: Multiple Problem Youth: Delinquency Substance Use and
Mental Health Problems. New York, Springer-Verlag, 1989.
24. Elliott DS: Youth Violence: An Overview. Boulder, CO, The Center for the Study and Preven-
tion of Violence, 1994.
25. Elliott DS: Serious violent offenders: Onset, developmental course, and termination-the
American Society of Criminology 1993 Presidential Address. Criminology 32:1–21, 1994.
26. Allen-Hagen B, Sickmund M, Snyder HN: Juveniles and Violence: Juvenile Offending and Victim-
ization. Washington, DC, Office of Juvenile Justice and Delinquency Prevention, US Depart-
ment of Justice, 1994.
27. Snyder HN, Sickmund M: Juvenile Offenders and Victims: A Focus on Violence (Statistics summa-
ry). Washington, DC, Office of Juvenile Justice and Delinquency Prevention, US Department
of Justice, 1995.
28. Valois RF, Vincent ML, McKeown RE, et al: Adolescent risk behaviors and the potential for
violence: A look at what's coming to campus. J Am College Health 41:141–147, 1993.
29. Osgood DW: Drugs, alcohol, and adolescent violence. Presented at the American Society of
Criminology annual meeting, Miami, FL, 1994.
30. Mitic WR: Adolescent drinking problems: Urban vs. rural differences in Nova Scotia. Cand J
Commun Mental Health 8:5–14, 1989.
31. Bureau of Justice Statistics: Fact Sheet: Drug-Related Crime. Washington DC, US Department of
Justice, 1994.
32. Carpenter C, Glassner B, Johnson BD, et al: Kids, Drugs, and Crime. Lexington, MA, Lex-
ington Books, 1988.
33. Altschuler DM, Brounstein PJ: Pattern of drug use, drug trafficking, and other delinquency
among inner-city adolescent males in Washington, DC. Criminology 29:589–622, 1991.
34. Hartstone E, Hansen KV: The violent juvenile offender: An empirical portrait, in Mathias RA,
DeMuro P, Allison RS (eds): Violent Juvenile Offenders: An Anthology. San Francisco, CA,
National Council on Crime and Delinquency, 1984, pp 83–112.
35. Tinklenberg JR, Murphy P, Murphy PL, et al: Drugs and criminal assaults by adolescents: A
replication study. J Psychoactive Drugs 13:277–287, 1981.
36. Fagan J, Weis J, Cheng YT, et al: Drug and alcohol use, violent delinquency, and social
bonding: Implications for theory and intervention. Washington, DC, National Institute of
Justice Drug and Crime Research Program, 1987.
37. Fagan J, Weis JG, Cheng Y-T: Delinquency and substance use among inner city students.
New York, New York City Criminal Justice Agency, 1988.
38. Kingery PM, Pruitt BE, Hurley RS: Violence and illegal drug use among adolescents: Evi-
dence from the US National Adolescent Student Health Survey. Int J Addict 27:1445–1464,
1992.
39. Watts WD, Wright LS: The relationship of alcohol, tobacco, marijuana, and other illegal drug
use to delinquency among Mexican-American, black, and white adolescent males. Adolescence
25:171–181, 1990.
40. Collins JJ: The relationship of problem drinking to individual offending sequences, in Blum-
stein A, Cohen J, Roth J, et al (eds): Criminal Careers and “Career Criminals.” Washington DC,
National Academic Press, 1986, vol 2, pp 89–120.
41. Loeber R: Natural histories of conduct problems, delinquency, and associated substance use,
in Lahey BB, Kazdin AE (eds): Advances in Clinical Child Psychology, New York, Plenum, 1988,
vol 11, pp 73–124.
42. Chaiken J, Chaiken M: Drugs and predatory crime, in Tonry M, Wilson JQ (eds): Drugs and
Crime. Chicago, The University of Chicago Press, 1990, vol 13, pp 203–239.
43. Robins LN: The adult development of the antisocial child. Semin Psychiatry 2:420–434, 1970.
44. Zucker RA: The concept of risk and the etiology of alcoholism: A probablistic–developmental
perspective, in Pittman DJ, White HR (eds): Society, Culture and Drinking Patterns Reexamined.
New Brunswick, NJ, Rutgers Center of Alcohol Studies, 1991, pp 513–532.
45. White HR: Early problem behavior and later drug problems. J Res Crime Delinq 29:412–429,
1992.
46. McCord JF: Alcohol in the service of aggression, in Gottheil E, Druley KA, Skoloda TE, et al
(eds): Alcohol, Drug Abuse and Aggression. Springfield, IL, Charles C. Thomas, 1983, pp 270–
279.
47. White HR, Hansel1 S: The moderating effects of gender and hostility on the alcohol-aggres-
sion relationship. J Res Crime Delinq 33:451–472, 1996.
48. Huesmann LR, Eron LD, Lefkowitz MM, et al: Stability of aggression over time and genera-
tions. Dev Psychol 20:1120–1134, 1984.
49. Olweus D: The stability of aggressive reaction patterns in human males: A review. Psychol
Bull 85:852–875, 1979.
50. Farrington DP: The development of offending and antisocial behavior from childhood: Key
findings from the Cambridge study in delinquent development. J Child Psychol Psychiatry
36:1–36, 1995.
51. Fillmore KM: Drinking and problem drinking in early adulthood and middle age: An explora-
tory 20-year follow-up study. Q J Stud Alcohol 35:819–840, 1974.
52. Virkkunen M: Arrests for drunkenness and recidivism in juvenile delinquents. Br J Addict
72:201–204, 1977.
53. Dembo R, Williams L, Getreu A, et al: A longitudinal study of the relationships among
marijuana/hashish use, cocaine use, and delinquency in a cohort of high risk youths. J Drug
Issues21:271–312, 1991.
54. Miller BA: Investigating links between childhood victimization and alcohol problems, in
Martin SE (ed): Alcohol and Interpersonal Violence: Fostering Multidisciplinary Perspectives.
Rockville, MD, National Institute of Health, 1993, NIAAA Research Monograph No. 24, pp
3–36.
55. Widom CS: Child abuse and alcohol use and abuse, in Martin SE (ed): Alcohol and Interpersonal
Violence: Fostering Multidisciplinary Perspectives. Rockville, MD, National Institute of Health,
1993, NIAAA Research Monograph No. 24, pp 291–314.
56. Thornberry TP: Violent families and youth violence. Office of Juvenile Justice and Delinquen-
cy Prevention, US Department of Justice, December, 1994.
57. Widom CS: The cycle of violence. Science 244:160–166, 1989.
58. Johnson JL, Perez-Bouchard L, Rebeta JL: Parental reports of behavior problems and stress in
sons of alcoholics. Alcohol Clin Exp Res 18:472, 1994.
59. Jessor R, Jessor S: Problem Behavior and Psychosocial Development—A Longitudinal Study of
Youth. New York, Academic Press, 1977.
60. White HR, Labouvie EW: Generality versus specificity of problem behavior: Psychological
and functional differences. J Drug Issues 24:55–74, 1994.
61. Gottfredson MR, Hirschi T: A General Theory of Crime. Standford, CA Standford University
Press, 1990.
62. White HR, Pandina RJ, LaGrange RL: Longitudinal predictors of serious substance use and
delinquency. Criminology 25:715–740, 1987.
63. White HR Marijuana use and delinquency: A test of the "independent cause" hypothesis. J
Drug Issues 21:231–256, 1991.
64. Osgood DW: Covariation among adolescent problem behaviors. Presented at American Soci-
ety of Criminology annual meeting, Baltimore, MD, 1990.
65. Johnson BD, Wish E, Schmeidler J, et al: The concentration of delinquent offending; Serious
drug involvement and high delinquency rates, in Johnson BD, Wish E (eds): Crime Rates
among Drug Abusing Offenders. New York, Interdisciplinary Research Center, Narcotic and
Drug Research, 1986, pp 106–143.
66. Hawkins JD, Catalano RF, Miller JY: Risk and protective factors for alcohol and other drug
problems in adolescence and early adulthood: Implications for substance abuse prevention.
Psychol Bull 112:64–105, 1992.
67. Kandel D: Drug and drinking behavior among youth, in Inkeles A, Smelser NJ, Turner RH
(eds): Annual Review of Sociology. Palo Alto, CA, Annual Reviews, 1980, vol 6, pp 235–285.
68. Tolan P, Guerra N: What Works in Reducing Adolescent Violence: An Empirical Review of the Field.
Boulder, CO, Center for the Study and Prevention of Violence, 1994.
69. Tremblay RE, Craig WM: Developmental crime prevention, in Tonry M, Farrington DP (eds):
Building a Safer Society: Strategic Approaches to Crime Prevention. Chicago, The University of
Chicago Press, 1995, vol 19, pp 151-236.
70. Brewer DD, Hawkins JD, Catalano RF, et al: Preventing serious, violent and chronic juvenile
offending: A review of evaluations of selected strategies in childhood, adolescence and the
community, in Howell JC, Krisberg B, Hawkins JD, et al (eds): Sourcebook on Serious, Violent,
and Chronic juvenile Offenders. Thousand Oaks, CA, Sage, in press, 1995, pp 61–141.
71. Hawkins JD, Catalano RF, Brewer DD: Preventing serious, violent and chronic delinquency
and crime: Effective strategies from conception to age six, in Howell JC, Krisberg B, Hawkins
JD, et al (eds): Sourcebook on Serious, Violent, and Chronic juvenile Offenders. Thousand Oaks,
CA, Sage, 1995, pp 47–60.
72. Hirschi T: Causes of Delinquency. Berkeley, University of California Press, 1969.
73. Sutherland EH, Cressey DR: Criminology. Philadelphia, Lippincott, 1978.
74. Elliott DS, Huizinga DH, Ageton SS: Explaining Delinquency and Drug Use. Beverly Hills, CA,
Sage, 1985.
75. Bandura A: Aggression: A Social Learning Analysis. Englewood Cliffs, NJ, Prentice-Hall, 1973.
76. White HR, Bates ME, Johnson V: Social reinforcement and alcohol consumption, in Cox WM
(ed): Why People Drink: Parameters of Alcohol as a Reinforcer. New York, Gardner Press, 1990, pp
233–261.
77. Abrams DB, Niaura RS: Social learning theory, in Blane HT, Leonard KE (eds): Psychological
Theories of Drinking and Alcoholism. New York, Guilford Press, 1987, pp 131–178.
78. Bandura A: Social Learning Theory. Englewood Cliffs, NJ, Prentice-Hall, 1977.
79. Hawkins JD, Arthur MW, Catalano RF: preventing substance abuse, in Tonry M, Farrington
DP (eds): Building a Safer Society: Strategic Approaches to Crime Prevention. Chicago, The Univer-
sity of Chicago Press, 1995, vol 19, pp 343–427.
80. Gorman DM: Do school-based social skills training programs prevent alcohol use among
young people? Addict Res, 4:191–210, 1996.
81. Gorman DM, Speer PW: Preventing alcohol abuse and alcohol-related problems through
community interventions: A review of evaluation studies. Psychol Health 11:95–131, 1995.
82. Moffitt TE: Adolescence-limited and life-course-persistent antisocial behavior: A develop-
mental taxonomy. Psychol Rev 100:674–701, 1993.
5
Alcohol and Violence-Related

Injuries in the Emergency Room
Cheryl J. Cherpitel
Abstract. This chapter reviews data on estimated blood alcohol concentration (BAC), self-re-
ported consumption, and drinking patterns and problems from emergency room (ER) studies of
alcohol and violence-related injury. These studies used probability samples of all injured patients
that were representative of the population served by the ER where the data were collected. Those
with violence-related injuries were more likely to be admitted to the ER with a positive BAC, to
report drinking prior to the event, and to report more frequent heavy drinking and alcohol-
related problems than those admitted to the same ER during the same time period with injuries
from other causes. Limitations to these ER studies, including representativeness of samples,
alcohol's presence and role in violence perpetration compared to violence victimization, the
presence of other psychoactive substances, and the actual risk at which alcohol places the indi-
vidual for injuries resulting from violence are discussed.
1. Introduction
Alcohol consumption has been found to be associated with both fatal and
nonfatal injuries resulting from violence.1–3 Data documenting alcohol's pres-
ence in nonfatal injuries have come primarily from studies of patients admit-
ted to hospital emergency rooms (ERs) for treatment. Emergency rooms are
the primary source of treatment for all types of injuries including those re-
lated to violence. Injuries resulting from violence that are serious enough to
require emergency room treatment are thought to account for over 20 times as
many cases as violence-related (nonsuicide) fatal injuries.4 While studies in
ERs provide more information on victims of violence than on perpetrators
Cheryl J. Cherpitel • Alcohol Research Group, Western Consortium for Public Health, Berkeley,
California 94709
105
(unless the perpetrator also sustains injuries requiring treatment), these stud-
ies can provide data on the estimated prevalence of alcohol’s presence in
events involving violence, drinking characteristics of those involved, and
other information on alcohol’s role in such events. This chapter will review
findings from ER studies of violence-related injuries in relation to prevalence
estimates of alcohol’s involvement in such injuries, as well as alcohol’s role in
the event and drinking patterns and problems of those with injuries resulting
from violence. The present work updates and expands upon an earlier review
article on the same topic.5
Alcohol is thought to be positively associated with severity of injury,
including injuries resulting from violence,6 although findings have been
found to vary considerably across studies. One study found 29% of those
with severe injuries related to violence reported drinking more than 10 drinks
in the 12 hr prior to the event, compared to 18% of those with minor injuries.7
Additionally, 30% of those with multiple injuries reported drinking at this
level compared to 19% of those with single injuries. These authors suggest
that loss of judgment induced by alcohol may prolong the violent event or the
victim may be less able to avoid the encounter, both of which would result in
more serious injury. To further explore alcohol’s association with severity of
violence-related injury, data will be compared on alcohol’s presence from
representative samples of fatal and nonfatal injuries resulting from violence,
both occurring in the same county during the same time from two regions of
the United States where per capita consumption of alcohol differ greatly—the
South and the West. These data will provide information on alcohol’s involve-
ment in violence-related events resulting in injury in relation to severity of
injury and typical drinking patterns that may influence the occurrence of such
events.
2. Prevalence Estimates of Blood Alcohol Concentration

Emergency room studies that are the most informative regarding alco-
hol’s presence in violence-related injuries are those which have obtained a
measure of estimated blood alcohol level (BAC) from blood, breath, saliva, or
urine soon after arrival in the ER. A number of these studies have included
representative samples of patients both with and without injuries related to
violence admitted to the ER during the same period of time (see Table I).
Among those studies in Table I that meet these criteria, the proportion of
patients with violence-related injuries who had positive BACs ranged from
17% (in a health maintenance organization) to 70% (in a rural district general
hospital in Scotland). While a large amount of variation exists, both in the
prevalence of positive BACs and in the proportion of those legally intoxicated
(BACs of 0.08 or higher), across all of these studies, those with violence-
related injuries were from two to five times more likely to be positive for
alcohol (and to be intoxicated) at the time of the ER visit than those with other
injuries. This also has been found to be true when age and gender are con-
5 • Alcohol and Violence-Related Injuries 107
Table I. ER Studies Measuring BACa Among Probability Samples of Patients
Alcohol Violent injuries: Nonviolent injuries:

Study (year, location) Age measure % pos BAC (N) % pos BAC (N)
1. Wechsler et al (1967)8 ≥16 Breath 56 (188) 22f (2701)

(US—MA) 39 ≥ . 05 11f ≥ .05
2. Peppiatt et al. ≥15 Blood 45 (29) 10f (671)
(1976/77)9 (UK)
3. Walsh & Macleod ≥15 Breath 70 (56) 14 f (698)
(1980)10 (Scotland) 50 ≥ .08 8f ≥ .08
4. Papoz et al ≥15 Blood within 64 (544) 32 f (4245)
(1982/83)11 (France) 6 hr 50 ≥ .08 18 f
5. Yates et al 12 (UK) ≥16 Blood or 60 (59) 14f (946)
breath 9 ≥ .08 3
6. Cherpitel (1984/85)13 ≥18 Breath 44 (57) 16 f (148)
(US-CA) within 6 hr 30 ≥ .10 11f ≥ .10
7. Cherpitel (1985)b ≥18 Breath 26 (102) 9 f (764)
(US-CA) within 6 hr 12 ≥ .10 3 ≥ .10
8. Cherpitel (1986)c ≥18 Breath 38 (438) 15f (1113)
(Mexico) within 6 hr 21 ≥ .10 8f ≥ .10
9. Cherpitel (1986/87)16 ≥18 Breath 25 (142) 14f (591)
(US-CA) within 6 hr 14 ≥ .10 6f
10. Cherpitel (1987/88)d ≥18 Breath 25 (81) 10f (1520)
(Spain) within 6 hr 16 ≥ .10 4 f ≥ .10
11. Cherpitel (1989)e ≥18 Breath 17 (12) 7 (355)
(US-CA) within 6 hr 0 ≥ .10 2 ≥ .10
12. Cherpitel (1992)19 ≥18 Breath 35 (54) 7f (210)
(US-MS) within 6 hr 11 ≥ .10 5
a BAC is recorded as mg %: Positive is ≥0.01 (10 mg. of alcohol per 100 ml of blood). .05, .08, .10 are BACs as
indicated.
b Reanalysis of data from ref. 14.
cReanalysis of data from ref. 15.
dReanalysis of data from ref. 17.
eReanalysis of data from ref. 18.
f P < 0.05 Comparison of positive breathalyzer readings between injured and noninjured in the same sample
using tests of significant difference between proportions.
trolled. One study that compared violence-related injuries with all other inju-
ries, separately for males and females under 30 and those 30 and older, found
those with injuries resulting from violence were significantly more likely to be
heavy problem drinkers than those with other injuries in each of the gender–
age groups.20 A study of representative samples of ER patients in Acapulco,
Mexico found those with positive breathalyzer readings less than 0.10 were
almost 13 times more likely than those with negative readings to have injuries
related to violence as opposed to injuries resulting from animal bites or work-
place or recreational accidents, when age, gender, other demographic charac-
teristics, and time of occurrence (weekday vs. weekend) were controlled.21
Surprisingly, those with breathalyzer readings over 0.10, however, were no
more likely to have sustained injuries resulting from violence.
Other ER studies of nonrepresentative samples of patients among whom
one might expect a larger proportion to be alcohol positive, e.g., those admit-
ted to the ER on weekend evenings22 or those so seriously injured as to
require blood typing,23 have generally found higher rates of alcohol's pres-
ence among both those with and without violence-related injuries (86 vs. 46%
and 71 vs. 33%, respectively) than studies of representative samples of pa-
tients.
These data suggest that while a substantial association appears to exist
between alcohol consumption and violence-related injuries treated in the ER,
the estimated prevalence of positive BACs at the time of arrival in the ER
varies considerably, even among representative samples of patients. Mea-
sures of both blood and breath were used to obtain BAC estimates; however,
this would not be expected to affect prevalence rates reported, since breath
analysis for alcohol has been found to be highly correlated with chemical
analysis of blood among cooperative patients (r = .96).24 As seen in Table I,
however, the age criteria for inclusion in samples of patients varied across
studies, from 15 and older to 18 and older, which may have influenced report-
ed prevalence estimates. The length of data collection in each site, which
varied from 1 week in each of eight ERs in Mexico City (Study 8)25 to 1 year in
one ER in California (Study 9),16 may also have influenced prevalence rates,
since a shorter length of data collection would have a higher probability of
being less representative of all patients attending the ER than a longer period
of data collection. The degree of ascertainment for obtaining BAC estimates
may also affect prevalence rates, particularly if patient refusal or severity of
injury are related to drinking in the event (for both those with violence-
related injuries and those with other injuries). While data were not available
on injury severity in these studies, completion rates for obtaining an esti-
mated BAC ranged from 99% (Study 3)10 to 72% (Study 9)16 (no ascertainment
rate was reported for Study 4). One ER study found that among those injured
who were not interviewed because of refusal, severity of injury, or for other
reasons, but on whom a BAC estimate was obtained, a larger proportion
tested positive than those who did participate in the study.26
Another factor that could influence the proportion of positive BACs
found is the time that elapsed between the injury and admission to the ER, as
well as the time that elapsed between admission to the ER and when the BAC
estimate was obtained. Although a BAC estimate was obtained within 6 hr of
ER admission for many of the studies listed in Table I, the length of time
between the event that resulted in injury and arrival in the ER is not known.
Blood alcohol concentration has been found to be negatively correlated with
the length of time between injury and arrival in the ER among all of those
seeking treatment for injuries, with those arriving at the ER within 3 hr of
injury over twice as likely to be alcohol positive as those arriving more than 6
hr after the event, when drinking after the event was controlled.8 Another
study, however, found no association between the likelihood of being alcohol
positive and the time elapsed between injury and arrival in the ER when
drinking after the event was controlled.13
The prevalence of alcohol-positive cases in an ER (across all types of

injury) has also been found to vary by the type of ER facility and the associ-
ated sociodemographic characteristics of the clientele served by the facility.27
For example, in Table I, the US studies in which the largest proportion of
alcohol-positive violence-related injuries was found were those carried out in
large urban level 1 trauma centers, where considerable proportions of those
attending the ER were indigent and heavier drinkers (Studies 1, 2, and 6). On
the other hand, the smallest proportion of alcohol-positive violence-related
injuries was found among those attending suburban county and community
hospital ERs (Studies 7 and 9) and a health maintenance organization (HMO)
emergency room (Study 11).
3. Self-Reports of Alcohol Consumption Prior to Injury

The last seven studies in Table I also obtained data on self-reported drink-
ing within 6 hr prior to the injury event (see Table II). In each of the studies in
Table II except one, the proportion of those with violence-related injuries was
significantly more likely to report drinking prior to the event than those with
other injuries. In the one study that did not find a significant difference
(perhaps due to the small number of patients with violence-related injuries)
those with injuries resulting from violence were still three to four times more
likely to report drinking prior to the event than those with other injuries.
Additional analyses of data from a combined California sample of patients
from Studies 2 and 4 in Table II found that among those with violence-related
injuries who reported drinking within the 6 hr prior to injury, 35% reported
consuming more than six drinks during this time (compared to 23% of those
with other injuries), and well over half (67%) reported less than an hour
Table II. Self-Reported Consumption Prior to Injury
Violent injuries: Nonviolent injuries:

Year of study % positive BAC (N) % positive BAC (N)
1. 1984/8513 55 (158) 27e (397)

2. 1985a 52 (116) 14e (826)
3. 1986b 52 (448) 18e (1136)
4. 1986/8716 49 (160) 22e (622)
5. 1987/88c 36 (87) 0e (1327)
6. 1989d 27 (15) 8 (361)
7. 199219 56 (57) 14e (214)
a Reanalysis of data from ref. 14.
b Reanalysis of data from ref. 15.
c Reanalysis of data from ref. 17.
d Reanalysis of data from ref. 18.
e P < 0.05 Comparison of positive breathalyzer readings between injured and non-
injured in the same sample using tests of significant difference between proportions.
Table III. Drinking Patterns and Problems among Drinkers (in percent)
Contra Costa County,

San Francisco, CAa CA19 Jackson, MS19
Violence No violence Violence No violence Violence No violence

(141) (337) (84) (1007) 48) (139)
Heavy drinkers 41 25b 22 15 17 12

Drunk > 1 time 37 24b 20 13 27 9b
per month
2 or more con- 34 18b 11 7 8 3
sequences
3 or more 35 21b 13 6 17 9
dependence
experiences
a
Reanalysis of data from ref. 14.
b P < 0.05 Comparison of positive breathalyzer readings between injured and noninjured in the same sample
using tests of significant difference between proportions.
lapsed between their last drink and the event.20 Of those who reported drink-
ing, 45% reported feeling drunk at the time of injury, and almost half of these
(46%) believed the event would not have happened if they had not been
drinking. Similar analyses of data from Mississippi (Table I, Study 12) found
32% of those with violence-related injuries who reported drinking during the
6 hr prior to the event consumed more than six drinks during this time
(compared to 11% of those with other injuries), and three fourths reported
less than an hour between the last drink and the event.19 A third of those with
violence-related injuries who had been drinking reported feeling drunk, and
of these, 60% believed the event would not have happened had they not been
drinking.
Also of interest in Table II is the fact that a larger proportion of both those
with and those without violence-related injuries reported drinking within 6 hr
prior to the event than were positive for estimated blood alcohol at the time of
ER arrival (see Table I). Several of these same studies have compared the
concordance of positive estimates for blood alcohol and self-reported con-
sumption prior to the event for those who arrived at the ER within 6 hr of
injury. While relatively small proportions of those who deny drinking prior to
injury register positive for estimated blood alcohol, ranging from 0.5%28 to
3.6%13 in US studies across all causes of injury, substantial proportions of
those who report drinking prior to the event have been found to register
negative for estimated blood alcohol. A comparison of violence-related inju-
ries among those from whom a BAC estimate was obtained within 6 hr of
injury and who reported no drinking following the event in comparable stud-
ies carried out in United States, Mexico, and Spain found that while no pa-
tients who denied drinking were positive for estimated BAC in the US sam-
ple, 4% in Mexico and 12% in Spain denied drinking while registering
positive.28 However, 42% of those in the United States, 13% in Mexico, and
27% in Spain reported drinking prior to the event involving violence while
registering negative for estimated BAC. The validity of self-reports has been
found to be high even when BAC estimates have been obtained after self-
reported consumption was obtained.29 While self-reports of alcohol use have
generally been found to be reasonably valid when compared to objective
measures, some variation in accuracy has been found, leading to both under-
reporting and overreporting.30–32 The social desirability of responses has been
found to affect validity of self-report drinking33; thus, differences in self-
reported consumption between those with and those without injuries result-
ing from violence may be conservative. Interestingly, in the Spain sample
(mentioned above), those with violence-related injuries were more likely than
those with injuries from falls or vehicular or other accidents to deny drinking
while having a positive BAC, while in the Mexico and US samples, those with
violence-related injuries were no more likely to deny drinking than those with
injuries from other causes.
4. Drinking Patterns and Alcohol-Related Problems

Findings from ER studies have also suggested that those with injuries
related to violence are more likely to be frequent heavy drinkers with more
alcohol-related problems compared to those with other injuries as well as to
those in the general population from which these patients come.12,20,34,35 One
study that compared representative samples of patients seen in the same
emergency rooms during the same period of time found those with violence-
related injuries twice as likely (30 vs. 15%) to report frequent heavy drinking
(drinking at least once a week and reporting five or more drinks at a sitting at
least weekly).20 These patients were also found to report more frequent
drunkenness, social consequences of drinking, alcohol-dependence experi-
Table IV. BACa in Fatal Compared to Nonfatal

Violence-Related Injuries
Contra Costa County Hinds County
ER Coroner ER Coroner
BAC (47) (38) (46) (37)
Positive 19 47b 37 54
≥.10 15 34b 11 27
a
BAC is recorded as mg %: Positive is ≥0.01 (10 mg of alcohol per 100 ml of blood).
b P < 0.05 Comparison of proportions between those breathalyzed with 6 hr of ER
admission and reported no drinking after the event and blood toxicology screen for
those whose blood was drawn or who died within 6 hr of injury, using tests of
significant difference between proportions.
ences, and treatment for alcohol problems than those with injuries not result-
ing from violence, and these differences were found within gender-by-age
(less than 30 and 30 and older) categories.
Table III shows the proportion who report heavy drinking, at least
monthly drunkenness, two or more social consequences of drinking, and
three or more dependence experiences among drinkers with and without
violence-related injuries from three ER studies, which, using similar meth-
ods, obtained representative samples of all patients admitted to the ER. In
this table, heavy drinking is defined as reporting drinking at least three times
per week with 12 or more drinks on at least one occasion during the last year.
Social consequences included reporting any of the following problems related
to drinking during the last year: problems with personal relationships, work,
police or other authorities, physical health, psychological health or mental
well-being. Experiences related to alcohol dependence included blackouts,
relief drinking, hands shaking alot the morning after drinking, binge drink-
ing, and feeling that one should cut down on his or her drinking or quit
altogether.
As seen in Table III, those with violence-related injuries were more likely
to report heavy drinking, more frequent drunkenness, social consequences of
drinking, and alcohol-dependence experiences than those with other injuries
within the same ER sample, although the proportion of those reporting heavy
and problem drinking varies considerably across the three studies. While
differences between those with and without injuries resulting from violence
were statistically significant only for the San Francisco sample, differences
were substantial in the other two samples, although not significant, possibly
due to the relatively small numbers of those with violence-related injuries in
these two samples. A comparison of those in the Contra Costa sample (which
is a merged sample of the county hospital, three of the six community hospi-
tals, and the three HMO ER, all of which were weighted to be representative
of those admitted to the ER in the entire county) to a general population
sample of the same county found both those with and without injuries re-
sulting from violence significantly more likely to report heavy and problem
drinking36
Because high rates of frequent heavy drinking and alcohol-related prob-
lems have been found among those with violence-related injuries in ER stud-
ies, it has been suggested that presenting to the ER with injuries resulting
from violence may be a sign of excessive alcohol consumption, particularly
among males over 25.35
Although, as a group, those with violence-related injuries in the ER have
high rates of heavy and problem drinking, estimated BAC at the time of ER
admission often fails to identify most problem drinkers across all injury
causes, ranging from 1637 to 31%.12 One study found that while two thirds of
injured patients who were BAC positive met criteria for alcohol dependence,
almost half of those who were BAC negative also met criteria for alcohol
dependence.38
5. Regional Comparisons of ER and Coroner Data

While alcohol is thought to be positively associated with severity of inju-
ry, including those resulting from violence,7 few data are available that have
compared representative samples of both fatal and nonfatal injuries resulting
from violence occurring in the same geographic locality during the same
period of time. Additionally, alcohol's association with both fatal and nonfatal
injury may be expected to vary from one region of the country to another as
regional drinking patterns vary.39 Table IV shows the proportion of those with
nonfatal (ER sample) and fatal (coroner sample) injuries who were alcohol
positive and intoxicated for two diverse regions of the United States: Contra
Costa County, California16 and Hinds County, Mississippi.40 Emergency
room and coroner data in each region included representative samples and
covered the same geographic area during the same period of time.
Similar data on fatal and nonfatal injuries from two geographically di-
verse regions of the United States are of interest because national alcohol
surveys consistently have found distinct regional differences in drinking pat-
terns and associated problems. The “dryer” regions of the United States
(those which have traditionally supported a strong pro-temperance senti-
ment), particularly the South, have been found to report higher rates of ab-
stention but also higher rates of frequent heavy drinking among those who do
drink and more problems related to drinking (e.g., alcohol-related aggres-
sion) at a given level of consumption compared to the “wetter” regions of the
country.39,41 Mississippi has had the strongest pro-temperance sentiment of
any of the Southern states, as evidenced by their being the last state to repeal
prohibition (in 1966) and allow for local county option for selling alcoholic
beverages. Presently, 44% of Mississippi's counties are dry (representing a
25% increase since 1988) compared to an average of 15% for the four sur-
rounding states.42 Mississippi reports a lower per capita consumption, based
on gallons of absolute alcohol per person aged 14 years and older (2.09) than
the Southern region as a whole (2.28), which is lower than the United States
(2.28) and California (2.55).43
As seen in Table IV, those with violence-related fatalities were more likely
to be alcohol positive and to be legally intoxicated than those with nonfatal
injuries in both samples; however, differences were statistically significant
only in the Contra Costa County sample. Both those with fatal and nonfatal
injuries were more likely to be alcohol positive, but less likely to be intoxicated
in the Mississippi sample compared to the California sample. These data
suggest that, while alcohol may have a greater association with violence-
related fatal than nonfatal injuries, the difference may be less pronounced in
dryer regions of the country. The data also suggest that alcohol may have a
greater association with both fatal and nonfatal violence-related injuries in
dryer areas than in wetter areas, but not at high levels of consumption. While
various explanations have been offered for the reasons for regional differ-
ences in drinking patterns and problems in the United States,44,45 the possi-
bility that areas of the United States that have relatively low rates of per capita
consumption have higher rates of alcohol-related injuries resulting from vio-
lence than areas with greater per capita consumption presents an interesting
paradox in relation to prevention of alcohol-related violent injuries. This is in
contrast to the positive temporal association that has been found between per
capita consumption and aggregate-level violence in a number of Scandinavian
countries.46 It is important to note, however, that the counties compared here
are not necessarily representative of the larger geographic area from which
they come. As seen in Table I, a great deal of variation exists in the proportion
of ER patients who are BAC positive, even in adjacent geographic areas (e.g.,
Studies 7 and 11), and this may also be true of coroner cases from adjacent
counties.
6. Limitations to ER Studies
While this review of alcohol and violence-related injuries in the ER
strongly supports the supposition that alcohol is related to such injuries
treated in the ER, there are several limitations to ER studies that affect their
usefulness for understanding alcohol's presence and role in injuries resulting
from violence. Even given representative samples of patients treated in the
ER, many victims of violence may obtain treatment from other sources or may
not have injuries severe enough to require any treatment. Comparisons of ER
data with data from the general population from which these patients come,
however, have shown that, while injured ER patients are more likely to be
frequent heavy and problem drinkers as compared to the general population
as a whole, they are similar in demographic and drinking characteristics to
those in the general population who have used the ER for treatment for an
injury during the last year.20,36 The greatest difference between the two
groups is the frequency of ER use, as might be expected, since those sampled
in an ER would tend to be more frequent users of the ER than those sampled
elsewhere. While factors associated with frequency of ER use may differ be-
tween these two groups, comparative findings between ER samples and the
general population are reassuring in relation to the representativeness of ER
samples to their counterparts in the general population. Another potential
bias in ER data is that of misclassification, in which an unknown proportion of
those with injuries related to violence may go unrecognized or be deliberately
misclassified by either the patient or ER staff because of the stigma that may
be associated with such injuries.
As mentioned earlier, ER studies, by their nature, are limited to victims of
violence, and little may be learned from these studies about the perpetrator of
the violent event unless the perpetrator is also injured and obtains ER treat-
ment. A review of both ER patients as well as those known to the criminal
justice system in Great Britain found a substantial association of alcohol with
both violence perpetration and violence victimization, with both increasing as
alcohol consumption increased.3 Additionally, alcohol's involvement may

vary by the nature of the event, e.g., whether family members or nonfamily
members are involved. One study of perpetration of violence-related events
during the preceding 3 months among patients admitted to a psychiatric ER
found those who reported such events were over twice as likely to report
alcohol abuse when the event involved nonfamily members as when the
event involved family members (45 vs. 22%).47
While estimated BAC at the time of admission to the ER provides quan-
titative data on alcohol consumption prior to the ER visit, this may actually
have little to do with alcohol consumption prior to or at the time of the
violence-related injury event, depending on the length of time between the
event and arrival at the ER, as well as drinking after the injury event. One
study found that only 56% of injury patients reported to the ER within 6 hr of
the event.26 Therefore, a positive BAC at the time of ER admission is not
necessarily an indication that a patient had been drinking prior to the violent
event, and conversely a negative BAC at the time of the ER visit does not
mean the patient had not been drinking. Although ER studies that obtain an
estimate of BAC may document the presence of alcohol at the time of the
event, the actual role that alcohol may have played in the event is not known,
i.e., whether alcohol was related to or caused the event, or whether alcohol
use was just coincidental to the event. While alcohol is thought to act as a
social disinhibitor48 leading to aggressive behavior in certain individuals,39
and alcohol is also known to affect motor coordination and reaction time and
to alter judgment—all of which could result in injury—the actual causative
mechanism for the association of alcohol and violence-related injury is not
known. Further, alcohol's role in perpetrating a violence-related event may be
different than alcohol's role in becoming the victim of such an event. Addi-
tionally, many individuals who drink also use psychoactive drugs, which
makes determining the independent role of alcohol in the event more diffi-
cult. In the Contra Costa County study in Table IV, of those in the ER sample
who were positive for alcohol, 31% were also positive for drugs, while 39% of
those in the coroner sample who were alcohol positive were also drug posi-
tive. It is also possible that associations of alcohol consumption and violence
may be the result of their co-occurrence in relation to time of day and day of
week. Drinking patterns have been found to vary in the United States, with
higher rates of heavier drinking occurring during weekend evening
hours.49,50 If violent-related events are also more likely to occur during week-
end evenings than at other times, the association of alcohol consumption and
violence may be the result of a temporal association rather than a causal
association. One ER study found, however, that regardless of the time of
admission to the ER, those with violence-related injuries were more likely to
be positive for estimated BAC and to report drinking prior to the event than
those with other injuries.20
While this review of ER studies suggests that alcohol may be a major risk
factor for injuries resulting from violence, the actual risk at which various
levels of drinking, or any drinking, places the individual for violence-related

injuries cannot be determine from these studies. One study of alcohol and
injury in the general population suggests that an individual's risk for injury
may increase at relatively low levels of drinking.51 Emergency room studies
that have found those with violence-related injuries more likely to be fre-
quent, heavy, and problem drinkers compared with those with other injuries
may actually underestimate alcohol's involvement in violence, since those
injured from other causes have also been found to report higher rates of heavy
and problem drinking than the noninjured as well as those in the general
population. We are also not able to determine the proportion of violence
resulting in injury that can be attributed to alcohol consumption, since we do
not know whether the perpetrator of the event had also been drinking at the
time. Only studies of alcohol and injury in representative samples from the
general population can begin to provide data to address these concerns. Nev-
ertheless, ER studies do provide a useful window for observing alcohol's
association with injuries that are treated in this setting, and they also provide
an important arena for the identification of those with alcohol-related prob-
lems who could benefit from an intervention or referral for problem drinking.
ACKNOWLEDGMENTS. This work was supported by the US National Institute

of Alcohol Abuse and Alcoholism Alcohol Research Center Grant (AA 05595).
References
1. Roizen J: Estimating alcohol involvement in serious events, in: Alcohol Consumption and Related
Problems. Washington, DC, US National Institute on Alcohol Abuse and Alcoholism, 1982, pp
179–219.
2. Roizen J: Alcohol and trauma, in Giesbrecht N, Gonzales R, Grant M, et al (eds): Drinking and
Casualties: Accidents, Poisonings and Violence in an International Perspective. London, Tavi-
stock/Helm, 1989, vol 1, pp 21–66.
3. Norton RN, Morgan MY: The role of alcohol in mortality and morbidity from interpersonal
violence. Alcohol Alcoholism 24:565–576, 1989.
4. Barancik J, Chatterjee B, Greene YC, et al: Northeastern Ohio Trauma Study: 1. Magnitude of
the problem. Am J Public Health 73:746–751, 1983.
5. Cherpitel CJ: Alcohol and injuries resulting from violence: A review of emergency room
studies. Addiction 89:157–165, 1994.
6. Honkanen R, Smith GS: Impact of acute alcohol intoxication on the severity of injury: A
cause-specific analysis of non-fatal trauma. lnjury 21:353–357, 1990.
7. Shepherd J, Irish M, Scully C, Leslie I: Alcohol intoxication and severity of injury in victims of
assault. Br Med J 296:1299, 1988.
8. Wechsler H, Kasey E, Thum D, Demone H: Alcohol level and home accidents. Public Health
Rep 84:1043–1050, 1969.
9. Peppiatt R, Evans R, Jordan P: Blood alcohol concentrations of patients attending an accident
and emergency department. Resuscitation 6:37–43, 1978.
10. Walsh ME, Macleod D: Breath alcohol analysis in the accident and emergency department.
lnjury 15:62–66, 1983.
11. Papoz L, Weill J, Gate C, et al: Biological markers of alcohol intake among 4,796 subjects
injured in accidents. Br Med J 292:1234–1237, 1986.
12. Yates D, Hadfield J, Peters K: The detection of problem drinkers in the accident and emergen-
cy departments. Br J Addict 82:163–167, 1987.
13. Cherpitel CJ: Breathalyzer and self-reports as measures of alcohol-related emergency room
admission. J Stud Alcohol 50:155–161, 1989.
14. Cherpitel CJ: Alcohol consumption and casualties: A comparison of two emergency room
populations. Br J Addict 83:1299–1307, 1988.
15. Cherpitel CJ, Rosovsky H: Alcohol consumption and casualties: A comparison of emergency
room populations in the United States and Mexico. J Stud Alcohol 51:319–326, 1990.
16. Cherpitel CJ: Alcohol and casualties: A comparison of emergency room and coroner data.
Alcohol Alcoholism 29:211–218, 1994.
17. Cherpitel CJ, Pares A, Rodés J: Drinking patterns and problems: A comparison of emergency
room populations in the United States and Spain. Drug Alcohol Depend 29:5–15, 1991.
18. Cherpitel CJ: Alcohol consumption among emergency room patients: Comparison of coun-
ty/community hospitals and an HMO. J Stud Alcohol 54:432–440, 1993.
19. Cherpitel CJ: Alcohol and Injuries Resulting from Violence: A Comparison of Emergency Room
Samples from Two Regions of the US. J Addict Dis, in press.
20. Cherpitel CJ: Alcohol and violence-related injuries: An emergency room study. Addiction
88:79–88, 1993.
21. Borges G, Garcia G, Gill A, Vandale S: Casualties in Acapulco: Results of a study of alcohol
use and emergency room care. Drug Alcohol Depend 36:1–7, 1994.
22. Holt S, Stuart I, Dixon J, et al: Alcohol and the emergency service patient. Br Med J 281:638–
640, 1980.
23. Honkanen R, Visuri T: Blood alcohol levels in a series of injured patients with special refer-
ence to accident and type of injury. Ann Chir Gynaecol 65:287–294, 1976.
24. Gibb K, Yee A, Johnson C, et al: Accuracy and usefulness of a breath alcohol analyzer. Ann
Emerg Med 13:516–520, 1984.
25. Rosovsky H, Lopez J, Narvaez A: Trauma and acute medical problems in emergency rooms:
Two dimensions of alcohol involvement? The International Workshop on Alcohol and the
Emergency Room, Trieste, Italy, 1990, October 15–17.
26. Cherpitel CJ: A study of alcohol use and injuries among emergency room patients, in Gies-
brecht N, Gonzales R, Grant M, et al (eds): Drinking and Casualties: Accidents, Poisonings and
Violence in an International Perspective. London: New York, Tavistock/Routledge, 1989, pp 288–
299.
27. Cherpitel CJ: Alcohol and injuries: A review of international emergency room studies. Addic-
tion 88:923–937, 1993.
28. Cherpitel CJ, Parés A, Rodés J, Rosovsky H: Validity of self-reported alcohol consumption
in the emergency room: Data from the U.S., Mexico and Spain. J Stud Alcohol 53:203–207,
1992.
29. Cherpitel CJ: Timing of the breath analyzer: Does it make a difference? J Stud Alcohol 54:517–
519, 1993.
30. Midanik LT: Validity of self-reported alcohol use: A literature review and assessment. Br J
Addict 83:1019–1029, 1988.
31. Midanik LT: Perspectives on the validity of self reported alcohol use. Br J Addict 84:1419–1424,
1989.
32. Babor TF, Stephens RS, Marlatt GA: Verbal report methods in clinical research on alcoholism:
Response bias and its minimization. J Stud Alcohol 48:410–424, 1987.
33. Embree BG, Whitehead PC: Validity and reliability of self-reported drinking behavior: Deal-
ing with the problem of response bias. J Stud Alcohol 54:334–344, 1993.
34. Cherpitel CJ: Drinking patterns and problems associated with injury status in emergency
room admissions. Alcoholism Clin Exp Res 12:105–110, 1988.
35. Shepherd J, Irish M, Scully C, Leslie I: Alcohol consumption among victims of violence and
among comparable UK populations. Br J Addic 84:1045–1051, 1989.
36. Cherpitel CJ: Alcohol and casualties: Comparison of county-wide emergency room data with
the county general population. Addiction 90:343–350, 1995.
37. Redmond AD, Richards S, Plumhett PK: The significance of random breath alcohol sampling
in the accident and emergency department. Alcohol Alcoholism 22:341–343, 1987.
38. Soderstrom CA, Dischinger PC, Smith GS, et al: Psychoactive substance dependence among
trauma center patients. JAMA 267:2756–2759, 1992.
39. Hilton ME: Regional diversity in United States drinking practices. Br J Addic 83:519–532,1988.
40. Cherpitel CJ: Alcohol in fatal and non-fatal injuries: A comparison of coroner and emergency
room data from the same county. Alcoholism Clin Exp Res 20:338–342, 1996.
41. Midanik LT, Clark WB: The demographic distribution of US drinking patterns in 1990: De-
scriptions and trends from 1984. Am J Public Health 84:1218–1222, 1994.
42. Holder H, Cherpitel C: The end of US prohibition: A case study of Mississippi. Contemp Drug
Problems, 23:301–330, 1996.
43. Williams GD, Clem DA, Dufour MC: Apparent per Capita Consumption: National, State and
Regional Trends: 1977–2992. Rockville, MD, National Institute on Alcohol Abuse and Alcohol-
ism, Alcohol Epidemiologic Data System, 1994.
44. Room R: Normative perspectives on alcohol use and problems. J Drug Issues 5:358–368, 1975.
45. Room R: Ambivalence as a sociological explanation: The case of cultural explanations of
alcohol problems. Am Sociol Rev 41:1047–1065, 1976.
46. Lenke L: Alcohol and crimes of violence: A causal analysis. Contemp Drug Problems 11:355–
365, 1982.
47. Gondolf EW: Characteristics of perpetrators of family and nonfamily assaults. Hosp Commu-
nity Psychiatry 41:191–193, 1990.
48. Room R, Collins G (eds): Alcohol and disinhibition: Nature and meaning of the link, NIAAA
Research Monograph No. 12, Publication No. (ADM) 83-1246. Washington D.C.: Govern-
ment Printing Office, 1983.
49. Arfken CL: Temporal pattern of alcohol consumption in the United States. Alcoholism Clin Exp
Res 12:137–142, 1988.
50. Cherpitel CJ, Flaminio D, Poldrugo F: Alcohol and casualties in the emergency room: A US-
Italy comparison of weekdays and weekend evenings. Addic Res 1:223–238, 1993.
51. Cherpitel CJ, Tam TW, Midanik LT, et al: Alcohol and non-fatal injury in the US general
population: A risk function analysis. Accident Anal Prevent 27:651–661, 1995.
II
Neurobiology
Richard A. Deitrich, Section Editor

Overview
Richard A. Deitrich
All of us have seen the “mean drunk” in which the entire personality of the
individual seems to change under the influence of alcohol. It is frightening to
realize that such an easily acquired substance can rip the veneer from an
individual’s behavior to expose the violent base underneath.
This section covers some of the basic data obtained from studies of ani-
mals regarding the relationship between ethanol and aggression. Chapter 6,
by Yudko et al., outlines the preclinical research on alcohol and aggression.
They point out how difficult such research can be and they establish the link
between the influence of ethanol on animal and human aggressive behavior.
Chapter 7, an excellent chapter by Miczek et al., reviews some of the
same material and then concentrates on the relationship between GABAA–
benzodiazepine receptor complex and aggression. A recurring theme in these
chapters is the individual variability and the need to recognize this in animal
experimentation. The specter that the genetic makeup of an individual may
alter their aggressive tendencies either in the drug-free condition or under the
influence of alcohol or other drugs is raised by these findings. This is a
controversial subject; however, the recent identification of a gene that influ-
ences risk-taking behavior in humans is a good example of how such studies
can be carried out.1,2,3
Chapter 8, by Virkkunen and Linnoila, details the evidence that there is a
relationship between early-onset alcoholism, violent behavior, and serotonin
mechanisms in the brain.
Finally, moving from rodent studies to higher animal models, Higley and
Linnoila outline in Chapter 9 the development of a nonhuman primate model
Richard A. Deitrich • Department of Pharmacology, Alcohol Research Center, University of

Colorado Health Sciences Center, Denver, Colorado 80262.
121
122 II • Neurobiology
of excessive ethanol consumption that simulates many of the characteristics of

alcoholism in humans.
Thus, these four chapters, while each reviews some of the ground to be
found in others, provide a thorough review of the evidence that ethanol
precipitates violent behavior in lower animals as well as in humans, and they
begin to provide explanations at the biochemical and genetic level of how that
may come about.
A major difference between human and animal ethanol-precipitated ag-
gression may be that at high ethanol levels animals are less aggressive, per-
haps because of the physical inability to attack others. However, with hu-
mans verbal abuse or use of a weapon is still possible, even at high ethanol
levels.
The lack of tolerance development to ethanol in precipitating aggressive
behavior may also be of importance here. It is critical to determine if this exists
in humans as well as in laboratory animals.
The well-known tendency of males to be the aggressor and females to be
the victim is covered in these chapters. The influence of testosterone or per-
haps other steroids is important. However, the fact that victims, male or
female, are also often intoxicated indicates that this may play a role in the
ultimate outcome of an encounter with an intoxicated male.
References
1. Benjamin J, Li L, Patterson C, Greenberg, BD, Murphy DL, Hamer DH: Population and
familial association between the D4 dopamine receptor gene and measures of Novelty Seek-
ing. Nat Genet 12:81–84, 1996.
2. Cloninger CR, Adolfsson R, Svrakic NM: Mapping genes for human personality. Nat Genet
12:3–4, 1996.
3. Ebstein RP, Novick O, Umansky R, Priel B, Osher Y, Blaine D, Bennett ER, Nemanov L, Katz
M, Belmaker RH: Dopamine D4 receptor (D4DR) exon III polymorphism associated with the
human personality trait of Novelty Seeking. Nat Genet 12:78–80, 1996.
6
Emerging Themes in Preclinical

Research on Alcohol and Aggression
Errol Yudko, D. Caroline Blanchard, J. Andy Henrie,
and Robert J. Blanchard
Abstract. Animal research into the alcohol–aggression relationship is based on a need to under-
stand this relationship in people, and its success depends on the degree to which animal models
can provide appropriate parallels to relevant human phenomena. Comparisons of human and
animal literature suggest that parallels may be found for the following: alcohol enhances aggres-
sion in some, but not all individuals; consumption increases the probability of victimization
(being attacked by a conspecific); alcohol reduces anxiety, and socially stressed individuals show
increased voluntary consumption; alcohol reduces avoidance of threatening situations or stimuli
and may place individuals at greater risk of being attacked; both anxiety reduction and decreased
avoidance of threat may increase the probability of involvement in violent situations. These
findings suggest that a variety of mechanisms may be involved in alcohol enhancement of
aggression.
Differences in effects of alcohol on human, as opposed to animal, aggression may reflect
specific human capabilities. Although high doses of alcohol consistently reduce aggression in
laboratory animals, this may reflect motoric and sedative effects that are not relevant for human
behavior, in which verbal aggression and aggression involving the use of weapons make motor
capability less important. Human voluntary alcohol consumption may also reflect response to
stressors that also simultaneously promote aggression, a situation not paralleled by animal stud-
ies in which the drug is administered rather than voluntarily consumed. Nonetheless, obtained
parallels suggest that animal experimentation using ecologically relevant situations can provide
highly generalizable analyses of the alcohol–aggression relationship.
Errol Yudko • Department of Psychology, University of Hawaii at Manoa, Honolulu, Hawaii

96822. D. Caroline Blanchard • Bekesy Laboratory of Neurobiology, Department of Anato-
my and Reproductive Biology, University of Hawaii at Manoa, Honolulu, Hawaii 96822.
J. Andy Henrie • Department of Chemistry, University of Hawaii at Manoa, Honolulu, Hawaii
96822. Robert J. Blanchard • Department of Psychology, Bekesy Laboratory of Neurobiol-
ogy, University of Hawaii at Manoa, Honolulu, Hawaii 96822.
123
Animal research on alcohol and aggression reflects the need and desire to
utilize a full spectrum of experimental methodologies in understanding the
complex interrelationships of alcohol consumption, aggression, and violence
in humans. These interrelationships have several different aspects, each of
which may suggest specific phenomena that need to be modeled and in some
cases have been modeled in animal research.
1. Alcohol Effects on Aggression and Violence

The existence of a strong positive correlational relationship between alco-
hol consumption, violence, and violent crime has been repeatedly demon-
strated.1–4 Moreover, this relationship also appears to be somewhat selective:
alcohol is associated with violent crime at a significantly higher level than it is
with nonviolent crime.5 Miller and Potter-Efron6 reported that alcohol is one
of a small group of substances of abuse (others include PCP, amphetamines,
and cocaine) that are regularly associated with aggressive patterns of be-
havior.
A number of recent studies also suggest that the involvement of alcohol
in family violence may be particularly potent. O’Farrell and Choquette7 re-
ported that a sample of married alcoholic men had a five to six times greater
prevalence of vioIent acts toward their wives than national norms. In a study
of domestic homicide cases, 85% of the cases (70% of the suspects and 45% of
the victims) had alcohol present,8 while over 50% of maritally violent men
self-reported that drinking accompanied their abusive events.9 In compari-
sons of violent and nonviolent distressed families seeking counseling,10 the
maritally violent men consumed more alcohol than distressed but nonviolent
men. Wesner et al.11 found that alcohol played a role in 82% of male “explo-
sive rage” episodes involving a spouse or girlfriend. In a study of violent
lesbian relationships, 64% reported that alcohol or drugs were taken before or
during incidents of battering, and the frequency of drinking significantly
correlated with committing abusive acts.12 Alcohol use is also a significant
predictor of abuse toward children.13
One complicating factor in this relationship is that alcohol use is associ-
ated with victimization as well as with the commission of violent acts. The
Slade et al.8 report cited above notes that victims as well as perpetrators of
domestic homicide are likely to have been drinking at the time. Both the
incidence and severity of spousal violence is higher for women alcoholics,14
even when other factors such as alcohol problems in the spouse, income,
parental violence, and parental alcohol problems are controlled.15 In a study
of women jailed for killing their abusive husbands, daily alcohol use by the
male victim was a particularly common finding.16
6 • Themes in Preclinical Research 125
2. Effects of Aggression, Violence, and Other Stressors

on Alcohol Use and Abuse
The alcohol–aggression relationship may also be bidirectional. A family
history of violence17 predicts alcohol abuse in individuals, and particularly
the combination of alcohol problems and spouse abuse.18 However, it seems
likely that this effect is, at least in part, nonspecific, with a history of other types
of severe stress such as childhood sexual exploitation19 also leading to in-
creased alcohol (and drug) abuse. Lindenberg et al.20 reviewed nine relatively
large-scale studies done between 1984 and 1991 on the relationship of the
magnitude or intensity of social stress to one or more types of alcohol or drug
use and abuse. Of these, six reported a positive relationship, two did not, and
one reported mixed findings. In addition, analysis of studies failing to find a
relationship between stress and substance use21 suggests that factors such as
the time span following stress, possible attrition from the sample of more
affected individuals, and the use of a control group that had also been through
stressful experiences may obscure this relationship. A number of
recent studies have reported high rates of alcohol or drug use or abuse in sub-
groups that experience unusually high stress levels, such as firefighters,22
homeless women,23 and human immunodeficiency virus (HIV)-positive ho-
mosexual men.24 It is notable that these groups are extremely divergent in char-
acteristics (e.g., social and employment status) other than high stress levels,
suggesting that stress is a major common factor in their alcohol or drug use.
In a recent longitudinal study that investigated the progressive use of
alcohol in adolescents and their subsequent aggressive behaviors, White et
al.25 found that early aggressive behavior, as compared to alcohol use, is a
better predictor of later alcohol use and alcohol-related aggression. The au-
thors conclude from these data that ”aggressive individuals become heavy
drinkers because of subcultural norms, situational contexts, self-medication,
or to give themselves an excuse to act aggressively” (p. 74). An additional
view is also suggested. Both aggression and heavy alcohol consumption may
be potentiated by a common factor such as stress. Thus, in addition to a
possible general potentiation of both behaviors as a function of early or chron-
ic stress, such a relationship could facilitate the expression of both behaviors
simultaneously in response to acute stress. In a recent article,26 the behavior
patterns of 183 maritally violent men were compared to those of maritally
nonviolent men. Maritally violent men consumed significantly more alcohol
than did maritally nonviolent men. Moreover, the maritally violent men also
reported that the major stressors in their lives were their wives and children.
This is consonant with the interpretation that stress can elicit both alcohol
consumption and violence, with these men both drinking in response to
stress and simultaneously acting aggressively toward the perceived source of
their stress.
These findings from research on humans, here only very briefly outlined,
suggest several hypotheses in need of experimental attention. Over the past

decade or so, quite a number of studies using laboratory animals have at-
tempted to examine these suggested relationships.
3. Does Administration of Alcohol Increase Aggression

in Animal Models?
In 1986, Berry and Smoothy27 summarized the literature on alcohol ef-

fects on aggression in animal models, concluding, first, that the specific test
situation used is a crucial factor in this relationship. Further, the potentiation
of aggression by alcohol was seen almost exclusively at low-to-moderate
doses, with higher doses attenuating this behavior. While some animals,
notably fish, cats, and primates, did show this relationship, potentiation of
aggression by alcohol was rare in rodents, although it did tend occur only at a
low (0.3 g/kg) to moderate (1.0 g/kg) dose range such as is effective in other
animals showing this relationship more strongly. Alcohol potentiation was
more common in situations involving chronic rather than acute administra-
tion, and it appeared to reflect both individual subject variation and situation-
al characteristics. Thus, in one study28 isolate aggressive mice showed some
enhancement of aggressive behavior at 0.3 g/kg alcohol and a decrease at 0.8
g/kg, while isolate timid/defensive mice showed an enhancement at 0.8 g/kg
and sociable isolates showed no enhancement at any dose.
The individual difference variable in the alcohol-aggression relationship
has received particular attention in recent work. One area of investigation
relates to initial levels of attack. Blanchard et al.29 reported that when rats
were divided on the basis of pretest aggression screening with an intruder
into their home cages into nonaggressive, moderately aggressive, and highly
aggressive groups, only the moderately aggressive animals showed a poten-
tiation of attack following alcohol (0.3 and 0.6 g/kg) administration.29 Nonag-
gressive animals continued to fail to show attack, while the highly aggressive
animals tended to show less aggression following alcohol administration.
This view that moderately but not highly aggressive rats show enhanced
aggression in response to ethanol administration is consonant with findings
in mice. When male mice were isolated (a treatment that increases aggression)
for 10 days, ethanol (0.8 mg/kg) decreased their aggression, but in less ag-
gressive 5-day isolates, the same dose of ethanol increased aggression.30
When male mice of different weights—light, average, or heavy (heavy mice
were significantly more aggressive than light mice but all were aggressive)—
were paired with male mice of equivalent weight, no significant ethanol effect
was evident.31 However, the nonsignificant differences did tend in the pre-
dicted direction, with light animals showing a twofold increase in aggression
at 0.8 mg/kg ethanol, while both average and heavy animals showed small
nonsignificant decreases in aggression.
An additional and perhaps interacting individual difference factor is tes-
tosterone level. Miczek and his colleagues have reported a number of findings
suggesting that testosterone levels may be important in the alcohol–aggres-
sion relationship, with higher testosterone levels tending to potentiate this
relationship in male mice32 and squirrel monkeys,33 but not in females.34 As
aggression is an androgen-dependent behavior in male rodents,35 tes-
tosterone deficiency may be an important factor in findings that alcohol does
not increase aggression in nonaggressive rats.
A range of studies of the effects of situational and opponent-related
factors in the alcohol–aggression relationship suggest that alcohol may en-
hance aggression because it reduces some type of inhibition on attack. Miczek
and O’Donnell36 reported that alcohol or chlordiazepoxide doses that failed to
potentiate attack by Swiss-Webster mice on intruders in their home cages did
so when tests were held in a neutral arena that tended to inhibit fighting
relative to that of residents in their home cages. An additional type of inhibi-
tion may be involved in the sharply lower level of attack by males toward
females in comparison to that seen to other males. In rat colonies acute alco-
hol administration produced a dramatic shift from male to female targets of
attack during group formation (at 0.6 and 1.2 g/kg)37 and increased attacks
toward familiar (but not unfamiliar) females in an established group.38 Inhibi-
tion based on size of opponent may also modulate the alcohol–aggression
relationship. Alcohol potentiated attack by lactating females on small, but not
large, male intruders into their nesting cages.39 Since even the small intruders
in this test were about the size of the females (and the large males about 50%
heavier), this finding suggests that alcohol can reduce some degree of inhibi-
tion, but is less effective with extremely potent inhibitory factors. This is
consonant with findings40 that attack by male rats on a conspecific show a
precipitous decline after cat exposure, and that alcohol doses in typical effec-
tive ranges (0.3–0.6), while tending to increase attack, fail to do so signifi-
cantly.
4. Alcohol and the Recipient of Attack
As an increasing volume of human literature suggests, alcohol may influ-

ence aggression in part through its effect on the recipient of attack. Adminis-
tration of alcohol to intruder mice and rats and to subordinate male squirrel
monkeys increased attack seen toward these animals by resident or dominant
male conspecifics.41 This particular mechanism of potentiation of attack oc-
curs at somewhat higher doses than does direct potentiation following ad-
ministration of alcohol to the attacker. Most studies of alcohol–aggression
relationships focus on attack by residents on intruders into their home cages
or of dominant on subordinate animals, models that tend to produce offen-
sive attack patterns. However, the recipient of attack by a conspecific typically
responds with a pattern of defensive behavior that includes defensive or
retaliatory attack components. The differentiation of offensive and defensive
forms of attack is now relatively well accepted and it involves differences in

virtually every component of these neurobehavioral systems. Offensive and
defensive attack involve different antecedent conditions, behavior patterns,
and specific targets for attack on the body of the opponent,42 as well as, in
many cases, differential response to drugs and other biological manipula-
tions.43,44 The question of the degree of fit between offensive (or defensive)
attack and human aggression has never been satisfactorily resolved. Indeed,
it has hardly been studied, and as yet there have not been any serious at-
tempts to provide detailed descriptions of human aggression in terms that
might be unequivocally related to either offensive or defensive attack modes.
However, it has been suggested45 that much of human aggression corre-
sponds better to a defensive than to an offensive mode of attack. Certainly,
given that fear tends to reduce offensive attack and promote defensive attack,
it is to be expected that human aggression often contains at least some compo-
nents of the latter.
The only direct test of the effect of alcohol on defensive attack measured a
pattern of defensive threat vocalizations, jump-attacks, and bites to an ap-
proaching threat stimulus (the human experimenter) and to a terminally anes-
thetized conspecific.44 These were potentiated by administration of moderate
(0.6 g/kg) doses of alcohol. Peterson and Pohorecky46 have reported that
ethanol administration, while increasing the frequency and severity of biting
attacks in a resident–intruder test, also shifts the distribution of wounds
made by dosed residents on intruders from the normal upper-back target site
for offensive attack (73% wounds for intruders bitten by vehicle control resi-
dents) to the lower back and ventrum (65% wounds made by ethanol-dosed
residents). This finding suggests that alcohol may blur the distinction be-
tween offensive and defensive attack modes that is normally seen in an expe-
rienced resident male, perhaps by potentiating defensive attack patterns.
Alternatively, since bites to the ventrum appear to be strongly inhibited in
rats,42 the finding of ventrum wounds may suggest again that alcohol acts to
reduce a variety of inhibitions relevant to aggressive behavior.
5. Does Social Stress Enhance Voluntary Alcohol

Consumption (VAC)?
The studies described above provide some indication that alcohol may
promote aggression in laboratory rodents and other animals, either through
direct effects on the attacker or, when administered to the target of an attack,
by increasing conspecific attack on the dosed animal and through potentia-
tion of defensive attack. However, consideration of the magnitude of the
potentiation obtained in animal studies relative to the magnitude of the alco-
hol–aggression/violence conjunction outlined in the human literature sug-
gests a considerable remaining discrepancy. Does this reflect a real limitation
on the usefulness of animal studies to model alcohol effects on aggressive or
violent behaviors, or does it indicate that the factors typically encountered in
violent human situations have not yet been sufficiently analyzed and then
duplicated in the animal work?
One aspect of the alcohol–aggression relationship that has as yet re-
ceived relatively little attention in animal models concerns social stress. The
human literature cited earlier strongly suggests that stress is a major risk
factor for substance use and addiction (e.g., refs. 17–25). This possibility,
difficult to test experimentally in people, may have extended implications
for the observed effects of alcohol on aggression/violence in human popula-
tions.
The relationship between social stress and voluntary consumption of
alcohol (and some other aggression-promoting substances) can be and has
been examined in laboratory rats and mice. Social grouping reliably increases
alcohol consumption in male rats,47 and those males that show the greatest
alcohol consumption are also characterized by withdrawal, inactivity, and
lowered dominance.48 A visible burrow system (VBS) providing tunnels and
chambers to mimic the natural habitat of wild rats produces particularly high
levels of fighting within rat groups, facilitating the development of a strong
dominance hierarchy, and thus presumably stress, among males.49 In rat
groups housed in a VBS, subordinate males showed increased voluntary
alcohol consumption (VAC) .50 When alcohol intake and defensive behaviors
were measured prior to VBS grouping, subordinate but not dominant males
showed an increase in VAC, indicating that the effect is not due to differential
alcohol preference for animals that become subordinate in grouping situa-
tions.51 This phenomenon is not restricted to rats. Hilakivi-Clark and Lister31
have reported enhanced VAC in subordinate but not dominant mice.
Are these subordinate males truly stressed? While grouping per se is not
likely to be a stressor in social animals such as rats, subordination reflects the
results of fighting within groups, and the stress effects of VBS housing on
subordinate males have been extensively documented. In the VBS, subordi-
nates show greatly enhanced defensive behavior and reduced social, sexual,
exploratory, and aggressive activity within the habitat.52 They also suffer
increased mortality,53 a phenomenon that may be similar to the “social stress
deaths” reported by Barnett54,55 among wild rats introduced to an established
colony. Early subordinate mortality increases with features that enhance
group aggression levels such as the presence of females or the provision of a
burrowing habitat.53 Moreover, individual subordinate mortality appears to
be related to the degree of stress experienced and expressed in individual
behavior. In somewhat less stressful open-bin colonies in which subordinate
mortality is spread out over the normal life span of rats, early-dying individu-
als can be predicted several hundred days in advance by a pattern of avoid-
ance of the dominant male and reduced sexual behavior.53 The period over
which individual subordinate mortality could be predicted is so long (about
one quarter of the life span of the rat) that it is highly unlikely that the animal
died of any bodily disease or wound that was directly altering its behavior
when the behavior change first appeared. Also, in behavioral tests of “emo-
tionality” outside the colony, subordinates show changes that sometimes
persist for weeks after the group is disbanded,56 providing further confirma-
tion of a severe behavioral stress response.
Subordinate males also show a number of physiological differences sug-
gestive of stress, including adrenal and spleen enlargement and reductions in
the weight of the thymus and testes.57 Basal corticosterone (CORT) levels are
increased and corticosterone-binding globulin (CBG) levels decreased, sug-
gesting that free CORT levels may have been dramatically increased in these
animals. Plasma testosterone levels are dramatically decreased and alterations
of hypothalamic–pituitary–adrenal (HPA) axis functioning have been found
in many subordinates.57 A variety of brain system changes, including specific
alterations of regional serotonin receptors,58 mineralocorticoid and glucocor-
ticoid receptors,59 corticotropin-releasing hormone receptors,60 and galanin
levels61 attest to widespread brain involvement in the chronic stress response
of these animals.
6. Stress and Substance Abuse of Other

Aggression-Impacting Substances
Alcohol is not the only substance for which stress facilitates self-adminis-
tration. Stressful rearing conditions have been shown to increase cocaine62
and amphetamine63 self-administration in rats, and repeated tail pinch also
enhanced vulnerability to self-administration of amphetamine. Offspring of
mothers stressed by a heavy schedule of restraint during gestation days 14–21
also showed enhanced amphetamine self-administration,64 suggesting that
maternal hypersecretion of CORT may have lasting effects on the fetus. Con-
sonant with this view, the restraint stress-based enhancement of locomotor
response to amphetamine and morphine was not obtained in adrenalec-
tomized animals with CORT implants.65
Rats vulnerable to amphetamine self-administration, either because of
individual differences66 or following prenatal stress,64 also show enhanced
locomotor response to amphetamine in a novel but not a familiar environ-
ment.67 These animals have higher magnitude and longer-lasting stress-in-
duced increases in nucleus accumbens concentrations of dopamine than do
animals not showing the enhanced locomotory response to amphetamine.68
The authors suggest that changes in these brain systems may constitute an
important neurobiological substrate of the predisposition to acquire amphet-
amine self-administration as well as other addictive behaviors. Rats that de-
veloped self-administration also tended to be those that showed an enhanced
locomotor response in a novel environment and acquired schedule-induced
polydipsia. However, when tested for schedule-induced polydipsia first (an
experience that is interpreted as having a coping function), the polydipsic rats
subsequently failed to acquire self-administration and had a reduced locomo-
tor response to novelty.69 It is particularly notable that cocaine and amphet-
amine, like alcohol, are among the relatively few substances that do appear to
be associated with aggression in people.6 Again, however, individual differ-

ence factors are important, in addition to the effects of stressful experience.
Taylor et al.70 reported that rats that show a high plasma catecholamine re-
sponse to stress show higher intake of a cocaine solution than do low cate-
cholamine stress responders.
The specificity of this relationship is not yet clear. Pohorecky et al.71
found that prenatal stress (gestational days 14–21) also increased sensitivity
to caffeine on measures of corner activity and rearing. While some studies,
notably those involving measures more clearly related to defensive attack,72,73
have found that caffeine increases aggression, others typically involving of-
fensive attack74,75 have reported no effect.
7. The Relationship of Aggression to the Predisposition to VAC

Some of the human literature described above26,76 suggests that aggres-
sive individuals may be predisposed (in particular situations?) to both alcohol
consumption and alcohol-related aggression. Laboratory animal research may
again provide a parallel. Voluntary alcohol consumption measured prior to
grouping in the VBS does not predict dominance-subordination status after
grouping. However, a positive correlation between pregrouping biting of an
intruder and postgrouping alcohol consumption for subordinates strongly
suggests that it is the more aggressive males that are most stressed by the
process of subordination. Notably, postgrouping VAC was not related to total
attack times in pregrouping tests with the intruder, and additional analyses
within this study strongly suggest that “aggressiveness” is better measured
through terminal components of the attack process such as biting rather than
overall attack times, which primarily measure lengthy but nondamage com-
ponents such as the lateral attack or chasing.51 In addition, behavioral an-
alyses of those subordinate males (nonresponders) that show a particular
change in HPA axis response, involving minimal or no CORT elevation to
acute stress,57 suggests that these are individuals that are both highly aggres-
sive and highly defensive.56 Rats bred to exhibit increased basal CORT77
response and decreased CORT output in response to stress, a pattern similar
to that of the nonresponder rats, spontaneously drink more ethanol than do
control rats bred from the same line. These results suggest that VAC may be,
at least in part, related to stress, and that in the VBS highly aggressive subor-
dinates are particularly stressed animals.
A single study using mouse colonies has reported similar, although not
identical, findings.78 Subordinate male mice generally increased alcohol in-
take, and this was particularly marked in the most severely wounded subor-
dinates, which showed an enhancement of the ratio of alcohol intake to total
fluid intake compared to dominants and controls during a 2-week period
following disbanding of the groups.78 Overall correlations between pregroup-
ing biting and postgrouping alcohol consumption for subordinates were not
reported, and, as in the Blanchard et al.51 study, total pregrouping attack

times tended not to be reliably correlated to postgrouping VAC. However,
the single reliable correlation (of 8 calculated) was in the opposite direction to
what might be expected if total attack times and bites are valid measures of
the same factor.
8. Is an Anxiolytic or Inhibition-Reducing Property of Alcohol an

Important Component of the Relationship between Stress,
Voluntary Alcohol Consumption, and Aggression?
The human literature suggests that stress increases voluntary alcohol

consumption and the animal literature provides experimental confirmation of
this. Alcohol also reduces the CORT response to a variety of stressors79–81 and
high-ethanol-consuming rats show greater preference for diazepam in com-
parison to rats with a lower preference for alcohol.82 These findings suggest
that anxiety reduction may be a mechanism in enhanced ethanol consump-
tion.
Preclinical research provides strong support for the view that alcohol is
an anxiolytic. In mice, alcohol produces an anxiolytic response in the black-
white box83 and in the elevated plus-maze.84 Lister and Hilakivi85 reported
anxiolytic effects of alcohol in rats in both a hole-board test and in the elevated
plus-maze. Alcohol withdrawal has anxiogenic effects that have been used to
provide a model for analysis of anxiolytic drugs (e.g., File et al.86). In two
defense test batteries measuring a range of natural defensive behaviors to
potential predators and conspecifics, alcohol provided a profile of effects on
proximate avoidance of potentially dangerous areas and risk assessment (in-
vestigation of possible danger) behaviors that was very similar to the profile
obtained with diazepam. Alcohol also reduced the cat-induced inhibition of
consummatory behavior.40
One aspect of these specific effects on defensive behaviors is particularly
notable. The dose-dependent reduction in proximate avoidance by alcohol is
so pronounced that alcohol-dosed rats may preferentially approach areas in
which they have seen a highly dangerous stimulus (i.e., cat). The effect of this
specific alteration of defensive behavior may be that alcohol produces an
inclination to approach, rather than avoid, potentially dangerous situations, a
factor that is likely to be involved in the findings of more conspecific attacks
on alcohol-drugged animals,41 as well as the enhanced levels of alcohol intake
for victims of human violence.6,8
The previously described findings of an alcohol enhancement of the pro-
portion of attacks by male rats toward females, and particularly toward famil-
iar females, suggests a slightly different interpretation: that alcohol reduces a
variety of inhibitions, some but possibly not all of which are associated with
anxiety. Certainly the reduction in approach to a potentially dangerous stimu-
lus can be viewed as reflecting inhibition, in this case an inhibition that
responds to anxiolytic drugs.87 Inhibition of attack on females has an obvious

adaptive potential, but little is known of the mechanisms involved; and there
is, in the absence of relevant data, no reason to assume that this is related to
anxiety or indeed to other types of inhibition. Some commonality of neuro-
behavioral inhibitory systems is strongly suggested by the work of Bohus,
Koolhaas, and their colleagues at Groningen, and some aspects of the biology
of these systems are beginning to be described.88–90 Nonetheless, an under-
standing of the relationships among different inhibitory systems and their
response to alcohol is just beginning to develop.
9. Relationship of Preclinical Studies of Alcohol Effects

to Human Alcohol-Aggression Phenomena
Briefly summarized, preclinical studies of alcohol and aggression, with
consideration also of the potential interaction of these with stress and re-
sponses to stress suggest the following. Alcohol at low to moderate doses
may potentiate both offensive and defensive attack. However, the former
appears to be potentiated in only a subpopulation of animals. Moreover,
alcohol strikingly reduces avoidance of potentially threatening or dangerous
situations. These three findings suggest a particularly potent scenario for
those situations in which both participants in a dyadic interaction have taken
alcohol. If potentially hostile or aversive interactions arise, one or both partici-
pants fail to show a normal avoidance response but instead approach. The
approach itself provides stimuli that may elicit either an offensive or a defen-
sive attack, depending on the situation and other motivational/emotional
factors of the individual responding. The attack by one participant then trig-
gers a defensive (retaliatory) attack by the other, heightened by the effects of
alcohol and triggering an escalating series of mutually aggressive acts.
There are differences, however, between the existing animal literature
and human experience that may reduce or obscure the relevance of animal
studies to an understanding of the effects on alcohol on human behavior.
First, in the vast majority of animal studies when alcohol effects are to be
evaluated, alcohol is administered to the animal, not taken voluntarily. If, as
both animal and human literature suggest, aggressiveness and alcohol prefer-
ence are positively correlated (under specific circumstances?), then the alco-
hol–aggression relationship may be more clearly seen when voluntary con-
sumption is allowed.
An additional difference, of possibly immense importance, is that aggres-
sion or violence in humans reflects two factors that are not, and are not likely
to be, of much significance in laboratory rodent studies. First, human aggres-
siveness is often, and quite reasonably, evaluated in terms of verbal rather
than physical violence. Hostile, insulting, and denigrating speech is capable
of doing lasting damage to relationships; to the mental well-being and devel-
opment of children, in particular, who are exposed to it; and, it comprises an
especially relevant stimulus for the elicitation of retaliatory hostility and

sometimes physical violence from interacting individuals. Second, human
aggression often involves weapon use. The reverse U-shaped dose–response
relationship consistently obtained in those animal studies that do show an
alcohol potentiation of aggression probably reflects both sedation and motor
impairment at higher doses. Indeed, this effect may be involved in the finding
that chronic rather than acute alcohol intake is more likely to produce en-
hanced aggression in animal studies,27 if chronic administration reduces the
motoric and sedative effects of a given dose of alcohol. Notoriously, the
possession of a weapon can reduce or eliminate the effects of motor impair-
ment and even sedation associated with high levels of alcohol intake, a com-
mon factor in human experience and one that is unlikely to be easily dupli-
cated in laboratory animal studies.
Overall, however, the animal literature does hold the promise of provid-
ing relatively comprehensive models for experimental investigation of many
aspects of the alcohol–aggression relationship. While relatively few, preclini-
cal studies incorporating attention to mechanisms such as stress and inhibi-
tory systems and their interactions with individual differences may provide a
more detailed understanding of the bidirectional relationship between alcohol
and aggression and may provide specific suggestions for improved control of
some of the consequences of this relationship.
ACKNOWLEDGMENT. Support for this research was provided by a grant from

the Howard Hughes Medical Institute through the Undergraduate Biological
Sciences Education Program.
References
1. Shupe LM: Alcohol and crimes: A study of the urine alcohol concentration found in 882
persons arrested during or immediately after the commission of a felony. J Criminal Law
Criminol 44:661–665, 1954.
2. Wolfgang ME, Strohm EB: The relationship between alcohol and criminal homicide. Q J Stud
Alcohol 17:411–425, 1956.
3. Pernanen K: Alcohol and crimes of violence, in Kissin B, Begleiter H (eds.): The Biology of
Alcoholism: Social Aspects of Alcoholism. New York, Plenum, 1976, pp 351–444.
4. Straus MA, Gelles RJ, Steinmetz SK: Behind Closed Doors: Violence in the American Family. New
York, Anchor/Doubleday, 1980.
5. Murdoch C, Pihl RO, Ross D: Alcohol and crimes of violence: Present issues. Int J Addict
25:1065–1081, 1990.
6. Miller MM, Potter-Efron RT: Aggression and violence associated with substance abuse. Chem
Depend Treat 3:1–36, 1989.
7. O’Farrell TJ, Choquette K: Marital violence in the year before and after spouse-involved
alcoholism treatment. Family Dynamics Addict Q 1:32–40, 1991.
8. Slade M, Daniel LJ, Heisler CJ: Application of forensic toxicology to the problem of domestic
violence. J Forensic Sci 36:708–713, 1991.
9. Fagan RW, Barnett OW, Patton JB: Reasons for alcohol use in maritally violent men. Am J
10. Russell MN, Llpov E, Phillips N, White B: Psychological profiles of violent and nonviolent
maritally distressed couples. Psychotherapy 26:81–87, 1989.
11. Wesner D, Patel C, Allen J: A study of explosive rage in male spouses counseled in an
Appalachian mental health clinic. J Counsel Dev 70:235–241, 1991.
12. Schilit R, Lle G, Montagne M: Substance use as a correlate of violence in intimate lesbian
relationships. J Homosex 19:51–65, 1990.
13. Suh EK, Abel EM: The impact of spousal violence on the children of the abused. J Independ
Social Work 4:27–34, 1990.
14. Miller, BA: The interrelationships between alcohol and drugs and family violence. In: De La
Rosa, M, Lambert, EY, and Gropper, BA (eds.), Drugs and Violence: Causes, Correlates, and
Consequences. NIDA Research Monograph No. 103, DHHS Publication No. (ADM) 91–1721,
Washington D.C.: Government Printing Office 1990, pp. 177–207.
15. Miller BA, Downs WR, Gondoli DM: Spousal violence among alcoholic women as compared
to a random household sample of women. J Stud Alcohol 50:533–540, 1989.
16. Foster LA, Veale CM, Fogel CI: Factors present when battered women kill. Special issue:
Family violence. Issues Ment Health Nurs 10:273–284, 1989.
17. Plutchik A, Plutchik R: Psychosocial correlates of alcoholism. Integrative Psychiatry 6:205–210,
1988.
18. Hastings Je, Hamberger LK: Personality characteristic of spouse abusers: A controlled com-
parison. Violence Victims 3:31–48, 1988.
19. Rew L: Long-term effects of childhood sexual exploitation. Issues Ment. Health Nurs. 10:229–
244, 1989.
20. Lindenberg CS, Gendrop SC, Reiskin HK: Empirical evidence for the social stress model of
substance abuse. Res Nurs Health 16:351–362, 1993.
21. Tennant C, Goulston K, Dent O: Clinical psychiatric illness in prisoners of war of the Japa-
nese: Forty years after release. Psychol Med 16:833–839, 1986.
22. Boxer PA, Wild D: Psychological distress and alcohol use among fire fighters. Scand J Work
Environ Health 19:121–125, 1993.
23. Smith EM, North CS, Spitznagel EL: Alcohol, drugs, and psychiatric comorbidity among
homeless women: An epidemiologic study. J Clin Psychiatry 54:82–87, 1993.
24. Atkinson JH Jr, Grant I, Kennedy CJ, et al: Prevalence of psychiatric disorders among men
infected with human immunodeficiency virus. A controlled study. Arch Gen Psychiatry
45:859–864, 1988.
25. White HR, Brick J, Hansel1 S: A longitudinal investigation of alcohol use and aggression in
adolescence. J Stud Alcohol (Suppl 11):62–77, 1993.
26. Barnett OW, Fagan RW: Alcohol use in male spouse abusers and their female partners. J
Family Violence 8:1–24, 1993.
27. Berry MS, Smoothy R: A critical evaluation of claimed relationships between alcohol intake
and aggression in infrahuman animals, in Brain PF (ed): Alcohol and Aggression. London,
Croom-Helm, 1986, pp 84–137.
28. Krsiak M, Elis J, Poschlova N, Masek K: Increased aggressiveness and lower brain seroton-
in levels in offspring of mice given alcohol during gestation. J Stud Alcohol 38:1696–1704,
1977.
29. Blanchard RJ, Hori K, Blanchard DC, Hall J: Ethanol effects on aggression of rats selected for
different levels of aggressiveness. Pharmacol Biochem Behav 27:641–644, 1987.
30. Hilakivi LA, Lister RG: Effects of ethanol on the social behavior of group-housed and isolated
mice. Alcoholism Clin Exp Res 13:622–625, 1989.
31. Hilakivi-Clarke LA, Lister RG: The role of body weight in resident–intruder aggression.
Aggressive Behav 18:281–288, 1992.
32. DeBold JF, Miczek KA: Testosterone modulates the effects of ethanol on male mouse aggres-
sion. Psychopharmacology 86:286–290, 1985.
33. Winslow JT, Miczek KA: Naltrexone blocks amphetamine-induced hyperactivity, but not
disruption of social and agonistic behavior in mice and squirrel monkeys. Psychopharmacology
96:493–499, 1988.
34. Lisciotto CA, Debold JF, Miczek KA: Sexual differentiation and the effects of alcohol on
aggressive behavior in mice. Pharmacol Biochem Behav 35:357–362, 1990.
35. Matochik JA, Sipos ML, Nyby JG, Barfield RJ: Intracranial androgenic activation of male-
typical behaviors in house mice: Motivation versus performance. Behav Brain Res 60:141–149,
1994.
36. Miczek KA, O’Donnell JM: Alcohol and chlordiazepoxide increase suppressed aggression in
mice. Psychopharmacology 69:39–44, 1980.
37. Blanchard RJ, Hori K, Flannelly K, Blanchard DC: The effects of ethanol on the offense and
defensive behaviors of male and female rats during group formation. Pharmacol Biochem Behav
26:61–64, 1987.
38. Blanchard RJ, Blanchard DC: The relationship between ethanol and aggression: Studies using
ethological models, in Olivier B, Mos J, Brain PF (eds.): Ethopharmacology of Agonistic Behavior
in Animals and Humans. Dordrecth/Boston/Lancaster, Martinus Nijhoff Publishers, 1987,
pp 145–161.
39. Blanchard DC, Flannelly K, Hori K, et al: Ethanol effects on female aggression vary with
opponent size and time within session. Pharmacol Biochem Behav 27:645–648, 1987.
40. Blanchard RJ, Blanchard DC, Flannelly KJ, Hori K: Ethanol effects on freezing and conspecific
attack in rats previously exposed to a cat. Behav Process 16:193–201, 1988.
41. Miczek KA, Winslow JT, Debold JF: Heightened aggressive behavior by animals interacting
with alcohol-treated conspecifics: Studies with mice, rats and squirrel monkeys. Pharmacol
Biochem Behav 20:349–353, 1984.
42. Blanchard RJ, Blanchard DC: Aggressive behavior in the rat. Behav Biol 21:197–224, 1977.
43. Blanchard RJ, Kleinschmidt CK, Flannelly KJ, Blanchard DC: Fear and aggression in the rat.
44. Blanchard R, Blanchard DC: Ethological models of fear and angry aggression. Clin Neurophar-
macol 9(Suppl 4):383–385, 1986.
45. Albert DJ, Walsh ML, Jonik RH: Aggression in humans: What is its biological foundation?
Neurosci Biobehav Rev 17:405–425, 1993.
46. Peterson JT, Pohorecky LA, Hamm MW: Neuroendocrine and beta-adrenoceptor response to
chronic ethanol and aggression in rats. Pharmacol Bichem Behav 34:247–253, 1989.
47. Ellison G, Staugaitis S, Crane P: A silicone delivery system for producing binge and continu-
ous ethanol intoxication in rats. Pharmacol Biochem Behav 14:207–211, 1981.
48. Ellison G, Levy A, Lorant N: Alcohol-preferring rats in colonies show withdrawal, inactivity,
and lowered dominance. Pharmacol Biochem Behav 18(Suppl 1):565–570, 1983.
49. Blanchard RJ, Blanchard DC: Antipredator defensive behaviors in a visible burrow system. J
Comp Psychol 103:70–82, 1989.
50. Blanchard RJ, Hori K, Tom P, Blanchard DC: Social structure and ethanol consumption in the
laboratory rat. Pharmacol Biochem Behav 28:437–442, 1987.
51. Blanchard RB, Flores T, Magee L, et al: Pregrouping aggression and defense scores influences
alcohol consumption for dominant and subordinate rats in the visible burrow systems. Ag-
gressive Behav 18:459–467, 1992.
52. Blanchard DC, Sakai RR, McEwen B, et al: Subordination stress: Behavioral, brain, and
neuroendocrine correlates. Behav Brain Res 58:113–121, 1993.
53. Blanchard RJ, Blanchard DC, Flannelly KJ: Social stress, mortality, and aggression in colonies
and burrowing habitats. Behav Process 11:209–213, 1985.
54. Barnett SA: Physiological effects of “social stress” in wild rats. 1. The adrenal cortex. J
Psychosom Res 3:1–11, 1958.
55. Barnett SA: Enigmatic death due to “social stress” a problem in the strategy of research.
Interdisciplinary Sci Rev 13:40–51, 1988.
56. Yudko E, Bjorenson K, Blanchard RJ: Behavioral differences of dominant and subordinate
rats in a visible burrow system. Abstracts of the Annual Meeting for the Society of Neuro-
science 1993. (Abstract)
57. Blanchard DC, Spencer RL, Weiss SM, et al: Visible burrow system as a model of chronic
social stress: Behavioral and Neuroendocrine correlates. Psychoneuroendocrinology 20:117–134,

1995.
58. McKittrick CR, Blanchard DC, Blanchard RJ, et al: Serotonin receptor binding in a colony
model of chronic social stress. Biol Psychiatry 37:383–393, 1995.
59. Chao HM, Blanchard DC, Blanchard RJ, et al: The effect of social stress on hippocampal gene
expression. Cell Mol Neurobiol 4(6):543–560. 1993.
60. Albeck DS, McKittrick CR, Blanchard DC, et al: Chronic social stress inhibits corticotropin-
releasing hormone mRNA in the paraventricular nucleus of a subpopulation of subordinates.
Abstracts of the Society of Neuroscience 1994. (Abstract)
61. Holmes PV, Blanchard DC, Blanchard RJ, et al: Chronic social stress increases levels of
preprogalanin mRNA in the rat locus coeruleus. Pharmacol Biochem Behav 50:655–660, 1995.
62. Schenk S, Hunt T, Klukowski G, Amit Z: Isolation housing decreases the effectiveness of
morphine in the conditioned taste aversion paradigm. Psychopharmacology 92:48–51, 1987.
63. Maccari S, Piazza PV, Deminere JM, Lemaire V: Life events-induced decrease of corticoste-
roid type I receptors is associated with reduced corticosterone feedback and enhanced vul-
nerability to amphetamine self-administration. Brain Res. 574:7–12, 1991.
64. Deminiere JM, Piazza PV, Guegan G, Abrous N: Increased locomotor response to novelty
and propensity to intravenous amphetamine self-administration in adult offspring of
stressed mothers. Brain Res 586:135–139, 1992.
65. Deroche V, Piazza PV, Casolini P, Maccari S: Stress-induced sensitization to amphetamine
and morphine psychomotor effects depend on stress-induced corticosterone secretion. Brain
Res 598:343–348, 1992.
66. Deminiere JM, Piazza PV, le Moal M, Simon H: Experimental approach to individual vul-
nerability to psychostimulant addiction. Neurosci Biobehav Rev 13:141–147, 1989.
67. Piazza PV, Calza L, Giardino L, Amato G: Chronic thioridazine treatment differently affects
DA receptors in striatum and in mesolimbo-cortical systems. Pharmacol Biochem Behav 35:937–
942, 1990.
68. Rouge-Pont F, Piazza PV, Kharouby M, le Moal M: Higher and longer stress-induced in-
crease in dopamine concentrations in the nucleus accumbens of animals predisposed to
amphetamine self-administration: A microdialysis study. Brain Res 602:169–174, 1993.
69. Piazza PV, Mittleman G, Deminiere JM, et al: Relationship between schedule-induced poly-
dipsia and amphetamine intravenous self-administraiton. Individual differences and role of
experience. Behav Brain Res 55:185–193, 1993.
70. Taylor J, Hams N, Vogel WH: Voluntary alcohol and cocain consumption in “low” and “high
stress plasma catecholamine responding rats. Phamcol Biochem Behav 37:359–363, 1990.
71. Pohorecky LA, Roberts P, Cotler S, Carbone JJ: Alteration of the effects of caffeine by prenatal
stress. Pharmacol Biochem Behav 33:55–62, 1989.
72. Eichelman B, Orenberg E, Hackley P, Barchas J: Methylxanthine-facilitated shock-induced
aggression in the rat. Psychopharmacology 56:305–308, 1978.
73. Emley GS, Hutchinson RR: Unique influences of ten drugs upon post-shock biting attack and
pre-shock manual responding. Pharmacol Biochem Behav 19:5–12, 1983.
74. Hansen S, Ferreira A: Food intake, aggression, and fear behavior in the mother rat: Control
by neural systems concerned with milk ejection and maternal behavior. Behav Neurosci
100:64–70, 1986.
75. File SE, Guardiola Lemaitre BJ: I-Fenfluramine in tests of dominance and anxiety in the rat.
Neuropsychobiology 20:205–211, 1988.
76. White HL, Ascher JA: Preclinical and early clinical studies with BW 1370U87, a reversible
competitive MAO-A inhibitor. Clin Neuropharmacol 15(Suppl 1 Pt A):343A–344A, 1992.
77. Prasad C, Prasad A: A relationship between increased voluntary alcohol preference and basal
hypercorticosteronemia associated with an attenuated rise in corticosterone output during
stress. Alcohol 12:59–63, 1995.
78. Hilakivi Clarke LA, Lister RG: Social status and voluntary alcohol consumption in mice:
Interaction with stress. Psychopharmacology 108:276–282, 1992.
79. Pohorecky LA, Rassi E, Weiss JM, Michalak M: Biochemical evidence for an interaction of
ethanol and stress: Preliminary studies. Alcoholism Clin Exp Res 4:423–426, 1980.
80. Brick J, Horowitz GP: Alcohol and morphine induced hypothermia in mice selected for
sensitivity in ethanol. Pharmacol Biochem Behav 16:473–479, 1982.
81. Thomas R, Beer R, Harris B, et al: GH responses to growth hormone releasing factor in
depression. J Affect Disord 16:133–137, 1989.
82. Wolfgramme J, Heyne A: Social behavior, dominance, and social deprivation of rats deter-
mine drug choice. Pharmacol Biochem Behav 38:389–385, 1991.
83. Costal1 B, Kelly ME, Naylor RJ, Onaivi ES: Actions of buspirone in a putative model of
anxiety in the mouse. J Pharm Pharmacol 40:494–500, 1988.
84. Melchior CL, Ritzmann RF: Dehydroepiandrosterone is an anxiolytic in mice on the plus
maze. Pharmacol Biochem Behav 47:437–441, 1994.
85. Lister RG, Hilakivi LA: The effects of novelty, isolation, light and ethanol on the social
behavior of mice. Psychopharmacology 96:181–187, 1988.
86. File SE, Zharkovsky A, Hitchcott PK: Effects of nitrendipine, chlordiazepoxide, flumazenil
and baclofen on the increased anxiety resulting from alcohol withdrawal. Prog Nurop-
sychopharmacol Biol Psychiatry 16:87–93, 1992.
87. Blanchard DC, Veniegas R: Alcohol and anxiety: Effects on offensive and defensive aggres-
sion. J Stud Aggression (Supplement 11):9–19, 1993.
88. Roozendaal B, Wiersma A, Driscoll P, Koolhaas JM: Vasopressinergic modulation of stress
responses in the central amygdala of the Roman high-avoidance and low-avoidance rat. Brain
Res 596:35–40, 1992.
89. de Ruiter C, Cohon L: Persoonlijkheidskenmerken van patienten met agorafobie: een studie
met het Comprehensive System voor de Rorschach. (Personality characteristics of patients
with agoraphobia: A study with the comprehensive system for the Rorschach Test.) Neder-
lands Tijdschrift voor de Psychologie en haar Grensgebieden 48:35–42, 1993.
90. Compaan JC, Wozniak A, De Ruiter AJH, et al: Aromatase activity in the preoptic area differs
between aggressive and nonaggressive male house mice. Brain Res Bull 35:1–7, 1994.
7
Alcohol, GABAA–Benzodiazepine
Receptor Complex, and Aggression
Klaus A. Miczek, Joseph F. DeBold, Annemoon
M.M. van Erp, and Walter Tornatzky
Abstract. Neurobiological investigations have become productive since experimental protocols

were developed that engender large increases in aggressive behavior after acute alcohol chal-
lenges in individual experimental animals. Recent developments extended the heightened ag-
gressive behavior to rats that self-administered alcohol shortly before the social confrontation.
Quantitative ethological analysis revealed that alcohol prolongs “bursts” of aggressive acts and
displays and disrupts communication between the aggressive animal and the opponent who
defends, submits, or flees. Pharmacological modulation of the GABA, receptor with ben-
zodiazepine agonists and neuroactive steroids results in dose-dependent biphasic changes in
aggressive behavior that mimic the dose–effect function of alcohol; benzodiazepines potentiate
the aggression-heightening effects of alcohol as well as the behaviorally suppressive effects; and
antagonists at benzodiazepine receptors prevented the aggression-heightening effects of alcohol.
The maturational and experiential origins for potentially distinctive GABA, receptor characteris-
tics in individuals who exhibit heightened aggressive behavior await identification.
1. Introduction
The dissection of the alcohol–aggression link remains a formidable challenge
to the social and neurobiological scientific communities. How cultural and
social conventions, sanctions, licensures, and expectations interact with an
individual’s drinking habits and social behavior and how the individual’s
physiological processes at the systems and cellular level are modified by
alcohol’s action on receptor molecules that are relevant to the drug’s effect on
Klaus A. Miczek • Departments of Psychology, Psychiatry, and Pharmacology, Tufts University,

Medford, Massachusetts 02155. Joseph F. DeBold, Annemoon M.M. van Erp, and Walter
Tornatzky • Department of Psychology, Tufts University, Medford, Massachusetts 02155.
139
aggressive behavior requires integration of information from anthropological,

sociological, psychological, and neurobiological sources. An important char-
acteristic of alcohol in a social context is the observation, portrayed artistically
in Greek symposia, Renaissance bacchanales, party scenes by the Dutch mas-
ters of the Golden Age, or an intimate tete-à-tete by Picasso, that most alcohol
drinking is associated with a convivial mood. Benchmark data on the amount
and pattern of alcohol drinking on the one hand and the incidence of violent
acts on the other are still lacking.1 Conversely, the consistently high and
persistent numbers of aggressive and violent acts that are associated with
alcohol have been documented for decades in several communities through-
out the world, as summarized in the first section of this volume.
The sheer magnitude of the public health problem of how alcohol is
associated with aggressive and violent behavior is overwhelming, in most
epidemiological analyses exceeding more than 50% of all incidences of these
types of behavior.2–4 From the perspective of violence research, it is most
instructive that many different kinds of aggressive and violent behavior are
linked to alcohol, ranging from murders, rapes, sexual violence, wife beat-
ings, child abuse, incestuous offenses, and felonies. In human experimental
research and in clinical studies, acute doses of alcohol have been shown to
increase feelings of hostility and behavioral indices of competitive and retalia-
tory tendencies.5–7 In recent reviews of the world literature on alcohol, drugs
of abuse, aggression, and violence that also encompassed evidence from ex-
perimental laboratory research, alcohol was found to be “the drug that is most
consistently and seriously linked to many types of aggressive and violent
behavior” (ref. 4, p. 389).
In order to gain insight into the precise sequence of events at the behav-
ioral, physiological, and neurobiological levels that characterize alcohol-
heightened aggression, it is necessary to glean from epidemiological and
criminal statistics whether or not the perpetrator or the victim of the violent
and aggressive behavior was under the influence of alcohol. Detailed infor-
mation is required to determine whether or not the alcohol-intoxicated indi-
vidual provokes more aggressive behavior than a sober individual. And if so,
how the alcohol intoxication contributes to the escalation of interactions that
ultimately result in violent and aggressive acts. Epidemiological data rarely
provide insight into these questions and they need to be complemented by
information that stems from situations in which the details of alcohol and the
social confrontation are specified.
Experimental research on alcohol and aggression has the advantages of
being able to (1) specify the precise amount of alcohol that is in the individual
who engages in aggressive behavior, as well as in the victim or target of
aggressive behavior at the relevant times of the initiation, execution, and
termination of the behavioral sequence; (2) delineate acute and chronic alco-
hol conditions; (3) disentangle the issue whether the perpetrator, victim, or
both were under the influence of alcohol; and (4) analyze the proximal and
triggering social events that are influenced by alcohol. Signals of provocation
and appeasement characterize social confrontations in humans and other
7 • GABAA–Benzodiazepine Receptor Complex 141
animal species.8 In the escalation of social confrontations, alcohol appears to

profoundly alter the communication of socially significant signals.9 Detailed
experimental analysis of the nature of the confrontation, the phases of escala-
tion, and the sending and receiving of social signals reveals how alcohol
modifies not only the sensations of sounds, smells, and movements, but
distorts the communicative processes. For example, low alcohol doses, given
orally to dominant squirrel monkeys confronting an opponent, increase ag-
gressive vocalizations concurrently with the display of aggressive acts10; by
contrast, the targets of these vocal and postural displays, when under the
influence of increasing doses of alcohol, are more frequently attacked.11 Rats
that confront an aggressive resident typically emit certain types of ultrasonic
vocalizations that reduce the probability of being attacked; when given in-
creasing doses of alcohol, these vocalizations become less frequent, and it
appears that this disruption of vocal signals may contribute to the higher
frequency of being attacked (K.A. Miczek, unpublished observations).
A major challenge and dilemma for experimental research on alcohol and
aggression continues to be the development and conduct of protocols that
lead to increased aggression that is potentially injurious, harmful, and painful
(e.g., ref. 12). Avoidance and reduction of pain, harm, and injury are the very
principles that govern experimental research. How does one apply these
principles to research protocols in which validity is achieved when the phar-
macological manipulations and behavioral phenomena approximate the
events that constitute the public health problem, namely alcohol-linked vio-
lence? At present, the most rational approach to this type of experimental
research is to evaluate the significance of each protocol in terms of its neuro-
biological, pharmacological, and behavioral validity and to judge its relevance
to the public health problem.
From an experimental viewpoint, one of the most perplexing facets of
alcohol’s impact on aggressive behavior is the enormous variability in out-
come. Even under conditions that control for pharmacological, genetic, situa-
tional, behavioral, and social factors, across individuals alcohol may increase
or decrease or leave aggressive behavior unaltered, and this individual vari-
ability extends from humans to other animal species. The study of individual
alcohol effects on aggressive behavior represents a genuine experimental
challenge necessitating experimental strategies that differ from the practice of
randomly assigning individuals to a test sample and averaging the test results
from all individuals in a sample. Such novel strategies have to capture rela-
tively rare, episodic events in individuals.
2. Alcohol and Aggressive Behavior in Animals:

Ethological Analysis
2.1. Individual Differences in Alcohol Effects on Aggressive Behavior

By now, it has become possible to identify major sources for apparently
opposing conclusions that have been reached about alcohol and aggression in
reviews of epidemiological, experimental human, and preclinical studies in

various animal species. A decade ago, a survey of alcohol effects on animal
aggression prompted the conclusion that alcohol is not a potent aggression-
stimulating drug (e.g., ref. 13). Of course, this conclusion needs to be con-
trasted with the conclusion that alcohol does indeed cause aggression, based
on a meta-analysis of human experimental studies14 and on the overwhelm-
ing epidemiological findings (see Chapters 1 and 2, this volume). How can
one explain these seemingly contradictory conclusions? It is difficult to accept
the notion that animal studies on alcohol and aggression may not be informa-
tive on the human condition.
As in humans, alcohol affects aggressive behavior patterns in individual
animals ranging in species from fish to nonhuman primates that differ from
other individuals in terms of the direction and magnitude of effect. In addi-
tion to the critical alcohol parameters such as dose and time course, it has
become apparent that environmental and behavioral determinants are of
paramount significance in whether or not alcohol enhances aggressive behav-
ior (e.g., refs. 3,4,15). When Siamese fighting fish or Cichlid fish confront a
territorial intruder, when isolated male mice encounter another male mouse,
when dogs or rats compete with a rival, when a resident rat opposes an
intruder into the colony, when a dominant squirrel monkey or macaque
threatens a rival, the administration of low acute doses of alcohol has been
reported to increase these aggressive behaviors (e.g., refs. 16–30). These find-
ings appear robust and impressive on account of their repeated demonstra-
tion in a wide range of animal species, yet only suppressive effects of alcohol on
aggressive behavior were reported in several other studies with mice, rats,
and monkeys (e.g., refs. 31–41). These contrasting effects appear particularly
perplexing since the parameters of alcohol administration and the animal
species are closely similar in reports of increased and decreased aggression.
One source for these apparent inconsistencies derives from the practice of
assigning subjects in experimental studies randomly to treatment groups and
to schedule experimental treatments in a systematically varied sequence. In
fact, most research samples of subjects contain individuals that show quali-
tatively opposite effects of alcohol on aggressive behavior, and pooling data
from individuals with divergent effects prompts the conclusion that lower
alcohol doses do not exert reliable effects on aggressive behavior.
During the past decade, studies with random-bred Swiss-Webster mice and
hooded rats of the Long-Evans strain revealed that approximately 15–25% of the
animals engage in significantly more aggressive behavior after being given low
acute alcohol doses than they exhibit under drug-free conditions.42,43 During
the past two decades, appropriate laboratory procedures have been developed
that engender a behavioral repertoire of agonistic behavior in rats and mice that
represents all salient elements characteristic of the colonial cohesive social
behavior in rats and dispersed territorial behavior of mice.44,45 When given the
opportunity to mark the boundaries of a specific locale and to establish resi-
dence, adult male mice and about two thirds of adult male rats of most common
laboratory strains will investigate, pursue, threaten, and attack an adult intruder
in a species-typical manner.46–49 Intruder mice and rats may initially retaliate,
then attempt to escape, and engage in defensive responses, and when escape is
barred, mice will display a characteristic upright defeat posture or, alternatively,
rats display a submissive supine posture.50
Quantitative ethological methods provide the means to adequately mea-
sure alcohol effects on the initiation, execution, and termination of aggressive
interactions in terms of frequency, duration, and temporal and sequential
pattern. These methods have supplanted the earlier used rating scales of
“aggressiveness” or “agressivity” and simple tallies of whether or not ani-
mals were fighting.3 With the aid of microprocessor-based encoding pro-
grams, each behavioral element such as acts, postures, displays, and vocal
signals during a confrontation between a resident and an intruder or between
social rivals can be measured accurately and precisely in terms of initiation
and termination (e.g., refs. 42,51).
Alcohol, when given in low acute doses, increased aggressive behavior
by resident rats and mice confronting an intruding opponent. These aggres-
sion-heightening effects of alcohol were large and repeatable, but limited to a
subgroup of individuals.42,43 As illustrated in Fig. 1, the alcohol dose–effect
curves for the entire population contained those for two subgroups that had
qualitatively different types of alcohol effects on aggressive behavior both in
rats and in mice. The highest proportion of mice and rats exhibiting height-
ened aggressive behavior was seen after administration of 1.0 g/kg ethanol,
by mouth; typically this subgroup [alcohol-heightened aggression (AHA)]
represents a quarter of animals in any given sample. Several behavioral ele-
ments of aggressive behavior, most prominently attack bites and sideways
threats, are more frequent after alcohol administration, whereas in alcohol-
suppressed animals (ASA) these behaviors are significantly decreased. When
alcohol and water vehicle are administered over the course of alternating test
days, the frequency of attack bites is significantly increased repeatedly in each
individual mouse and rat of the AHA subgroups after the alcohol administra-
tion (Fig. 2). Whether the heightened aggression after alcohol is a phasic,
spikelike phenomenon or a consistent characteristic for a specific individual
has to await more detailed long-term analysis. Preliminary observations have
not revealed any evidence for tolerance or sensitization to the aggression-
heightening effects of alcohol.
Aggressive acts belong to those behavioral and biological functions that
occur in an episodic fashion. In addition to seasonal peaks and troughs, a
microanalysis of the temporal organization of aggressive behavior reveals
bursts or epochs of rapidly succeeding aggressive acts that are separated by
long gaps between bursts. Figure 3 depicts an event record of consecutive
aggressive acts by a resident rat confronting an intruder and, in addition,
summarizes more than 20,000 intervals between consecutive aggressive acts
in a log-survivor plot. It is apparent that more than 85% of all aggressive acts
are separated by very short intervals, i.e., they constitute aggressive bursts,
Figure 1. Effects of alcohol on frequency of attacks. (A) Alcohol dose–effect curve for
the total population (n = 93). (B) Four selected individual alcohol dose–effect curves
showing increases in attack frequency at several alcohol doses. (C) Four selected
individual alcohol dose–effect curves showing suppression of attack behavior at all
alcohol doses. (Data in part from Miczek et al.42)
and the remaining intervals represent the gaps between aggressive bursts. In
experimentally conducted resident–intruder confrontations, administration
of alcohol to the resident aggressive animals lengthened the burst of aggres-
sive behavior without affecting the latency to initiate nor the gaps between
consecutive bursts (Fig. 4). At a behavioral level of analysis, prolonged ag-
gressive bursts may be due to alcohol rendering the individual impaired in
terminating highly energy-demanding exertions, diverting important ener-
gies from other behavioral demands, and unable to recognize signals of sub-
mission and appeasement.
By contrast to the effects on the burstlike characteristics of aggressive
behavior, administration of alcohol did not modulate the intricate sequential
organization of aggressive behavior in resident-intruder confrontations. As
depicted in Fig. 5, a lag sequential analysis of the salient elements of aggres-
sive behavior by resident rats reveals a sequence of the elements “pursuit” →
“sideways threat” → “attack bite” → “aggressive posture” that is two to four
times more probable than chance. Neither low, activating doses of ethanol
nor higher sedating doses altered this high-probability sequence of behavioral
Figure 2. A subgroup of rats (n = 5) and mice (n = 6) that showed enhanced attack

bites upon repeated administration of alcohol. (Top) All mice received 1.0 g/kg alcohol
in both tests. (Bottom) Rats no. 1, 2, and 4 showed peak effects at 0.1 g/kg alcohol, and
rats no. 3 and 5 showed peak effects at 0.3 g/kg alcohol. (a) Individual alcohol and
vehicle tests. (b) Group means for first and second administration of peak aggression-
heightening dose of alcohol. (Data in part from Miczek et al.42)
A. Time Line
min
B. Criterion for Burst Length C. Burst Composition
Figure 3. Temporal analysis of aggressive behavior. (A) The start and end times of all
elements of aggressive behavior (pursuit, sideways threat, attack bite, aggressive pos-
ture) by a resident rat directed toward an intruder in real time during a selected
confrontation. Each behavioral event is depicted as an upward deflection from the
time line. (B) Burst criterion. Log survivor plot of all intervals between consecutive
occurrences of aggressive behavior as a function of interval size. Note the logarithmic
scale on the y-axis. Regression lines are fitted against the steep and flat portions of the
data curve. The intersection of the two regression lines defines the maximal interval
size that is considered to be part of a “burst.” The burst criterion in the present studies
was 6.6 seconds. (C) Definition of “burster.” The ratio between behavioral elements
within an aggressive bursts (i.e., within 6.6 seconds or less since a previous aggressive
element) to those elements outside of the burst criterion. Ratios that are smaller than
10 identify 77 out of 389 5-min confrontations with the least clear burst characteristics.
(Miczek et al.42)
elements in a burst. It is most remarkable that individuals who fight at a rate

more than twice than normal or who are barely capable of fighting emit
aggressive behavior in a sequentially organized fashion.
In primates, individual differences in the effects of alcohol on aggression
have become apparent in studies with multigenerational groups of squirrel
monkeys (Fig. 6). Acute alcohol doses increase a dominant monkey’s aggres-
sive behavior toward rivals within the social group as well as from unfamiliar
groups; however, this effect is not seen in lower-ranking group mem-
bers.10,29,30,52,53 The aggression-heightening effects extend behaviorally from
dominance displays such as bipedal stances and shaking of branches to
threats in the form of chin thrusts and genital displays as well as aggressive
vocalizations.9,10 Increases in aggressive behavior after alcohol did not de-
7 • GABAA–Benzodiazepine Receptor Complex
Figure 4. Alcohol effects on the temporal pattern of aggressive behavior in two AHA rats that show peak in-
creases in parameters of aggression at 0.3 g/kg alcohol. (A) Aggressive events in real time during tests with vehicle,
0.1, 0.3, 1.0 and 3.0 g/kg alcohol. Each upward deflection indicates the occurrence of an aggressive event (i.e.,
pursuit, sideways threat, attack bite, aggressive posture). (B) Alcohol effects on frequency of attacks (solid tri-
147
angles), aggressive elements in bursts (open squares), and time in burst (solid circle) as a function of dose. (Miczek
et al.42)
n=93
Event Lag
Figure 5. Lag sequential analysis of elements of aggressive behavior. Pursuit, sideways threat, attack bite, and aggressive posture
are depicted on the basis of photographs (top). The probability of each specific behavioral element following (lag +1, +2 . . . +n) and
preceding (lag –1, –2 . . . –n) as the first, second, third, fourth, and fifth next element to another specific behavioral element is
shown with the expected level of random sequence (stipled horizontal lines). Each probability is determined according to the rate of its
occurrence. Specifically, the expected mean of the transition from pursuit to threat is 33.0 + 4.4% (mean + 95% confidence levels),
while the actual occurrence of this transition at lag 1 is 68.8%; the expected mean of the transition from threat to attack bites is 13.3 +
0.6%, while the actual occurrence at lag 1 is 32.6%; the expected mean of the transition from attack bites to aggressive posture is 8.4 +
0.8%, while the actual occurrence at lag 1 is 38.5%. (Miczek et al.42)
AGGRESSIVE BEHAVIOR
A. Dose -Effect B. Time Course (0.6 g/kg)
Log Ethanol Dose (g/kg) Time Since Injection (min)
Figure 6. (A) The frequency of aggressive behavior (grasps, display, displacements)

during the 40-min period starting 5 min after alcohol administration to dominant (n =
5) and subordinate (n = 6) members of groups of captive, free-ranging squirrel mon-
keys. Vertical lines in each data point indicate ±SEM. (B) The frequency of aggressive
behaviors measured in consecutive 20-min segments of a 2-hr observation. The data
represent the effects of 0.6 g/kg alcohol on the aggressive behavior of dominant male
squirrel monkeys (n = 5). The shaded area represents the mean ± 1 SEM of five water
vehicle control tests for each of the five dominant monkeys. (Reprinted from Winslow
and Miczek30)
pend on the baseline rate of this behavior in a particular individual monkey;

as a matter of fact, the very low level of aggressive displays in subordinate
monkeys failed to be “disinhibited” by alcohol, whereas the markedly higher
rates of dominant monkeys were even further increased by alcohol. Phar-
macologically, the aggression-heightening effects are seen at low acute eth-
anol doses (0.3, 0.5, 0.6 g/kg, by mouth), and these effects are critically
determined by the circulating levels of testosterone.30
An intriguing question traces the source for the social-status-dependent
alcohol effects in squirrel monkeys and in rats. At present, it is unclear at
which developmental stage alcohol begins to engender aggression-heighten-
ing effects and whether these effects are a life-long characteristic. It is possible
that dominant monkeys or resident rats acquire the sensitivity to the aggres-
sion-heightening effects of alcohol by repeated experiences with aggressive
encounters and by repeated exposures to alcohol. If in fact distinctive experi-
ential factors are critical for the aggression-heightening effects of alcohol, then
it appears feasible to identify such decisive events in the natal group and in
the postpubertal period.
2.2. Animal Models of Alcohol Self-Administration and Aggression

As illustrated in the preceding account, a major advantage of experimen-
tal research on the determinants of the link between alcohol and aggression is
the control over the precise pharmacological conditions of alcohol. This is
traditionally accomplished by administering alcohol at a specific dose and
time. Of course, it is evident from the epidemiological statistics of violent and
aggressive behavior that these acts are committed by alcohol-intoxicated indi-
viduals who have self-administered the drug; this route of administration
needs to be implemented in laboratory studies in order to enhance the validi-
ty of experimental protocols. During the last decade, experimental prepara-
tions have been developed that achieve voluntary intake of alcohol at intox-
icating doses, in both primate and rodent species (e.g., refs. 54,63,73,164).
The application of these methodologies enables a detailed analysis of the
conditions under which alcohol self-administration may lead to increased
aggressive behavior and, vice versa, how situations of social confrontation
influence alcohol self-administration.
Historically, the first approach to the study of social factors on alcohol
self-administration has been to alter experimentally the housing conditions,
comparing alcohol self-administration in singly housed versus pair- or group-
housed animals. In general, individually housed rats have been shown to
drink more than group-housed counterparts.55–57 Yet, social factors have also
been shown to increase alcohol self-administration. For example, Randall and
Lester58 speculated that “social facilitation” or “peer pressure” might explain
why DBA mice, which usually avoid alcohol, consume significantly more of a
10% alcohol solution when housed with C57BL mice, which drink alcohol
more readily. Similarly, cats drink more of a 10% alcohol–milk solution in the
presence of a companion than when singly housed.59 Unfortunately, often
the past and present social behavior of alcohol-drinking group members is not
assessed, rendering conclusions as to the impact of social housing and domi-
nance status uncertain,60–62 whereas the effect of alcohol directly on aggres-
sive behavior was not studied at all.
Studies in primates are even more scarce, chiefly due to difficulties in
implementing alcohol self-administration in primate colonies. For instance,
Crowley and Andrew63 investigated alcohol self-administration in ma-
caques, but did not find a correlation between social status and alcohol con-
sumption. In this valiant effort to induce alcoholiclike drinking, blood alcohol
concentrations (BACs) in excess of 100 mg/dl were achieved only when food
access was scheduled after alcohol availability. Curiously, no ill effects such
as wounding or increased aggression were noted in alcohol drinking individ-
uals. Interestingly, Ervin et al.64 reported increased orientation to external
stimuli, increased incidence of “stereotyped aggression,” and decreased fre-
quency of affiliative behaviors in stable social groups of vervet monkeys dur-

ing periods of alcohol availability as compared to control periods. Stereotyped
aggression was defined as aggressive overtures, including threats, slapping
at, and physical displacement, which were neither responded to nor followed
up by the instigator. Alcohol was presented in increasing concentrations from
7.5 to 20% in a 3% sucrose solution. It appeared that young animals drank
more than older ones, whereas neither sexually mature males nor dominant
females drank substantial quantities of alcohol. Unfortunately, no blood alco-
hol concentrations after self-administration were measured; however, 20–
25% of alcohol-preferring animals were reported to drink occasionally to atax-
ia, while 5–8% drank to unconsciousness on at least one occasion. In another
study, the effects of rearing conditions on alcohol self-administration in rhe-
sus monkeys were established.65 Monkeys that were reared among peers
drank significantly more of an aspartame-sweetened 7% alcohol solution than
monkeys reared with their mothers. Social isolation increased the level of
alcohol consumption in mother-reared monkeys up to a similar level as peer-
reared monkeys. Most monkeys were reported to drink alcohol in sufficient
quantities to demonstrate signs of intoxication with ataxia, sway, and vomit-
ing; BACs measured after intubation with a similar quantity as self-adminis-
tered on the day of greatest intake ranged from about 40 to 380 mg/dl, with a
mean BAC of about 275 mg/dl. In sum, these studies show that alcohol self-
administration can be implemented in group-housed primate species; how-
ever, the effect of social housing and dominance status on self-administration
or aggressive behavior remains unclear.
Other studies in rats tried to delineate a possible link between dominance
status and alcohol self-administration. For instance, Ellison and co-workers66
found that a larger proportion of individuals that had lived in large single-sex
rat colonies under conditions of abundant and varied food consumed 10%
alcohol in preference over water, as compared to single housed animals.
Notably, these increases in alcohol self-administration were observed in a
subpopulation of individuals, many of whom appeared to be less dominant
within the colony.67 In concordance with this observation, it was shown that
subordinate group members in male–female rat colonies lapped at bottles
containing 4 and 8% alcohol more often than dominant group members, and
females did so more often than males.68 Although both studies in rat colonies
addressed the effect of dominance status on alcohol self-administration, they
did not provide actual blood alcohol concentrations, which renders conclu-
sions about individual alcohol intake somewhat tentative. Still, these observa-
tions suggest that lower-ranking individuals, who are often the recipients of
aggression, are more likely to increase alcohol consumption than the individ-
uals displaying aggressive behavior. However, the timing of self-administra-
tion with respect to the experience of “social stress” appears to be of crucial
importance, as many studies have shown that exposure to stress (e.g., foot
shock, restraint, or social stress) reduces alcohol self-administration during
and immediately following stress, and increases in alcohol intake are only
observed after termination of the stress period (e.g., refs. 69–71; van Erp et
al., submitted).
Recently, methodological advances have allowed a more systematic in-
vestigation of orally self-administered alcohol on aggressive behavior in ani-
mals” (van Erp et al., submitted). The key developments are (1) the induction
of concentration-dependent oral alcohol self-administration in laboratory rats
under conditions that preserve the social structure and that avoid body
weight reductions, and (2) the induction of aggressive behavior in laboratory
rats that is very similar to the behavioral repertoire seen by the feral counter-
parts. Under the present conditions, male Long-Evans rats, each housed with
a female, attacked a male intruder in their home cage during 5-min confronta-
tions.49 After establishing a baseline of reliable aggressive behavior toward
intruders during repeated confrontations over the course of several weeks,
“resident” rats were conditioned to drink a 10% alcohol solution during a
daily 15-min access period in their home cage, using a sucrose substitution
technique.73 After ethanol intake stabilized for approximately 10 days, the
resident rats confronted an intruder twice per week, 5 min after consuming
alcohol. Behavioral data from confrontations that followed alcohol self-ad-
ministration were grouped according to the BAC, as measured in samples
that were taken immediately after the confrontation. The contemporary con-
trol level of aggressive behavior for a specific individual animal was assessed
during confrontations with an intruder that were scheduled before or more
than 3 hr after the ethanol access period, i.e., at a time when no BAC was
detectable.
Under these conditions, resident rats drank up to 1.0 g/kg during the
daily 15-min access period, resulting most often in BAC levels of 20–40 mg/dl
with a range of 10–80 mg/dl. As discussed above, earlier studies had found
aggression-heightening effects at lower acute ethanol doses, the focus was on
the ascending limb of the ethanol dose–effect curve, and neither intoxication
nor sedation were seen. As seen previously with experimenter-administered
ethanol,42 aggressive behavior increased after ethanol self-administration in
certain individuals and remained unchanged in others (van Erp et al., submit-
ted). The magnitude of the increase in the frequency and duration of attacks
and threats ranged from 40 to 90% above the control level (see Fig. 7). It is
evident that individual differences in the effect of alcohol on aggressive be-
havior can be studied in rats, using limited-access self-administration and
short-term aggressive confrontations. Although the overall group does not
show an effect of alcohol on aggressive behavior, it appears that there is a
subgroup of individuals that shows enhanced aggression after alcohol self-
administration.
In sum, effects of alcohol on aggressive behavior have been studied in
rodent and primate species using alcohol self-administration. Social factors
under certain conditions may enhance alcohol self-administration and, vice
versa, alcohol has been shown to enhance aggressive behavior in a sub-
Figure 7. Alcohol’s effect on aggressive behavior in resident rats. Rats were assigned
to the alcohol-enhanced aggression (AHA) group (n = 6) or the alcohol-nonenhanced
aggression (ANA) group (n = 6), based on individual behavior during intruder con-
frontations with and without alcohol self-administration preceding the confrontation.
Data are expressed as percentage change of control (confrontations without alcohol).
The AHA group significantly differs from the ANA group for all the behaviors shown
(frequency of aggressive postureion top, sideways threat and bite; duration of aggres-
sive postureion top and total time spent on aggression). Data were analyzed with a
one way analysis of variance (P < 0.05).
population of individuals. The most useful information derives from studies

that establish self-administration in group-housed individuals, while mon-
itoring individual alcohol intake and BACs. In addition, it is important to
simultaneously assess social interactions between individuals in these groups
and identify hierarchies and dominance status, in order to determine the
effect of social factors on alcohol self-administration. Integrating these essen-
tial pharmacological and ethological features is bound to yield the most valid
information on the interaction between aggression and alcohol self-adminis-
tration. In identifying individuals who are more prone to increased alcohol
intake and to increased aggression after alcohol intake, it is possible to assess
pharmacological and neurochemical differences between these individuals

and those that are not prone to those effects.
3. Pharmacological Evidence for Interactions between Alcohol and

γ-Aminobutyric Acid-A Receptors and Aggression
3.1. Benzodiazepine Receptor
Several of alcohol's behavioral effects are linked to its action on subtypes
of serotonin (5-HT), N-methyl-D-aspartate (NMDA), and γ-aminobutyric acid
(GABA) receptors (e.g., refs. 74–76). Alcohol has been shown to potentiate
behavioral, electrophysiological, and biochemical GABAergic responses.77
Here, we focus on the GABAA–benzodiazepine receptor complex as a site of
action for alcohol in modulating aggressive behavior. Benzodiazepines,
which are allosteric modulators of this complex, and alcohol share similar
anxiolytic, amnesic, hypnotic, sedative, anticonvulsant, and ataxic effects.
Further evidence linking ethanol and GABAA comes from studies showing
that pharmacological blockade of the benzodiazepine site on the GABAA–
benzodiazepine receptor complex can antagonize several of ethanol's physi-
ological, behavioral,78–80 and neurochemical effects (e.g., refs. 81,82). In
terms of aggression, mice that were selectively bred for high or low levels of
aggressive behavior could be further differentiated by divergent GABA-de-
pendent chloride uptake into cortical neurons, benzodiazepine receptor bind-
ing, and behavioral effects of chlordiazepoxide treatment.83
3.1.1. Effects of Alcohol on Aggression as Compared to Those of Ben-
zodiazepine Agonists. A closely similar behavioral profile of effects by ben-
zodiazepine receptor agonists and alcohol suggests, at least indirectly, poten-
tially shared mechanisms for aggression-modulating effects (e.g., ref. 84). In
the first decade of benzodiazepine research, alcohol and benzodiazepines
were shown to reduce aggressive behavior, mostly due to their sedative and
motor incoordinating properties at intermediate to higher doses in many
animal species.85–88 This evidence supported the characterization of ben-
zodiazepines as antiaggressive drugs that could “tame” even feral animals.
Exploring a wider dose range, both chlordiazepoxide and diazepam, on
the one hand, and alcohol on the other were found to share biphasic effects
on aggressive behavior, with higher doses decreasing the frequency of threats
and attacks and lower doses increasing these behaviors (e.g., refs.
19,21,89,90). For example, during confrontations between a resident and an
intruder in rats and mice, the frequency of pursuits, attack bites, threatening
acts, and postures was increased following administration of low acute alco-
hol doses ranging from 0.3 to 1.0 g/kg orally, while doses above 1.7 g/kg
decreased these behaviors as well as prolonged inactivity (see Fig. 10).21,53,91
Similarly, low doses of chlordiazepoxide (<10 mg/kg) and diazepam (0.3
mg/kg) increase the frequency of several behavioral elements of the aggres-
sive repertoire, and higher doses have suppressive effects on these acts and
postures as well as on all active elements of motor activity.90,92
However, not all benzodiazepine receptor agonists engender identical
profiles of effects on aggressive, defensive, flight, and social behavior in
rodents or primates, including humans. Alprazolam or oxazepam, two des-
chloro-phenyl derivatives, seem to be more selective in their reduction in
defensive and escape activities and also in their antiaggressive effects in mice
(e.g., ref. 92). This selective antiaggressive effect of alprazolam (>0.3 mg/kg,
intraperitoneal) was confirmed in resident mice tested in an experimental
protocol93 that was designed to dissociate between effects on conditioned
performance and aggressive behavior following the same drug treatment.
Moreover, the same mice showed large increases in aggressive behavior after
administration of low acute alcohol doses (e.g., 1.0 g/kg, orally).43
Recently, a comparison of the effects of chlordiazepoxide (0.3–10 mg/kg)
and alcohol (0.1–1.0 g/kg) on vocalizations and threat displays during aggres-
sive confrontations between rival dominant male squirrel monkeys showed a
similar result. As illustrated in Fig. 8, threat peeps, the most prominent
squirrel monkey call in social contexts, and aggressive displays such as genital
threats, bipedal stances, and branch shakes were increased by low doses of
alcohol and chlordiazepoxide.10 Aggressive displays were in turn reduced by
the moderate-to-high ataxic alcohol doses, but neither alcohol nor chlor-
diazepoxide reduced aggressive vocalizations in this behaviorally demanding
situation.
These similar effects of alcohol and benzodiazepines on offensive-type
aggressive behavior in different species may suggest shared mechanisms of
action between these drugs. In order to test this possibility, it would be
necessary to systematically investigate the interactions between alcohol and
various types of benzodiazepine receptor agonists for their mutually poten-
tiating effects on aggressive behavior. In male resident mice confronting an
intruder, alcohol or chlordiazepoxide had no significant effects on aggressive
behavior in nonsedative doses.21 However, when attack rates were sup-
pressed by conducting the agonistic confrontation in a neutral cage, a low
dose of alcohol (0.3 g/kg orally) more than doubled the frequency of attacks
and threats. In this latter condition, alcohol's proaggressive effect was further
enhanced by chlordiazepoxide (5.0 mg/kg orally) pretreatment. These obser-
vations, if substantiated by systematic leftward shifts in the alcohol dose–
effect curve, would in fact be consistent with the proposal of a common site of
action for the aggression-heightening effects of alcohol and benzodiazepines.
3.1.2. Interactions with Benzodiazepine Receptor Antagonists. More direct

support for the GABAA–benzodiazepine receptor complex as a site of action
for alcohol's aggression-heightening effects can be accrued from studies with
benzodiazepine receptor antagonists? The rate of several aggressive acts and
postures of dominant squirrel monkeys and resident male rats during the
dyadic confrontations described above increases after administration of low
A. Threat Peeps
B. Aggressive Displays
Figure 8. Effects of chlordiazepoxide and ethyl alcohol on (A) rate of threat peeps
during the first minute of aggressive confrontations (B) frequency of aggressive behav-
iors. Asterisk indicate significant (P < 0.05) differences when compared to vehicle
control. (Reprinted with permission from Weerts et al.10)
acute alcohol doses (0.1–0.3 g/kg, orally). In a series of experiments, the

alcohol dose–effect curves were redetermined after rats and monkeys were
pretreated with one of two benzodiazepine receptor antagonists, ZK 93426 (3
mg/kg) or flumazenil (10 mg/kg). Figure 9 illustrates that both antagonists,
the imidazobenzodiazepine as well as the β-carboline derivative, selectively
prevented the aggression-heightening effects of lower alcohol doses, without
having an effect on the aggression-decreasing and sedative effects of higher
alcohol doses.
However, quantitative ethological analysis of the salient elements in the
behavioral repertoire of socially housed squirrel monkeys revealed subtle
differences in the way both benzodiazepine receptor antagonists modulated
the sedative and motor-incoordinating effects of alcohol. Flumazenil, given
before low alcohol doses (0.1–0.3 g/kg), actually decreased active locomotion
and increased inactive postures to the low levels that are typically seen after
higher alcohol doses (1.5 g/kg).53 Flumazenil administration before high alco-
hol doses (1.5 g/kg) did not antagonize the characteristic motor incoordina-
tion, as measured by the incidence of staggering, but further intensified these
effects. By contrast, ZK 93426 pretreatment prevented motor incoordination
by higher alcohol doses. These data suggest a remarkably specific antagonism
of alcohol’s aggression-heightening effects, as distinct from other behavioral,
i.e., sedative or motor-incoordinating effects, by the two benzodiazepine re-
ceptor antagonists.
The intrinsic effects of benzodiazepine receptor antagonists are of consid-
erable interest in the interpretation of the interactive effects between alcohol
and these substances. In a primate species such as squirrel monkeys that is
particularly sensitive to the effects of GABAA–benzodiazepine receptor ma-
nipulations, it is apparent that benzodiazepine receptor antagonists can en-
gender readily behavioral effects such as increased food intake or decreased
social interactions that have been interpreted to reflect either “anxiolytic” or
“anxiogenic” properties.94 Specific doses of ZK 93426 and flumazenil can
decrease aggressive and social behavior in rats and in squirrel monkeys.94–98
For example, higher doses of flumazenil (10 mg/kg) reduced isolation-in-
duced aggression in mice, suggesting a partial benzodiazepine agonistlike
activity of the compound.99 Considering the continuously ongoing regulation
of the GABAA–benzodiazepine receptor complex by endogenous and exog-
enous factors, it may not be at all surprising that antagonist drugs can pro-
duce effects that are either similar to partial agonist or partial inverse agonist
substances. These observations caution against any simple modulation of
alcohol effects on aggressive behavior by benzodiazepine receptor antagonist.
3.1.3. Interactions with Partial Benzodiazepine Agonists. The benzodiaze-

pine receptor partfal inverse agonist Ro15-4513, introduced in the mid-1980s,
has been studied for its prevention or reversal of physiological and behavioral
alcohol effects.100 Alcohol’s sedative, ataxic, muscle relaxant, and hypnotic
effects were demonstrated to be at least partially and transiently reduced by
a. In Dyad Test
b. In Colony
Dose (g/kg)
Figure 9. (A) Effects of alcohol (ETOH) and ZK 93426 (3 mg/kg) pretreatment on
aggressive threats and displays in dominant male squirrel monkeys (n = 6) in dyadic
confrontations. (B) Effects of ETOH and flumazenil (10 mg/kg) pretreatment on ag-
gressive threats, grasps, and displays in dominant male squirrel monkeys (n = 5)
directed toward untreated group members. * Represents P < 0.05 compared to vehicle
control. ** Represents P < 0.05 compared to vehicle control and the same dose of
ETOH alone. (Reprinted from Weerts et al.53)
Ro15-4513.78,80,101,102 Pretreatment with benzodiazepine receptor inverse ago-

nists reduces the enhancing effects of alcohol on behavior that is suppressed
by bright light103,104 or electric shock.103–106 Alcohol-induced reductions in
exploratory motor behaviors107,108 and social interactions between two famil-
iar rats in a novel brightly lit arena28 were prevented by inverse agonists.
However, in attempts to antagonize the aggression-heightening effects of
alcohol in squirrel monkeys, Ro15-4513 pretreatment produced tremors and
seizures.53,109 Similarly, Ro15-4513 also potentiates seizures in mice after alco-
hol withdrawal and leads to electrophysiological indices of seizurelike activity
in rats.110,111 The risk of inducing epileptogenic activity in a significant pro-
portion of individuals renders benzodiazepine receptor partial inverse ago-
nists such as Ro15-4513 unusable and the entire strategy as problematic for
eventual clinical development.
In sum, there is evidence for the GABAA–benzodiazepine receptor com-
plex as a major site that differentiates highly aggressive animals from less
aggressive individuals.83 Additionally, there are data that suggest that this
site is mainly involved in the potentiation or the reversal of alcohol's aggres-
sion-heightening effects by benzodiazepine agonists or antagonists, respec-
tively.
A cautionary note is in order: the GABAA–benzodiazepine receptor com-
plex interacts with monoamines, steroids, and peptides, which in turn are
influenced by many endogenous and exogenous factors. Considering the fact
that alcohol does not act only at GABAA receptors, it remains to be deter-
mined if alcohol's interactions with the GABAA–benzodiazepine receptor
complex is the most important component of alcohol's proaggression effect.
3.2. Alcohol Interactions with Neurosteroids

3.2.1. Neurosteroids. Another class of compounds that can affect the GABAA
receptor complex are certain C21 steroids. Although the genomic mechanism
of steroid action mediated by intracellular receptors has been known for over
30 years,112 recently it has been demonstrated that some steroids can act
extracellularly at the GABAA receptor complex. Some of these GABAA-active
steroids have been found to be allosteric positive modulators of the GABAA
receptor complex in vitro, while others are negative modulators. Among the
most well documented of the positive modulators is the naturally occurring
metabolite of progesterone, allopregnanolone (5α-pregnan-3α-ol-20-one).
This steroid has been shown to competitively inhibit [35S] t-butylbicy-clophos-
phorothionate (TBPS) from binding to the GABAA chloride ion channel in
brain. 113–115 Additional research with rat cortical synaptoneurosomes has
shown that allopregnanolone also increases the binding of the GABAA ago-
nist [3H]muscimol,116 and nanomolar concentrations of the steroid potenti-
ates the stimulatory effect of muscimol on 36Cl– flux.117 Moreover, this is a
stereospecific effect, since its 3β isomer is inactive.118 In addition to increasing
muscimol binding, allopregnanolone enhances the binding of the benzo-
diazepine agonist flunitrazepam.113,117 Another example of a steroid that pos-
itively modulates the GABAA complex is the synthetic steroid alphaxalone.

This compound had been used for its anesthetic properties, but it was also
one of the first steroids shown to acutely enhance [3H]muscimol binding in
rat brain membrane preparations.119
Behaviorally, steroids like allopregnanolone and alphaxalone have also
been shown to share effects with other positive modulators of GABAA such as
benzodiazepines, barbiturates, and ethanol. For example, they have anxioly-
tic,120–124 anticonvulsant,125 anesthetic,126–128 and analgesic actions129,130
when administered to rats. These effects have also been found to be quite
potent. Allopregnanolone has anxiolytic effects in the elevated plus-
maze123,131 and on lick suppression,124 at dosages of 10–20 mg/kg IP. The
same dose range protects mice from pentylene tetrazole-induced seizures,132
and loss of the righting reflex has been reported to have an ED50 of 1.05
mg/kg with intravenous administration.133 The same effects are generally
reduced by GABAA antagonists. For example, bicuculline reduces the analge-
sic effects of allopregnanolone.134
However, unlike with ethanol and benzodiazepines, there has been very
little work on the effects of GABAA-active steroids on aggression. The few
papers that have appeared on this topic report effects that are consistent with
the other GABAA-active compounds. For example, we have found that acute
administration of allopregnanolone to male mice has biphasic effects on ag-
gression that are quite similar to those of ethanol.135 As shown in Fig. 10, 10
mg/kg allopregnanolone was seen to increase the frequency of the salient
components of male mouse aggressive behavior, while this effect disappeared
at higher doses, and at 30 mg/kg, allopregnanolone reduced aggressive be-
havior. Such a biphasic dose–effect curve is also very similar in pattern to the
behavioral responses to ethanol and to diazepam (Fig. 10). The close sim-
ilarity in the dose-effect curves of these three drugs extends the proposal
presented earlier in this chapter that their shared effects on the GABAA recep-
tor complex underlies their actions on aggression.
Allopregnanolone is not the only neuroactive steroid with effects on
aggression. Others have found that a negative modulator of GABAA, dehy-
droepiandrosterone, inhibits aggression in either male or female mice toward
lactating females.136,137
3.2.2. Interactions with Ethanol. Given that ethanol and these steroids have
converging actions on the GABAA–benzodiazepine receptor complex,76 it
should not be surprising that when administered jointly, the effects of either
compound are altered. For example, Büküsoglu, Thalhammer, and Kriegerr133
evaluated the role of the interactive effects of ethanol and allopregnanolone in
a test for analgesic loss of the righting response (LRR). They found that when
male albino mice were administered ethanol (0.5 and 1.1 g/kg) and allopreg-
nanolone (0.3–3.0 mg/kg) intravenously, ethanol enhanced the steroids anal-
gesic effects in a dose-dependent manner. In addition, 5β-pregnan-3α-01-20-
one, another positive modulator of GABAA, enhances the effects of ethanol
on locomotion, body temperature, and sleep duration in mice.138
7 • GABAA-Benzodiazepine Receptor Complex 161
Rat
Mouse
Dose (mg/kg)
Figure 10. Changes in the frequency of attack bites, expressed as percent of control
(100% = control, dashed horizontal line), as a function of dose diazepam (DZP) and
ethanol (ETOH) in resident rats confronting an intruder (top), and of allopreg-
nanolone and ethanol (ETOH) in resident mice confronting an intruder (bottom).
Vertical lines in the data points indicate ± 1 SEM.
However, neuroactive steroids do not always interact with ethanol ef-

fects in ways that would be predicted based on the results of in vitro measures
of their actions on the GABAA complex. Some of the steroids that act as
negative modulators of GABAA in membrane preparations have not been
found to reduce the effects of ethanol. For example, pregnenolone sulfate
does not reduce the anticonvulsant or hypnotic effects of ethanol as would be

predicted based on its ability to reduce the effects of GABA on Cl– flux.138 In
fact, pregnenolone sulfate actually enhances ethanol-induced hypothermia.
Another negative modulator of GABAA, dehydroepiandrosterone sulfate, has
also been reported to enhance ethanol-induced hypothermia.139 Some of
these unexpected effects may be due to in vivo metabolic conversion of the
administered steroid into another steroid form with different activities. Rapid
metabolism of neuroactive steroids to less active forms has also been used to
explain the far greater analgesic potency of intraventricularly administered
allopregnanolone in comparison to peripheral administration.140 In addition,
some of the steroids that are negative modulators of GABAA also interact with
other neurotransmitter receptors. For example, pregnenolone sulfate has
been shown to potentiate NMDA-mediated cellular responses in vitro141,142
and to affect glycine receptors.143 Ethanol can also alter these receptors in at
least some neurons,144 but it is less clear if these are the same neurons that are
responsive to pregnenolone sulfate. These additional effects may complicate
the behavioral outcome of modulating alcohol effects by neuroactive steroids:
4. Neurochemical Mechanisms for Alcohol–GABAA

Interactions and Aggression
Ideally, evidence for a critical role of GABAA receptors in the aggression-
heightening effects of alcohol should be derived from direct measurements of
GABAA receptor functions in individuals who exhibit increased aggressive
behavior while under the influence of alcohol. Unfortunately, such data are
currently not available, and the role of GABAA receptors has to be inferred
from studies that (1) demonstrate alcohol actions on these receptors at rele-
vant doses, (2) implicate GABA in the regulation of relevant aggressive be-
havior patterns, and (3) correlate alcohol effects on aggressive behavior with
those of the effects of modulators and direct agonists and antagonists at
GABAA receptors.
An attractive and inviting reason for implicating GABAA receptors in the
actions of alcohol is the evidence demonstrating chloride flux changes due to
alcohol in cortical synaptosomes at nanomolar concentrations that may corre-
spond to the oral doses that increase aggressive behavior.78.145–147 So far, it is
not clear whether or not individual differences in the in vitro chloride flux
measurements at increasing alcohol concentrations actually exist and whether
such differences would be relevant to the aggression-heightening effects.
However, consistent with such reasoning is evidence from mice that were
selected to be highly aggressive and that differ from those that were selected
for a low level of aggressive behavior in several neurochemical systems,148
most notably in muscimol-activated chloride flux in cortical tissue.83
As the two most apparent alternative possibilities, one can conceptualize
the GABAA receptors as being either in parallel with several concurrently
activated receptors or as the initiation site for a cascade of subsequent mecha-

nisms that are required for the aggression-heightening effects of alcohol.
Pharmacological strategies have been of limited use in deciding between
GABAA receptors' place in sequential or parallel models, since blockade of
many receptor types can modify alcohol's behavioral effects, including those
on aggressive behavior. For example, in addition to the earlier-discussed
benzodiazepine receptor antagonists, antagonists at opioid, serotonergic, do-
paminergic, and noradrenergic receptors can block alcohol's behavioral ef-
fects (e.g., refs. 149–151). Yet, so far, only modulators of the GABAA receptor
such as benzodiazepines, barbiturates, and neuroactive steroids mimic the
biphasic dose–effect function of alcohol on aggressive behavior and actually
can potentiate the effects of alcohol on these types of behavior.152 To decide
whether or not modulatory sites of a subpopulation of GABAA receptors are
indeed one of the critical sites for intervention in reducing alcohol-heightened
aggressive behavior has to await the development of more specific phar-
macological tools.
An important characteristic of the GABAA receptor complex is its protein
structure, most commonly being composed of two α, two β, and one γ sub-
unit, but there is variety in the identity of each subunit in this mix. This
heterogeneity in subunit composition has pharmacological implications. For
example, the γ2L subunit has been reported to be required for ethanol sensi-
tivity,153 but the identity of the α and β subunits may also affect re-
sponse.144,154 In addition, subunit composition of the GABAA receptor not
only can differentially affect sensitivity to different compounds, but subunit
composition also varies across brain regions.155–157 For example, the effect of
ethanol on neural activity also varies across brain regions.158 This may have
implications for the effects of ethanol on aggression. Some brain regions have
been specifically implicated in particular forms of aggressive behavior in labo-
ratory animals.159 It is possible that regional differences in GABAA subunit
composition and in ethanol, benzodiazepine, and steroid sensitivity may pro-
vide a way to refine where alcohol acts in the brain to affect aggression.
There is some evidence tracing individual differences in physiological
and behavioral alcohol effects to differences in GABAA receptor activation.
For example, mice that have been selected to be resistant to alcohol with-
drawal seizures differ from those who are prone to these seizures in terms of
binding and functional characteristics of GABAA receptors.160 Whether or not
individual differences in increased aggressive behavior after alcohol can be
linked to distinctive characteristics in subpopulations of GABAA receptors
remains an experimental challenge.
One of the obvious problems in manipulating directly GABAA receptors
in order to alter alcohol's aggression-heightening effects is the pervasive ef-
fects on many additional behavioral and physiological functions that range
from convulsive to motor disorders.161–163 Targeting modulatory sites on cer-
tain GABAA receptors with benzodiazepines, barbiturates, and neuroactive
steroids may offer a wider range of tools to modify ethanol effects on aggres-
sive behavior with fewer side effects than direct GABA receptor interven-
tions.
ACKNOWLEDGMENTS. The experimental research from the authors’ laboratory
was supported by USPHS research grants AA 05122 and DA 02632. We are
grateful to Dr. H. T. Barros and J. T. Sopko for their excellent assistance.
References
1. Reiss AJ Jr, Roth JA: Understanding and Preventing Violence. Washington, DC, National Acad-
emy Press, 1993.
2. Miczek KA, Haney M, Tidey J, et al: Neurochemistry and pharmacotherapeutic manage-
ment of violence and aggression, in Reiss AJ, Miczek KA, Roth JA (eds): Understanding and
Preventing Violence: Biobehavioral Influences on Violence. Washington, DC, National Academy
Press, 1994, vol 2, pp 244–514.
3. Miczek KA: The psychopharmacology of aggression, in Iversen LL, Iversen SD, Snyder SH
(eds): Handbook of Psychopharmacology. New York, Plenum, 1987, vol 19, pp 183–328.
4. Miczek KA, DeBold JF, Haney M, et al: Alcohol, drugs of abuse, aggression and violence, in
Reiss AJ, Roth JA (eds): Understanding and Preventing Violence: Social Influences on Violence.
Washington, DC, National Academy Press, 1994, vol 3, pp 377-570.
5. Cherek DR, Steinberg JL, Manno BR: Effects of alcohol on human aggressive behavior.
J Stud Alcohol 46:321–328, 1985.
6. Shuntich RJ, Taylor SP: The effects of alcohol on human physical aggression. J Exp Res
Person 6:34–38, 1972.
7. Zeichner A, Pihl RO: Effects of alcohol and behavior contingencies on human aggression. J
Abnorm Psychol 88:153–160, 1978.
8. Huntingford FA, Turner AK: Animal Conflict. London, Chapman and Hall, 1987.
9. Pedigo NW, Yamamura HI, Nelson DL: Discrimination of multiple 3H-5-hydroxytrptamine
binding sites by the neuroleptic spiperone in rat brain. J Neurochem 36:220–226, 1981.
10. Weerts EM, Miczek KA: Primate vocalizations during social separation and aggression:
Effects of alcohol and benzodiazepines. Psychopharmacology, 127:255–264, 1996.
11. Miczek KA, Winslow JT, DeBold JF: Heightened aggressive behavior by animals interacting
with alcohol-treated conspecifics: Studies with mice, rats and squirrel monkeys. Pharmacol
Biochem Behav 20:349–353, 1984.
12. Huntingford FA: Some ethical issues raised by studies of predation and aggression. Anim
Behav 32:210–215, 1984.
and aggression in infra-human animals, in Brain PF (ed): Alcohol and Aggression. London,
Croom Helm, 1986, pp 84–137.
14. Bushman BJ, Cooper HM: Effects of alcohol on human aggression: Integrative research
review. Psychol B 107:341–354, 1990.
15. Cherek DR, Steinberg JL: Effects of drugs on human aggressive behavior, in Burrows GD,
Werry JS (eds): Advances in Human Psychopharmacology. Greenwich, CN, JAI Press, 1987, vol
4, pp 239–290.
16. Raynes AE, Ryback RS: Effect of alcohol and congeners on aggressive response in Betta
splendens. Q J Stud Alcohol 5:130–135, 1970.
17. Ellman GL, Herz MJ, Peeke HVS: Ethanol in a cichlid fish: Blood levels and aggressive
behavior. Proc West Pharmacol Soc 15:92–95, 1972.
18. Chance MRA, Mackintosh JH, Dixon AK: The effects of ethyl alcohol on social encounters
between mice. J Alcohol 8:90–93, 1973.
19. Miczek KA, Barry H III: Effects of alcohol on attack and defensive-submissive reactions in
rats. Psychopharmacology 52:231–237, 1977.
20. Pettijohn TF: The effects of alcohol on agonistic behavior in the Telomian dog. Psychophar-
macology 60:295–301, 1979.
21. Miczek KA, O’Donnell JM: Alcohol and chlordiazepoxide increase suppressed aggression in
mice. Psychopharmacology 69:39–44, 1980.
22. Peeke HVS, Cutler L, Ellman G, et al: Effects of alcohol, cogeners, and acetaldehyde of
aggressive behavior of the convict cichlid. Psychopharmacology 75:245–247, 1981.
23. Yoshimura H: Pharmaco-ethological analysis of agonistic behavior between resident and
intruder mice: Effects of ethyl alcohol. Folia Pharmacol Japonica 81:135–141, 1983.
24. DeBold JF, Miczek KA: Testosterone modulates the effects of ethanol on male mouse aggres-
sion. Psychopharmacology 86:286–290, 1985.
25. Blanchard RJ, Hori K, Flannelly K, Blanchard DC: The effects of ethanol on the offense and
defensive behaviors of male and female rats during group formation. Pharmacol Biochem
Behav 26:61–64, 1987.
26. Blanchard KJ, Flannelly KJ, Hori K, Blanchard DC: Ethanol effects on female aggression
vary with opponent size and time within session. Pharmacol Biochem Behav 27:645–648, 1987.
27. Blanchard RJ, Hori K, Blanchard DC, Hall J: Ethanol effects on aggression of rats selected for
different levels of aggressiveness. Pharmacol Biochem Behav 27:641–644, 1987.
28. Lister RG, Hilakivi LA: The effects of novelty, isolation, light and ethanol on the social
behavior of mice. Psychopharmacology 96:181–187, 1988.
29. Winslow JT, Ellingboe J, Miczek KA: Effects of alcohol on aggressive behavior in squirrel
monkeys: Influence of testosterone and social context. Psychopharmacology 95:356–363, 1988.
30. Winslow JT, Miczek KA: Androgen dependency of alcohol effects on aggressive behavior: A
seasonal rhythm in high-ranking squirrel monkeys. Psychopharmacology 95:92–98, 1988.
31. Krsiak M, Borgesova M: Effect of alcohol on behaviour of pairs of rats. Psychopharmacologia
32:201–209, 1973.
32. Crowley TJ, Stynes AJ, Hydinger M, Kaufman IC: Ethanol, methamphetamine, pentobarbi-
tal, morphine, and monkey social behavior. Arch Gen Psychiatry 31:829–838, 1974.
33. Lagerspetz KMJ, Ekqvist K: Failure to induce aggression in inhibited and in genetically non-
aggressive mice through injections of ethyl alcohol. Aggressive Behav 4:105–113, 1978.
34. Smoothy R, Bowden NJ, Berry MS: Ethanol and social behaviour in naive Swiss mice.
35. Smoothy R, Berry MS: Effects of ethanol on behavior of aggressive mice from two different
strains: A comparison of simple and complex behavioral assessments. Pharmacol Biochem
Behav19:645–653, 1983.
36. Smoothy R, Berry MS, Brain PF: Acute influences of ethanol on murine social aggression:
Effects of dose, strain and fighting experience (Abstract). Aggressive Behav 9:119–120,
1983.
37. Benton D, Smoothy R: The relationship between blood alcohol levels and aggression in
mice. Physiol Behav 33:757–760, 1984.
38. Smoothy R, Berry MS: Effects of ethanol on murine aggression assessed by biting of an
inanimate target. Psychopharnracology 83:268–271, 1984.
39. Everill B, Berry MS: Effects of ethanol on aggression in three inbred strains of mice. Physiol
Behav39:45–51, 1987.
40. Berry MS: Ethanol-induced enhancement of defensive behavior in different models of mur-
ine aggression. J Stud Alcohol (Suppl 11):156–162, 1993.
41. Mos J, Olivier B: Differential effects of selected psychoactive drugs on dominant and subor-
dinate male rats housed in a colony. Neurosci Res Commun 2:29–36, 1988.
42. Miczek KA, Weerts EM, Tornatzky W, et al: Alcohol and “bursts” of aggressive behavior:
Ethological analysis of individual differences in rats. Psychopharmacology 107:551–563, 1992.
43. Wheatley MD: The hypothalamus and affective behavior in cats. Arch Neurol Psychiatry
52:296–316, 1944.
44. Barnett SA: The Rut. A Study in Behavior. Chicago, University of Chicago Press, 1975.
45. Crowcroft P: Mice All Over. London, Foulis, 1966.
46. Blanchard RJ, Blanchard CD: Aggressive behavior in the rat. Behav Biol 21:197–224, 1977.
47. Brain PF: Differentiating types of attack and defense in rodents, in Brain PF, Benton D (eds):
Multidisciplinary Approaches to Aggression Research, Amsterdam, Elsevier, 1981, pp 53–78.
48. Miczek KA, O’Donnell JM: Intruder-evoked aggression in isolated and nonisolated mice:
Effects of psychomotor stimulants and L-dopa. Psychopharmacology 57:47–55, 1978.
49. Miczek KA: A new test for aggression in rats without aversive stimulation: Differential
effects of D-amphetamine and cocaine. Psychopharmacology 60:253–259, 1979.
50. Miczek KA, Thompson ML, Shuster L: Opioid-like analgesia in defeated mice. Science
215:1520–1522, 1982.
51. Miczek KA: Ethological analysis of drug action on aggression, defense and defeat, in Spieg-
elstein MY, Levy A (eds): Behavioral Models and the Analysis of Drug Action. Amsterdam,
Elsevier, 1982, pp 225-239.
52. Winslow JT, Miczek KA: Social status as determinant of alcohol effects on aggressive behav-
ior in squirrel monkeys (Saimiri sciureus). Psychopharmacology 85:167–172, 1985.
53. Weerts EM, Tornatzky W, Miczek KA: Prevention of the proaggressive effects of alcohol by
benzodiazepine receptor antagonists in rats and in squirrel monkeys. Psychopharmacology
111:144–152, 1993.
54. Meisch RA, Carrol ME: Establishment of Orally Delivered Drugs for Rhesus Monkeys: Some
Relations to Human Drug Dependence. Washington, DC, US Government Printing Office,
1981.
55. Deatherage G: Effect of housing density on alcohol intake in the rat. Physiol Behav 9:55–57,
1972.
56. Parker LF, Radow BL: Isolation stress and volitional ethanol consumption in the rat. Physiol
Behav 12:1–3, 1974.
57. Wolffgramm J: Free choice ethanol intake of laboratory rats under different social condi-
tions. Psychopharmacology 101:233–239, 1990.
58. Randall CL, Lester D: Social modification of alcohol consumption in inbred mice. Science
189:149–151, 1975.
59. Wyrwicka W, Long AM: The effect of companion on consumption of ethanol solution in
cats. Pavlovian J Biol Sci 18:49–53, 1983.
60. Heminway DA, Furumoto L: Population density and alcohol consumption in the rat. Q J
Stud Alcohol 33:794–799, 1972.
61. Kulkosky PJ, Zellner DA, Hyson RL, Riley AL: Ethanol consumption of rats in individual,
group, and colonial housing conditions. Physiol Psychol 8:56–60, 1980.
62. Rockman GE, Gibson JEM: Effects of duration and timing of environmental enrichment on
voluntary ethanol intake in rats. Pharmacol Biochem Behav 41:689–693, 1992.
63. Crowley TJ, Andrews AE: Alcoholic-like drinking in simian social groups. Psychopharmacol-
ogy 92:196–205, 1987.
64. Ervin FR, Palmour RM, Young SN, et al: Voluntary consumption of beverage alcohol by
vervet monkeys: Population screening, descriptive behavior and biochemical measures.
Pharmacol Biochem Behav 36:367–373, 1990.
65. Higley JD, Hasert MF, Suomi SJ, Linnoila M: Nonhuman primate model of alcohol abuse-
Effects of early experience, personality, and stress on alcohol consumption. Proc Nat Acad
Sci USA 88:7261–7265, 1991.
66. Ellison G, Daniel F, Zoraster R Delayed increases in alcohol consumption occur in rat
colonies but not in isolated rats after injections of monoamine neurotoxins. Exp Neurol
65:608–615, 1979.
67. Ellison G: Stress and alcohol intake: The socio-pharmacological approach. Physiol Behav
40:387–392, 1987.
68. Blanchard RJ, Hori K, Tom P, Blanchard DC: Social structure and ethanol consumption in
the laboratory rat. Pharmacol Biochem Behav 28:437–442, 1987.
69. Anisman H, Waller TG: Effects of inescapable shock and shock-produced conflict on self-
selection of alcohol in rats. Pharmacol Biochem Behav 2:27–33, 1974.
70. Mills KC, Bean JW, Hutcheson JS: Shock-induced ethanol consumption in rats. Pharmacol
Biochem Behav 6:107–115, 1977.
71. Rawleigh JM, Kushner MG, Fiszdon J, Carroll ME: The effects of restraint-stress on volun-
tary ethanol consumption in rats (submitted).
72. Van Erp AMM, Samson HH, Miczek KA: Alcohol self-administration, dopamine and ag-
gression in rats. Neurosci Abstr 20:1614, 1994.
73. Samson HH: Initiation of ethanol reinforcement using a sucrose-substitution procedure in
food- and water-sated rats. Alcohol Clin Exp 10:436–442, 1986.
74. Deitrich RA, Dunwiddie TV, Harris RA, Erwin VG: Mechanism of action of ethanol: Initial
central nervous system actions. Pharmacol Rev 41:489–537, 1989.
75. Tabakoff B, Hoffman PL: Biochemical pharmacology of alcohol, in Meltzer HY (ed): Psy-
chopharmacology: The Third Generation of Progress. New York, Raven Press, 1987, pp 1521–
1526.
76. Lueddens H, Korpi ER: Biological function of GABAA/benzodiazepine receptor hetero-
geneity. J Psychiatr Res 29:77–94, 1995.
77. Korpi ER: Role of GABAA receptors in the actions of alcohol and in alcoholism: Recent
advances. Alcohol Alcoholism 29:115–129, 1994.
78. Suzdak PD, Glowa JR, Crawley JN, et al: A selective imidazobenzodiazepine antagonist of
ethanol in the rat. Science 234:1243–1247, 1986.
79. Deacon RMJ, Buddhram P, Thompson TA, Gardner CR: Differential interactions of Ro
15-4513 with benzodiazepines, ethanol and pentobarbital. Eur J Pharmacol 180:283–290, 1990.
80. Syapin PJ, Jones BL, Kobayashi LS, et al: Interactions between benzodiazepine antagonists,
inverse agonists, and acute behavioral effects of ethanol in mice. Brain Res B 24:705–709,
1990.
81. Mehta AK, Ticku MK: Ethanol potentiation of GABAergic transmission in cultured spinal
cord neurons involves gamma-aminobutyric acidA-gated chloride channels. Journal of Phar-
macology and Experimental Therapeutics 246:558–564, 1988.
82. Harris RA, Allan AM, Daniell LC, Nixon C: Antagonism of ethanol and pentobarbital
actions by benzodiazepine inverse agonists: Neurochemical studies. Journal of Pharmacology
and Experimental Therapeutics 247:1012–1017, 1988.
83. Weerts EM, Miller LG, Hood KE, Miczek KA: Increased GABAA-dependent chloride uptake
in mice selectively bred for low aggressive behavior. Psychopharmacology 108:196–204, 1992.
84. Miczek KA, Krsiak M: Drug effects on agonistic behavior, in Thompson T, Dews PB (eds):
Advances in Behavioral Pharmacology. New York, Academic Press, 1979, vol 2, pp 87-162.
85. Heise GA, Boff E: Taming action of chlordiazepoxide. Fed Proc 20:393, 1961.
86. Heuschele WP: Chlordiazepoxide for calming zoo animals. J Am Vet Med Assoc 139:996–998,
1961.
87. Christmas AJ, Maxwell DR: A comparison of the effects of some benzodiazepines and other
drugs on aggressive and exploratory behaviour in mice and rats. Neuropharmacology 9:17–29,
1970.
88. Langfeldt T, Ursin H: Differential action of diazepam on flight and defense behavior in the
cat. Psychopharmacologia 19:61–66, 1971.
89. Miczek KA: Intraspecies aggression in rats: Effects of D-amphetamine and chlordiazepoxide.
Psychopharmacologia 39:275–301, 1974.
90. Olivier B, Mos J, Miczek KA: Ethopharmacological studies of anxiolytics and aggression. Eur
Neuropsychopharmacol 1:97–100, 1991.
91. Krsiak M: Effect of ethanol on aggression and timidity in mice. Psychopharmacology 51:75–80,
1976.
92. Krsiak M, Sulcova A: Differential effects of six structurally related benzodiazepines on some
ethological measures of timidity, aggression and locomotion in mice. Psychopharmacology
101:396–402, 1990.
93. Miczek KA, Haney M: Psychomotor stimulant effects of D-amphetamine, MDMA and PCP:
Aggressive and schedule-controlled behavior in mice. Psychopharmacology 115:358–365, 1994.
94. Weerts EM, Tornatzky W, Miczek KA: "Anxiolytic" and "anxiogenic" benzodiazepines and
beta-carbolines: Effects on aggressive and social behavior in rats and squirrel monkeys.
Psychopharmacology 110:451–459, 1993.
95. Rodgers RJ, Waters AJ: Benzodiazepines and their antagonists: A pharmacoethological
analysis with particular reference to effects on "aggression." Neurosci B 9:21–35, 1985.
96. File SE, Lister RG, Nutt DJ: The axiogenic action of benzodiazepine antagonists. Neurophar-
macology 21 : 1033–1037, 1982.
97. File SE, Lister RG, Nutt DJ: Intrinsic actions of benzodiazepine antagonists. Neurosci Lett
32:165–168, 1982.
98. File SE, Pellow S: Intrinsic actions of the benzodiazepine receptor antagonist Ro 15-1788.
99. Skolnick P, Reed GF, Paul SM: Benzodiazepine-receptor mediated inhibition of isolation-
induced aggression in mice. Pharmacol Biochem Behav 23:17–20, 1985.
100. Lister RG, Nutt DJ: Alcohol antagonists—the continuing quest. Alcohol Clin Exp Res, 12:566–
569, 1988.
101. Deacon RMJ, Guy AP, Gardner CR: Effects of selected imidazopyrimidine ligands for ben-
zodiazepine receptors in rodent models of anxiety and behavioural impairment. Drug Dev
Res 22:321–329, 1991.
102. Bonetti EP, Burkard WP, Gabl M, Mohler H: The partial inverse benzodiazepine agonist Ro
15-4513 antagonizes acute alcohol effects in mice and rats. Br J Pharmacol 86 (Suppl):463–467,
1985.
103. Belzung C, Misslin R, Vogel E: The benzodiazepine inverse agonists beta-CCM and RO
15-3505 both reverse the anxiolytic effects in mice. Life Sci 42:1765–1772, 1988.
104. Misslin R, Belzung C, Vogel E: Interaction of RO 15-4513 and ethanol on the behaviour of
mice: Antagonistic or additive effects? Psychopharmacology 94:392–396, 1988.
105. Koob GF, Percy L, Britton KT: The effects of Ro 15-4513 on the behavioral actions of ethanol
in an operant reaction time task and a conflict test. Pharmacol Biochem Behav 31:757–760,1989.
106. Glowa JR, Crawley J, Suzdak PD, Paul SM: Ethanol and the GABA receptor complex:
Studies with the partial inverse benzodiazepine receptor agonist Ro 15-4513. Pharmacol
Biochem Behav 31:767–772, 1989.
107. Lister RG: Interactions of RO 15-4513 with diazepam, sodium pentobarbital and ethanol in a
holeboard test. Pharmacol Biochem Behav 28:75–79, 1987.
108. Lister RG: The benzodiazepine receptor inverse agonists FG 7142 and RO 15-4513 both
reverse some of the behavioral effects of ethanol in a holeboard test. Life Sci 41:1481–1489,
1987.
109. Bell R, Hobson H: 5-HT1a receptor influences on rodent social and agonistic behavior: A
review and empirical study. Neurosci B 18:325–328, 1994.
110. Britton KT, Ehlers CL, Koob GF: Is ethanol antagonist Ro 15-4513 selective for ethanol?
Science 239:648–649, 1988.
111. Lister RG, Karanian JW: RO 15-4513 induces seizures in DBA/2 mice undergoing alcohol
withdrawal. Alcohol 4:409–411, 1987.
112. Jensen EV, Jacobsen HI: Basic guides to the mechanism of estrogen action. Recent Prog Horm
Res 18:387–414, 1962.
113. Majewska MD, Harrison NL, Schwartz RD, et al: Steroid hormone metabolites are barbitu-
rate-like modulators of the GABA receptor. Science 232:1004–1007, 1986.
114. Gee KW, Chang WC, Brinton RE, McEwen BS: GABA-dependent modulation of the Cl–
ionophore by steroids in rat brain. Eur J Pharmacol 136:419–423, 1987.
115. Gee KW, Bolger MB, Brinton RE, et al: Steroid modulation of the chloride ionophore in rat
brain: Structure–activity requirements, regional difference and mechanism of action. Journal
of Pharmacology and Experimental Therapeutics 246:803–812, 1988.
116. Lopez-Colome AM, McCarthy M, Beyer C: Enhancement of 3H-muscimol binding to brain
synaptic membranes by progesterone and related pregnanes. Eur J Pharmacol 176:297–303,
1990.
117. Morrow AL, Suzdak PD, Paul SM: Steroid hormone metabolites potentiate GABA receptor-
mediated chloride ion flux with nanomolar potency. Eur J Pharmacol 142:483–485, 1987.
118. Morrow AL, Pace JR, Purdy RH, Paul SM: Characterization of steroid interactions with
γ-aminobutyric acid receptor-gated chloride ion channels: Evidence for multiple steroid
recognition sites. Mol Pharmacol 37:263–270, 1990.
119. Simmonds MA, Turner JP, Harrison NL: Interactions of steroids with the GABA-A receptor
complex. Neuropharmacology 23:877–878, 1984.
120. Crawley JN, Glowa JR, Majewska MD, Paul SM: Anxiolytic activity of an endogenous
adrenal stereoid. Brain Res 398:382–385, 1986.
121. Britton KT, Page M, Baldwin H, Koob GF: Anxiolytic activity of steroid anesthetic alphax-
alone. Journal of Pharmacology and Experimental Therapeutics 258:124–129, 1991.
122. Norberg L, Wahlstrom G, Backstrom T: The anaesthetic potency of 3a-hydroxy-5a-preg-
nan-20-one and 3a-hydroxy-5b-pregnan-20-one determined with an intravenous EEG-
threshold method in male rats. Pharmacol Toxicol 61:42–47, 1987.
123. Bitran D, Hilvers RJ, Kellogg CK: Anxiolytic effects of 3a-hydroxy-5a β -pregnan-20-one:
Endogenous metabolites of progesterone that are active at the GABAA receptor. Brain Res
561:157–161, 1991.
124. Wieland S, Lan NC, Mirasedeghi S, Gee KW: Anxiolytic activity of the progesterone metab-
olite 5alpha-pregnan-3alpha-ol-20-one. Brain Res 565:263–268, 1991.
125. Belelli D, Gee KW: 5a-pregnan-3a, 20a-diol behaves like a partial agonist in the modulation
of GABA-stimulated chloride uptake by synaptoneurosomes. Eur J Pharmacol 167:173–176,
1989.
126. Gyermek L, Genther G, Fleming N: Some effects of progesterone and related steroids on the
central nervous system. Int J Neuropharmacol 6:191–198, 1967.
127. Seeman P: The membrane actions of anesthetics and tranquilizers. Pharmacol Rev 24:583–
655, 1972.
128. Kubli-Garfias C, Cervantes M, Beyer C: Changes in multiunit activity and EEG induced by the
administration of natural progestins to flaxedil immobilized cats. Brain Res 114:71–81, 1976.
129. Mendelson WB, Martin JV, Perlis M, et al: Sleep induction by an adrenal steroid in the rat.
130. Majewska MD: Actions of steroids on neuron: role in personality, mood, stress, and dis-
ease. Integr Psychol 5:258–273, 1987.
131. Wieland S, Belluzzi JD, Stein L, Lan NC: Comparative behavioral characterization of the
neuroactive steroids 3alpha-OH, 5alpha-pregnan-20-one and 3alpha-OH, 5beta-pregnan-20-
one in rodents. Psychopharmacology 118:65–71, 1995.
132. Kokate TG, Svensson BE, Rogawski MA: Anticonvulsant activity of neurosteroids: Correla-
tion with gamma-aminobutyric acid-evoked chloride current potentiation. Journal of Phar-
macology and Experimental Therapeutics 270:1223–1229, 1994.
133. Biikiisoglu C, Thalhammer JG, Krieger NR: Analgesia with anesthetic steroids and ethanol.
Anesth Analg 77:27–31, 1993.
134. Kavaliers M, Wiebe JP: Analgesic effects of the progesterone metabolite, 3a-hydroxy-5a-
pregnan-20-one and possible modes of action in mice. Brain Res 414:393–398, 1987.
135. DeBold JF, Barros H, So S, Miczek KA: The effects of allopregnanolone and alcohol on
intramale aggression in mice. Alcohol Clin Exp Res 19:11A, 1995.
136. Haug M, Young J, Robel P, Baulieu EE: Neonatal testosterone administration potentiates the
inhibitory effect of dehydroepiandrosterone on the display of aggression by spayed female
mice towards lactating intruders. C R Acad Sci III 312:511–516, 1991.
137. Young J, Corpechot C, Haug M, et al: Suppressive effects of dehydroepiandrosterone and
3-b-methyl-androst-5-en-17-one on attack towards lactating female intruders by castrated
male mice .2. Brain neurosteroids. Biochem Biophys Res Commun 174:892–897, 1991.
138. Melchior CL, Allen PM: Interaction of pregnanolone and pregnenolone sulfate with ethanol
and pentobarbital. Pharmacol Biochem Behav 42:605–611, 1992.
139. Melchior CL, Ritzmann RF: Dehydroepiandrosterone enhances the hypnotic and hypother-
mic effects of ethanol and pentobarbital. Pharmacol Biochem Behav 43:223–227, 1992.
140. Frye CA, Duncan JE: Progesterone metabolites, effective at the GABA(a) receptor complex,
attenuate pain sensitivity in rats. Brain Res 643:194–203, 1994.
141. Wu FS, Gibbs TT, Farb DH: Pregnenolone sulfate: A positive allosteric modulator at the
n -methyl-D-aspartate receptor. Mol Pharmacol 40:333–336, 1991.
142. Irwin RP, Lin S-Z, Rogawski MA, et al: Steroid potentiation and inhibition of N-methyla-
aspartate receptor-mediated intracellular Ca++ responses: Structure-activity studies. journal
of Pharmacology and Experimental Therapeutics 271:677–682, 1994.
143. Farb DH, Gibbs TT, Wu FS, et al: Steroid modulation of amino acid neurotransmitter recep-
tors, in Biggio G, Concas A, Costa E (eds): GABAergic Synaptic Neurotransmission. New York,
Raven Press, 1992, pp 119–131.
144. Criswell HE, Simson PE, Duncan GE, et al: Molecular basis for regionally specific action of
ethanol on gamma-aminobutyric acid-A receptors: Generalization to other ligand-gated ion
channels. Journal of Pharmacology and Experimental Therapeutics 267:522–537, 1993.
145. Suzdak PD, Schwartz RD, Skolnick P, Paul SM: Ethanol stimulates gamma-aminobutyric
acid receptor-mediated chloride transport in rat brain synaptoneurosomes. Proc Natl Acad
Sci USA 83:4071–4075, 1986.
146. Allan AM, Harris RA: Gamma-aminobutyric acid and alcohol actions: Neurochemical stud-
ies of long sleep and short sleep mice. Life Sci 39:2005–2015, 1986.
147. Leidenheimer NJ, Harris RA: Acute effects of ethanol on GABA alpha receptor function:
Molecular and physiological determinants, in Biggio G, Concas A, Costa E (eds): GABAergic
Synaptic Transmission. New York, Raven Press, 1992, pp 269–279.
148. Lewis MH, Gariepy JL, Gendreau P, et al: Social reactivity and D1 dopamine receptors:
Studies in mice selectively bred for high and low levels of aggression. Neuropsychopharmacol-
ogy 10:115–122, 1994.
149. Grant KA: Emerging neurochemical concepts in the actions of ethanol at ligand-gated ion
channels. Behav Pharmacol 5:383–404, 1994.
150. Hubbell CL, Marglin SH, Spitalnic SJ, et al: Opioidergic, serotonergic, and dopaminergic
manipulations and rats' intake of a sweetened alcoholic beverage. Alcohol 8:355–367, 1991.
151. Nutt DJ, Peters TJ: Alcohol: The drug. Br Med Bull 50:5–17, 1994.
152. Miczek KA, DeBold JF, van Erp AMM: Neuropharmacological characteristics of individual
differences in alcohol effects on aggression in rodents and primates. Behav Pharmacol 5:407–
421, 1994.
153. Wafford KA, Burnett DM, Leidenheimer NJ, et al: Ethanol sensitivity of the GABAA receptor
expressed in Xenopus oocytes requires 8 amino acids contained in the g2L subunit. Neuron
7:27–33, 1991.
154. Criswell HE, Simson PE, Knapp DJ, et al: Effect of zolpidem on gamma-aminobutyric acid
(GABA)-induced inhibition predicts the interaction of ethanol with GABA on individual
neurons in several rat brain regions. Journal of Pharmacology and Experimental Therapeutics
273:526–536, 1995.
155. Inglefield JR, Sieghart W, Kellogg CK: Immunohistochemical and neurochemical evidence
for GABA(A) receptor heterogeneity between the hypothalamus and cortex. J Chem Neuro-
anat 7:243–252, 1994.
156. Wisden W, Laurie DJ, Monyer H, Seeburg PH: The distribution of 13-GABAA receptor
subunit messenger RNAs in the rat brain 1. Telencephalon, diencephalon, mesencephalon.
J Neurosci 12:1040–1062, 1992.
157. Zimprich F, Zezula J, Sieghart W, Lassmann H: Immunohistochemical localization of the
alpha-1, alpha-2 and alpha-3 subunit of the GABAA receptor in the rat brain. Neurosci Lett
127:125–128,1991.
158. Givens BS, Breese GR: Electrophysiological evidence that ethanol alters function of medial
septal without affecting lateral septal function. Journal of Pharmacology and Experimental Thera-
peutics 254:528–538, 1990.
159. Siegel A, Mirsky AF: The neurobiology of violence and aggression, in Reiss AJ Jr, Miczek
KA, Roth JA (eds): Understanding and Preventing Violence. Biobehavioral Influences. Washing-
ton, DC, National Academy Press, 1994, vol 2, pp 59–172.
160. Samson HH, Harris RA: Neurobiology of alcohol abuse. Trends in Pharmacological Sciences
13:206–211, 1992.
161. Mohler H: GABAergic synaptic transmission-Regulation by drugs. Arzneim Forsch 42(-1):

211–214, 1992.
162. Paredes RG, Agmo A: GABA and behavior: The role of receptor subtypes. Neurosci B 16:145–
170, 1992.
163. Costa E: The allosteric modulation of GABA(a) receptors: Seventeen years of research.
Neuropsychopharmacology 4:225–235, 1991.
164. Samson HH, Tolliver GA, Lumeng L, Li TK: Ethanol reinforcement in the alcohol non-
preferring rat: Initiation using behavioral techniques without food restriction. Alcohol Clin
Exp Res 13:378–385, 1989.
8
Serotonin in Early-Onset Alcoholism

Matti Virkkunen and Markku Linnoila
Abstract. This chapter examines current, common schemes to subgroup alcoholics to arrive at
relatively homogeneous groups of patients to facilitate psychobiological and molecular genetic
studies. Early-onset, male-limited alcoholism is commonly associated with antisocial personality
disorder or antisocial behavioral traits. It is often preceded by early-onset aggressiveness, which
is followed by conduct disorder. Early-onset alcoholism among men is associated with low
central serotonin turnover rate. The data concerning platelet MAO activity and serotonin uptake
to platelets among early-onset alcoholics are conflicting. Recent molecular genetic and brain
imaging studies on early-onset alcoholics are preliminary but appear very promising.
Alcoholism is the most common mental disorder among men in the United
States. Its prevalence is highest among the 24 to 40 age group.1 Like many
mental disorders, alcoholism is a heterogeneous condition.
1. Subgrouping Alcoholics
Various classification schemes have been developed to obtain relatively
homogeneous subgroups of alcoholics to facilitate scientific research. A gen-
erally accepted classification of alcoholics takes advantage of the common
comorbidity of many mental disorders with alcoholism among treatment-
seeking patient samples. Patients whose alcoholism started before the symp-
toms of other mental disorders appeared or who have only alcoholism are
classified as primary alcoholics. In secondary alcoholics, the onset of another
mental disorder precedes the onset of alcoholism.2
Matti Virkkunen • Department of Psychiatry, Helsinki University Central Hospital, Helsinki

00180, Finland. Markku Linnoila • Laboratory of Clinical Studies, Division of Intramural
Clinical and Biological Research, National Institute on Alcohol Abuse and Alcoholism, Bethesda,
Maryland 20892-1256.
173
The Achilles heel of this classification system lies in the possibility that an
age-appropriate expression of genetic vulnerability toward alcoholism, which
precedes the onset of alcoholism, is different from alcoholism. This may be
the case for a subgroup of boys with attention deficit disorder,3 conduct
disorder, or excessive early-onset aggressiveness.4 Thus, the above classifica-
tion scheme may be of limited utility for genetic studies on subtypes of alco-
holism.
An innovative, large-scale, adoption study based on a general population
sample in Sweden, conducted by Cloninger et al.,5 has been of great heuristic
value for the field. This study defined two relatively homogenous subgroups
of alcoholics called type 1 and type 2. This subgrouping scheme has been
successfully utilized in a number of studies on personality, biochemistry, and
genetics of alcoholics. A significant problem with this classification scheme,
however, is the lack of validated and agreed upon diagnostic criteria for the
subgroups applicable to samples of treatment-seeking alcoholics.6
2. Type 2 Alcoholism and Early-Onset Alcoholism

Cloninger et al.,5 in a large-scale Swedish adoption study of 862 men and
913 women, defined male-limited, type 2 alcoholism. It is characterized by
high heritability from fathers to sons, early onset, and is related to antisocial,
often violent behavioral traits. Twenty-five percent of alcoholic men in the
sample of Cloninger et al.5 were deemed to have type 2 alcoholism. Seventy-
five percent of men and all women in this cohort fulfilled criteria for type 1
alcoholism. These patients were characterized by a relatively late age of onset
of excessive drinking, anxious personality traits, and adverse environmen-
tal conditions during development. A recent, separate adoption study con-
ducted in Sweden by Sigrardsson et al.7 included 577 men and 660 women.
The findings confirmed all of the previously reported distinctions between
these two subtypes of alcoholism in terms of age of onset, patterns of inheri-
tance, and differences in gender distribution and personality traits. Clon-
inger8 has repeatedly emphasized that type 2 alcoholism is not only associ-
ated with teenage onset of alcohol abuse, but also with recurrent criminal and
often violent behaviors.
There have been several attempts to apply the findings of Cloninger and
colleagues5,7,8 to treatment-seeking alcoholics. Babor et al.9 proposed the ter-
minology of types A and B to subgroups of treatment-seeking alcoholics with
characteristics similar to Cloninger and co-workers’s subgroups, but without
taking into account family history. von Knorring et al.10 operationalized a
two-group classification scheme based primarily on age of onset criteria and
secondarily on complications thought to be consequences of antisocial behav-
ioral traits. Even though von Knorring et al.10 did not use paternal family
history as a criterion, early-onset alcoholics were labeled as type 2 and late-
onset alcoholics as type 1. In clinical samples, Irwin et al.11 and Lamparski et
8 • Serotonin in Early-Onset Alcoholism 175
al.6 were only partially successful in applying the findings of Cloninger et al.5
Thus, there is no consensus about the exact criteria for type 2 alcoholism to be
used in subgrouping clinical samples of alcoholics.
Gilligan et al.12 and Sullivan et al.13 have developed sets of clinical criteria
for the type 1 and type 2 subgrouping. Anthenelli et al.14 found that type 1
and 2 subgroups classified according to the criteria of Sullivan et al.13 signifi-
cantly overlap the subgroups diagnosed according to the primary–secondary
alcoholism scheme. There was a 73% concordance between the type 2 sub-
group and the subgroup with primary antisocial personality disorder and
secondary alcoholism in the Anthenelli et al.14 study.
Most researchers are of the opinion that, among male inpatients, early
onset of alcoholism is the central feature of the type 2 criteria.10–15 Further-
more, there seems to be a reasonable consensus that this form of alcoholism is
common among the paternal male relatives, and the fathers often have an
antisocial personality disorder or marked antisocial behavioral traits.
Hill16 has proposed a third type of alcoholism that affects primarily men
and is characterized by early onset, alcoholic fathers free of an antisocial
personality disorder, and brothers with repeated episodes of fighting while
intoxicated. Furthermore, Hill16 found no evidence that these fathers would
have explosive behavioral traits without satisfying criteria of antisocial per-
sonality disorder, different from the findings of Virkkunen et al.17,18 Turner et
al.19 used a pattern analysis of inheritance of alcoholism to demonstrate that
both father and mother may contribute equally to early age of onset of alco-
holism in the offspring. Although the mother may or may not be an alcoholic,
her brothers and father often are. This finding is consistent with X-linked
transmission of vulnerability to alcoholism,20 but differs from models pro-
posed by Cloninger et al.5,21 Consistent with both models, Linnoila et al.22
found that a majority of early-onset, impulsive violent male alcoholics had
impulsive alcoholic fathers. However, a minority had a mother who herself
was not an alcoholic but had an alcoholic father and brothers.
Type 2 alcoholics have been reported to be prone to violence during
intoxication.8,10,17,21,23,24 According to Moss et al.,25 the strongest correlate of
aggressivity in the son is the father's negative affectivity in the Multidimen-
sional Personality Questionnaire, which correlates positively with the pro-
pensity of the father for physical aggression. This is consistent with a large
proportion of this subgroup of alcoholics actually having antisocial person-
ality disorder.11 The propensity for violent behavior among a subgroup of
type 2 alcoholics is highlighted by the finding that among all perpetrators of
recidivist homicides in Finland, 85% had type 2 alcoholism and antisocial
personality disorder.24
3. Antisocial Personality Disorder and Early-Onset Alcoholism

There is a consensus on antisocial personality disorder being very com-
mon among men with early-onset alcoholism. Alcoholics with antisocial per-
sonality disorder are younger at the onset of alcohol-related problems and

they experience more adverse social consequences of drinking, especially
while under the influence of alcohol. They are also younger at first treatment
and have a generally worse outcome than alcoholics who do not meet criteria
for antisocial personality disorder at the time of or before the onset of alcohol-
ism.26–30
According to Schuckit et al.,31 it is difficult to identify any other reliable
personality profile than antisocial personality disorder that is associated with
an individual's risk of alcoholism. Hesselbrock32 found that, in the United
States, 60 to 80% of patients with antisocial personality disorder have con-
comitant alcohol and/or other substance dependence. Antisocial personality
disorder and type 2 alcoholism clearly overlap in many population samples,
and some of the clinical relevance of the type 1/type 2 classification is lost if
individuals with antisocial personality disorder are excluded.33 There are,
however, a number of early-onset alcoholics, both men and women, who do
not fulfill the diagnostic criteria of antisocial personality disorder.34,35
The same personality features, high novelty seeking, low harm avoid-
ance, and low reward dependence on the Tridimensional Personality Ques-
tionnaire,8,21 are characteristic of both type 2 alcoholism and antisocial per-
sonality disorder. Hesselbrock and Hesselbrock36 found men with antisocial
personality disorder to be higher than men without antisocial personality
disorder on measures of impulsivity, sensation seeking, and monotony
avoidance but not on harm avoidance and reward dependence. As young
adults, individuals characterized by these personality traits are at high risk to
commit violent crimes37 and to receive the psychiatric diagnosis of antisocial
personality disorder.38 Furthermore, they have recently been found to have a
prominent character trait termed uncooperativeness, which is defined as a lack
of empathy, social tolerance, compassion, and moral principles.39
Cadoret et al.40,41 have published adoption studies that point to high
heritability of antisocial personality disorder (i.e., a biological parent with
antisocial traits is at high risk to produce an adoptee with antisocial person-
ality disorder). The antisocial personality disorder is hypothesized to be con-
ducive to the development of alcoholism. The conclusion concerning the
direction of the effect is based on the fact that, in most patients, antisocial
personality disorder manifests itself in childhood and adolescence as conduct
disorder. This diagnosis precedes the characteristically early onset of sub-
stance abuse in these individuals.
4. Is Early-Onset Aggressivity Specifically Conducive to the

Development of Early-Onset Alcohol Abuse?
In a prospective study on precursors to problem drinking among young

adults in Finland, Pulkkinen and Pitkanen4 found that early aggression mea-
sured at the age of 8 contributed specifically but indirectly to adult problem
drinking. A high rating of early aggression preceded the development of a

full-blown conduct disorder at the age of 14 when the youngsters were inves-
tigated the second time. An alcoholism screening test, the CAGE question-
naire,42 was administered and peer and teacher ratings were collected at both
ages among the original sample that was drawn in 1968 from 369 second-
grade students. They were followed until adulthood. Criminal records were
investigated at the ages of 20 and 26. Problem drinking was defined by the
CAGE questionnaire and arrests for alcohol abuse and disorderly or criminal
conduct while intoxicated. A factor analysis confirmed the existence of a
factor “problem drinking,” which differed clearly from “social drinking” and
“controlled drinking.” Among women, the score on the scale for problem
drinking was unrelated to aggression, but it correlated positively with anxi-
ety. Among males, it correlated positively with aggression at the 8-, 14-, and
26-year interviews and negatively with the scores of social anxiety and proso-
ciality. Thus, among men, early aggressiveness preceded conduct disorder,
which preceded antisocial personality disorder and alcoholism. These pro-
spective findings are compatible with the Cloninger et al.5 findings on type 2,
male-limited alcoholism.
5. Psychobiology of Early-Onset Alcoholism

There are a number of studies that support the idea that a child’s early
behavioral characteristics, especially aggression and conduct problems
among boys, predict problem drinking.43–49 According to Pulkkinen,50 poor
impulse control is the primary cause of early-onset aggressivity. Because low
brain serotonin turnover rate, as indicated by a low cerebrospinal fluid (CSF)
5-hydroxyindoleacetic acid (5-HIAA) concentration, prospectively predicts
habitual violence and impaired impulse control among rhesus monkeys,51
young alcoholic violent men,17,52 and adolescent males with disruptive be-
havioral disorders,53 it may be a biochemical risk marker for both early-onset
aggressiveness and early-onset alcohol abuse.
In addition to low CSF 5-HIAA concentration, low CSF ACTH,54 low
urinary 24-hr cortisol secretion,55 and high CSF free testosterone54 are com-
mon among habitually violent offenders who have antisocial personality dis-
order and early-onset alcoholism. Additionally, Bergman and Brismar56 have
found that abusive and suicidal behaviors among von Knorring’s type 2 alco-
holics are often associated with high testosterone and low cortisol outputs
among the violent subgroup but not among the suicidal subgroup free of
interpersonally violent behavior. Low CSF ACTH and 24-hr urinary cortisol
may be features associated with low trait anxiety, which is one of the person-
ality features characteristic of early-onset male problem drinkers.
King et al.57 also found that young substance abusers, most of whom had
alcohol problems, had low plasma cortisol levels when measured at 8:00 AM
after an overnight fast. Moreover, they observed that low plasma cortisol
levels were associated with impulsivity among normal controls. Thus, they
speculated that impulsivity per se might influence plasma cortisol concentra-
tions. Future psychobiological studies on early-onset alcoholism should in-
clude indicators of hypothalamic-pituitary-gonadal and hypothalamic–pitu-
itary–adrenal axes activities.
6. Early-Onset Alcoholism and Antisocial Personality Disorder

Are Associated with Reduced Brain Serotonin Turnover
Rate and Transmission
6.1. Serotonin Turnover Rate (CSF 5-HIAA)
There is an increasing body of evidence that reduced central serotonin
turnover rate is associated with early-onset alcoholism and antisocial person-
ality traits.8 We have found among alcoholics that a majority of the violent
type 2 patients are characterized by a low CSF 5-HIAA concentration.23,52,58,59
In prospective follow-up studies, low CSF 5-HIAA and low blood glucose
nadir after a glucose challenge were predictive of recidivist violent crimes
under the influence of alcohol among early-onset male alcoholic violent of-
fenders.17 Low CSF 5-HIAA was particularly associated with a family history
positive for paternal alcoholism and violence22 compatible with Cloninger’s
findings.5
In our most recent follow-up study, alcoholic recidivists who committed
new violent crimes after release from prison had significantly lower CSF
5-HIAA, and 3-methoxy-4-hydroxyphenylglycol (MHPG) (main central me-
tabolite of norepinephrine concentrations.52 Low CSF 5-HIAA concentrations
were associated with family histories positive for both paternal violence and
alcoholism but not with family histories positive for paternal alcoholism but
negative for paternal violence. Again, these findings are compatible with
family histories characteristic of type 2 alcoholism as defined by Cloninger et
al.5 In a prospective follow-up study among American children and adoles-
cents with disruptive behavioral disorders (mainly conduct disorder), Kruesi
et al.53 also found low CSF 5-HIAA and HVA to be predictive of future violent
behavior.
Among the impulsive alcoholic violent offenders, high novelty seeking,
low monotony avoidance, high irritability, and low socialization scores on the
Karolinska Scales of Personality are associated with low CSF 5-HIAA concen-
trations.18 However, Cloninger8 postulates that although a central se-
rotonergic deficit is consistently correlated with antisocial character traits, it is
only indirectly and inconsistently associated with personality traits such as
high novelty seeking,60 low harm avoidance,60 other measures of impul-
sivity,18 or history of aggressiveness.18
In addition to alcoholic violent patients with antisocial personality disor-
der, alcoholic violent patients with impulsive personality disorders other than
antisocial personality disorder, who repeatedly exhibit explosive behaviors
under the influence of alcohol, also have extremely low CSF 5-HIAA concen-
trations. Typically, these patients have not, however, fulfilled all the diag-
nostic criteria for attention deficit or conduct disorders during childhood or
early adolescence, and their problems do not become manifest until after
puberty. Usually, by 20 years of age, they have developed alcoholism. Biolog-
ically, many of these patients are characterized by a profound disturbance of
day and night activity rhythm.54 These differences between patients with
antisocial personality and intermittent explosive disorder emphasize the im-
portant role early impulsive aggressivity can play in the very early onset of
alcohol abuse among patients with antisocial personality disorder.
Even among alcoholics without personality disorders, those with an early
onset (≤ 25 years) have a lower mean CSF 5-HIAA concentration than late-
onset (> 25 years) alcoholics.35 It is important to note, however, that the
mean CSF 5-HIAA concentration among these American patients with early-
onset alcoholism was much higher (85.6 ± 34.7 nmole/liter) than the mean
CSF 5-HIAA concentration among the violent antisocial and explosive crimi-
nal alcoholics (58.8 ± 25.2 nmoleiliter) in Finland who often also had a history
positive for suicide attempts.54 A history positive for suicide attempts has
been repeatedly associated with extremely low CSF 5-HIAA concentration
among the alcoholic violent Finns.17,52,54,61 This finding was not, however,
replicated in our two recent American studies on nonviolent alcoholics, most
of whom did not have personality disorders.35,58 Among late-onset alcoholics
in the American study,35 the mean CSF 5-HIAA concentration was 103.6 ±
38.9 nmole/liter.
Also, our earlier American study58 reported a lower mean CSF 5-HIAA
concentration among early- as compared to late-onset alcoholics, but the find-
ing was not statistically significant. The major differences between the two
studies were the larger sample size and the elimination of between-assay
variability in the Fils-Aime et al.35 study and a different age of onset criterion.
The age of onset criterion in the Fils-Aime et al.35 study was selected to
directly test the postulates of Cloninger et al.5 and Irwin et al.11
Many other investigators have reported reduced CSF 5-HIAA concentra-
tions in abstinent alcoholics, but they have not addressed the issue of sub-
grouping. Ballenger et al.62 studied abstinent, young male alcoholics and
reported one of the largest differences in CSF 5-HIAA between alcoholics and
controls. The demographics of their sample made it very likely that a large
proportion of their patients were early-onset alcoholics. Banki63 found low
CSF 5-HIAA in both recently abstinent alcoholic men and women, and Tak-
ahashi et al.64 found low CSF 5-HIAA concentrations only in alcoholics who
had experienced severe withdrawals.
6.2. Serotonin Receptor Challenges in Early- versus Late-Onset

Male Alcoholics
Recent pharmacological challenge studies65–67 suggest that serotonergic
dysregulation may be more prominent among von Knorring’s type 2 than
among type 1 alcoholics and healthy subjects. Meta-chlorophenylpiperazine

(mCPP), a primary serotonin (5-HT2c) receptor agonist, has been found to
produce blunted prolactin and cortisol responses in alcoholics compared to
healthy volunteers but to elicit alcohol-like cues mainly among von Knor-
ring’s type 2 alcoholics.65–67 In addition to the 5-HT2c receptor, mCPP may
also affect 5-HT1a, 5-HT1d, 5-HT2a, 5-HT3, and 5-HT7 receptors.66 Blunted
hormonal responses to mCPP have also been observed in patients with anti-
social personality disorder and substance abuse.68
6.3. Abnormal Tryptophan–Large Neutral Amino Acid Ratio in Plasma

Fernstrom and Wurtman69,70 reported in the early 1970s that tryp-
tophan–large neutral amino acid (LNAA) ratio in plasma regulates the avail-
ability of tryptophan to the brain among laboratory rodents. They suggested
that the rate of entry of tryptophan regulated the rate of serotonin synthesis
in the brain. The physiological importance of this system under steady-state
conditions, however, has not been extensively elucidated in humans.
Buydens-Branchey et al.71 observed that among alcoholics who had started
to abuse alcohol before 20 years of age there was an association between a low
tryptophan–LNAA ratio in plasma and depressive and aggressive behaviors.
The ratio was measured 1 day after cessation of drinking and the ratio in-
creased progressively for at least 2 to 3 weeks after detoxification. Many of the
subjects who had exhibited early-onset alcohol-seeking behavior had also
committed crimes of violence and had a history of paternal alcoholism.
Branchey et al.72 also found that the tendency to experience blackouts,
i.e., to lose one’s memory under the influence of alcohol, was associated with
a low plasma tryptophan concentration quantified after 2 weeks abstinence.
On the other hand, Buydens-Branchey et al.71 and Virkkunen and Narvanen73
found an elevated tryptophan–LNAA ratio among violent male alcoholics.
Interestingly, Fils-Aime et al.35 found a nonsignificant negative correla-
tion between CSF 5-HIAA and tryptophan concentrations among early-onset
alcoholics. According to these data, under steady-state conditions in humans,
tryptophan crossing the blood–brain barrier may not be the rate-limiting step
for serotonin synthesis.70 These findings are in accordance with our recent
postulate that in humans, under steady-state conditions, the activity of tryp-
tophan hydroxylase rather than the availability of tryptophan may be rate-
limiting for the turnover rate of central serotonin.74
6.4. Platelet Monoamine Oxidase

Monoamine oxidase (MAO) is the major enzyme that metabolizes bio-
genic monoamine neurotransmitters, including serotonin, by oxidative deam-
ination. Two forms of MAO, A and B, are widely distributed in the human
body, with high activity of the B form in blood platelets.75,76 There is a rela-
tively large number of published studies correlating behavioral characteristics
with platelet MAO activity. A lack of correlation between platelet and brain
MAO activities renders the meaning of platelet MAO findings uncertain for
CNS disorders.77 Clinical and experimental findings support, however, the
hypothesis that platelet MAO activity may be a marker of the functional
capacity of the central serotonin system.78,79
Many authors have reported that platelet MAO activity is lower among
alcoholics than healthy volunteers.80–88 von Knorring et al.10,89,90 and Sullivan
et al.13 both found that early-onset male alcoholics had low mean platelet
MAO activity compared to late-onset alcoholics and healthy volunteers, who
did not differ from each other. In these studies, type 2 alcoholics were charac-
terized by early onset of alcoholism, alcoholism in first-degree relatives, and a
high prevalence of drug abuse and social complications. Similar to these
findings, low platelet MAO activity has also been reported in patients with
antisocial personality and borderline personality disorders.91,92 Yates et al.85
and Anthenelli et al.,86 however, did not find platelet MAO activity differ-
ences between type 1 and 2 alcoholics.
Platelet MAO activity has been found to be negatively correlated with the
personality traits of impulsivity and sensation seeking,93 which are among
the traits associated with early-onset alcohol abuse among men, as reviewed
above. These are also traits that, according to genetic studies, have high
heritability and characterize many children with conduct disorder who have
often been exposed to family discord and disruption.94 As reviewed above,
impaired impulse control has been repeatedly associated with reduced brain
serotonin turnover rate.
Moreover, symptoms of conduct disorder commonly precede early alco-
hol problems and the illness often progresses to an antisocial personality dis-
order. Therefore, it is surprising that platelet MAO activity has been reported
to be higher in sons of substance-abusing fathers who have conduct disorder
than in sons of substance-abusing fathers who do not have conduct disorder
or among sons of nonsubstance-abusing fathers who have conduct disor-
der.95 Furthermore, Stoff et al.96 found that impulsivity, which is often an
important clinical feature of conduct disorder, was positively correlated with
platelet MAO activity. Shekim et al.,97 on the other hand, found that both low
and high platelet MAO activities were associated with impulsivity among
healthy children. Bowden et al.98 found lower platelet MAO activity among
youths with conduct and attention deficit disorder than among youths with
only attention deficit disorder.
Interestingly, among adults with personality disorders, prolactin re-
sponse to dl-fenfluramine has been found to be inversely correlated with
measures of aggression, motor impulsivity, assaultiveness, and irritability.99
In a study by Halperin et al.,100 the prolactin response to a dl-fenfluramine
challenge among youth with attention deficit disorder and a history of exhib-
iting aggressive behaviors was positively correlated with aggressiveness. The
subjects in the study were too young to have developed alcohol abuse.
The cause of the differences between the findings among the adults and
children and adolescents is unclear.101 It is possible that physiological

changes associated with puberty or increasing alcohol and other substance
abuse reverse the relationships between the platelet MAO activity, prolactin
responsiveness to fenfluramine, and behavior. Also, it is possible that al-
though by definition symptoms of conduct disorder always precede antiso-
cial personality disorder,102 boys with conduct disorder and high platelet
MAO activity may differ from boys with conduct disorder who are more likely
to become alcoholics with antisocial personality disorder with increasing age.
This suggestion is compatible with the finding of Gabel et al.101 that only a
very small proportion of the youths in their study had positive histories for
substance abuse at the mean age of 12.3 ± 2.0 years.
6.5. Uptake of Serotonin to Platelets

Many peripheral indicators of serotonin metabolism are currently being
investigated among alcoholics and especially among early-onset alcoholics.33
Blood platelets have been thought to be a representative peripheral model for
central serotonergic neuronal function because the kinetics of serotonin in
platelets resemble the kinetics of serotonin in synaptosomes.103 Thus far,
preliminary findings suggest that family-history-positive and family-history-
negative young men may have different platelet serotonin uptake rates. Fami-
ly history positives had higher maximum velocity (Vmax) of the uptake of
serotonin than family history negatives.104 Furthermore, Ernouf et al.105
found that descendants of alcoholics of both genders had a higher Vmax for
serotonin uptake than healthy age-matched offspring of nonalcoholics. Stud-
ies on alcoholics have also reported increased serotonin uptake by platelets of
alcoholics compared to controls.106,107 Only the directly quantified serotonin
uptake rate was higher among alcoholics than healthy volunteers, whereas no
differences between alcoholics and healthy volunteers were found for platelet
paroxetine-binding variables.107 Neiman et al.,108 however, found an elevated
affinity constant (K m) for serotonin in alcoholics only during withdrawal but a
normal rate of uptake in alcoholics during continued abstinence. Further-
more, Kent et al.109 found platelet serotonin uptake to be reduced by 18% on
the average in male alcoholics who had been abstinent for 2 weeks as com-
pared to healthy volunteers.
At this time, there are no studies available specifically investigating sero-
tonin uptake among patients with early-onset alcoholism, type 2 alcoholism
or alcoholism and antisocial personality disorder. Platelet serotonin uptake,
however, was found to be reduced among habitually violent impulsive of-
fenders,110 but this preliminary study contains no information concerning
alcohol use among the subjects.
Despite the conflicting reports, higher Vmax values among alcoholics and
their descendants as compared to healthy volunteers, if confirmed in future
studies, are particularly interesting because platelet serotonin uptake in major
depression has been reported to have a reduced Vmax in about 75% of the
studies.103 However, the issue of whether platelet uptake reflects serotonin
uptake in the brain remains unresolved.
6.6. Molecular Genetic Findings Associated with Early-Onset Alcoholism

Molecular genetic findings concerning early-onset or type 2 alcoholism
are preliminary, but several groups are pursuing this line of investigation.
Our approach, with Dr. David Goldman at the National Institute on Alcohol
Abuse and Alcoholism has been to directly scan the coding sequences of
serotonergic candidate genes for structural variants.
We have reported polymorphisms of several serotonin receptor
genes111–113 and the tryptophan hydroxylase (TPH) gene114 among violent
alcoholics. However, the only common amino acid substitution producing
polymorphism (5-HTcys23–ser23) is in the 5-HT2c receptor gene. The rarer vari-
ant has an allelic frequency of about 13% among Caucasians.112 Among Finns,
5-HT2c–ser23 showed a negative association in alcoholics as compared to
healthy volunteers matched for age, sex, and socioeconomic class, but it was
not associated with antisocial personality disorder or CSF 5-HIAA concentra-
tions.115 Thus, the only abundant structural variant discovered so far does
not seem to play a role in the vulnerability toward early-onset or type 2
alcoholism.115
An intronic polymorphism was identified in the TPH gene that may
influence serotonergic activity, and it has been genotyped as a single-strand
conformational polymorphism. Our preliminary data suggest an association
between the TPH “L” allele, severity of impulsivity, relatively low CSF
5-HIAA, and a history positive for suicide attempts among impulsive, violent
Finnish alcoholics.74 Because of the association between low CSF 5-HIAA and
the “L” allele, this polymorphism may be associated with type 2 and early-
onset alcoholism. Further mechanistic studies, however, are needed to eluci-
date the meaning, if any, of these associations because the polymorphism is
intronic. Analysis of the TPH promoter has identified several regions required
for tissue-specific and basal expression. Furthermore, Nielsen et al.114 have
discovered a novel transcription factor, which is being cloned and sequenced,
that regulates negatively TPH gene expression.
Vanyukov et al.116 have found an association between the vulnerability
toward early-onset alcoholism and substance abuse and a recently discovered
dinucleotide repeat-length polymorphism of the MAOA gene. A significant
correlation between the presence or absence off early-onset alcoholism and
substance abuse and the length of the MAOCA-1 repeat was found among
men but not among women. “Long” alleles (repeat length above 115 base
pairs) were associated with an early age of onset of substance abuse. At the
present time, these findings are difficult to interpret. This is because it is low-
platelet MAO activity that has been found to be associated with early-onset
alcoholism,10,13,89,90 and the MAO in platelets is MAOB, which is coded by a

different gene. The molecular genetic data on early-onset or type 2 alcoholism
are still very preliminary but promising.
6.7. Brain Neuroimaging Findings in Alcoholism

Using single-photon emission computed tomography (SPECT), Tiihonen
et al.117 found that patients with the most abundant type of alcoholism, which
in their study clinically resembles type 1 alcoholism as described by Cloninger
et al.,5,21 can be distinguished from violent patients with early-onset alcohol-
ism and from healthy volunteers by increased density of dopamine transpor-
ters in the caudate putamen. Cloninger8 regards this finding as the first iden-
tification of a specific neuroregulatory deficit in type 1 alcoholics. Among
early-onset alcoholics who had committed very severe violent crimes under
the influence of alcohol, the striatal density of the dopamine uptake sites was
only slightly higher than among controls. At this time, there are no brain-
imaging findings that have been identified to be characteristic of early-onset
or type 2 alcoholism.
7. Conclusion
Alcoholism is a common, heterogeneous disorder that, according to
Cloninger et al.,5 can be usefully divided into two relatively homogeneous
subgroups. The subgroup that consists of 25% of alcoholic men and is charac-
terized by early onset, antisocial personality traits, and a high degree of
heritability from fathers to sons may have a reduced central serotonin turn-
over rate. This disorder may share the same genetic background as antisocial
personality disorder. Early aggressiveness among boys, which is conducive to
the development of conduct disorder, may predispose these individuals to
the development of early-onset alcoholism and other substance abuse. Be-
cause of the high heritability and relative homogeneity of this disorder, it is a
prime target for molecular genetic investigations.
ACKNOWLEDGMENTS. The authors are grateful to David Goldman, MD and
Michael Eckardt, PhD for their thoughtful review and constructive comments
and to Ms. Andrea Hobbs for typing and editorial assistance.
References
1. Robins LN, Helzer JE, Weissman MM, et al: Lifetime prevalence of specific psychiatric
disorders in three sites. Arch Gen Psychiatry 41:948–958, 1984.
2. Schuckit M, Pitts FN, Reich T, et al: Alcoholism I: Two types of alcoholism in women. Arch
Gen Psychiatry 20:301–306, 1969.
3. Tarter RE, McBride H, Buonpane N, Schneider DU: Differentiation of alcoholics. Arch Gen
Psychiatry 34:761–768, 1977.
4. Pulkkinen L, Pitkanen T: A prospective study of the precursors to problem drinking in

young adulthood. J Stud Alcohol 55:578–587, 1994.
5. Cloninger CR, Bohman M, Sigvardsson S: Inheritance of alcohol abuse: Cross-fostering
analysis of adopted men. Arch Gen Psychiatry 38:861–868, 1981.
6. Lamparski DM, Roy A, Nutt DJ, Linnoila M: The criteria of Cloninger et al and von Knorring
et al for subgrouping alcoholics: A comparison in a clinical population. Acta Psychiatr Scand
84:497–502, 1991.
7. Sigrardsson S, Bohman M, Cloninger CR: Replication of the Stockholm adoption study of
alcoholism: Confirmatory cross-fostering analysis. Arch Gen Psychiatry 53:681–687, 1996.
8. Cloninger CR: The psychobiological regulation of social cooperation. Nature Med 7:623–625,
1995.
9. Babor TF, Hofmann M, DelBoca FK, et al: Types of alcoholics. I. Evidence for an empirically
derived typology based on indicators of vulnerability and severity. Arch Gen Psychiatry
49:599–608, 1992.
10. von Knorring AL, Bohman M, von Knorring L, Oreland L: Platelet MAO activity as a
biological marker in subgroups of alcoholism. Acta Psychiatr Scand 72:51–58, 1985.
11. Irwin M, Schuckit MA, Smith TL: Clinical importance of age at onset in type 1 and 2 primary
alcoholics. Arch Gen Psychiatry 47:320–324, 1990.
12. Gilligan SB, Reich T, Cloninger CR: Alcohol-related symptoms in heterogenous families of
hospitalized alcoholics. Alcohol Clin Exp Res 12:671–678, 1988.
13. Sullivan JL, Baenziger JC, Wagner DL, et al: : Platelet MAO in subtypes of alcoholism. Biol
Psychiatry 27:911–922, 1990.
14. Anthenelli RM, Smith TL, Irwin MR, Schuckit MA; A comparative study of criteria for
subgrouping alcoholics: The primary/secondary diagnostic scheme versus variations of the
type 1/type 2 criteria. Am J Psychiatry 151:1468–1474, 1994.
15. Penick EC, Powell BJ, Nickel EJ, et al: Examination of Cloninger’s type 1 and 2 alcoholism
with a sample of men alcoholics in treatment. Alcohol Clin Exp Res 14:623–629, 1990.
16. Hill SY: Absence of paternal sociopathy in the etiology of severe alcoholism: Is there a type 3
alcoholism? J Stud Alcohol 53:161–169, 1992.
17. Virkkunen M, DeJong J, Bartko J, et al: Relationship of psychobiological variables to recidiv-
ism in violent offenders and impulsive fire setters: A follow-up study. Arch Gen Psychiatry
46:600–603, 1989.
18. Virkkunen M, Kallio E, Rawlings R, et al: Personality profiles and state aggressiveness in
Finnish alcoholic violent offenders, fire setters and healthy volunteers. Arch Gen Psychiatry
51:28–33, 1994.
19. Turner WM, Cutter HSG, Worobec TG, et al: Family history models of alcoholism: Age of
onset, consequences and dependence. J Stud Alcohol 54:164–171, 1993.
20. Kaij L, Dock J: Grandsons of alcoholics: A test of sex-linked transmission of alcohol abuse.
Arch Gen Psychiatry 39:1379–1381, 1975.
21. Cloninger CR: Neurogenetic adaptive mechanisms in alcoholism. Science 236:410–416, 1987.
22. Linnoila M, DeJong J, Virkkunen M: Family history of alcoholism in violent offenders and
impulsive fire setters. Arch Gen Psychiatry 46:613–616, 1989.
23. Virkkunen M, Linnoila M: Brain serotonin, type 2 alcoholism and impulsive violence. J Stud
Alcohol 11:163–169, 1993.
24. Tiihonen J, Hakola P: Psychiatric disorders and homicide recidivism. Am J Psychiatry
151:436–438, 1994.
25. Moss HB, Mezzich A, Yao JK, et al: Aggressivity among sons of substance-abusing fathers:
Association with psychiatric disorder in the father and son, paternal personality, pubertal
development and socioeconomic status. Am J Drug Alcohol Abuse 21:195–208, 1995.
26. Virkkunen M: Alcoholism and antisocial personality. Acta Psychiatr Scand 59:493–501, 1979.
27. Cadoret RJ, Troughton E, Widmer R: Clinical differences between antisocial and primary
alcoholics. Compr Psychiatry 25:1–8, 1984.
28. Hesselbrock VM, Hesselbrock MN, Stabenau JR: Alcoholism in men patients subtyped by
family history and antisocial personality. J Stud Alcohol 46:59–64, 1985.
29. Schuckit MA: The clinical implications of primary diagnostic groups among alcoholics. Arch
Gen Psychiatry 42:1043–1049, 1985.
30. Liskow B, Powell BJ, Nickel E, Penick E: Antisocial alcoholics: Are there clinically significant
diagnostic subgroups? J Stud Alcohol 52:62–69, 1991.
31. Schuckit MA, Klein J, Twitchell G, Smith T: Personality test scores as predictors of alcohol-
ism almost a decade later. Am J Psychiatry 151:1038–1042. 1994.
32. Hesselbrock MN: Gender comparison of antisocial personality disorder and depression in
alcoholism. J Subst Abuse 3:205–219, 1991.
33. Schuckit MA: A clinical model of genetic influences in alcohol dependence. J Stud Alcohol
55:5–17, 1994.
34. Helmi S, Penick EG, Nickel EJ, et al: Early onset male alcoholics with and without antisocial
personality disorder, in 148th Annual Meeting Proceedings of the American Psychiatric Associa-
tion. Miami, APA Press, 1995, p. 91.
35. Fils-Aime M-L, Eckardt MJ, George DT, et al: Early-onset alcoholics have lower cerebro-
spinal fluid 5-hydroxyindoleacetic acid levels than late-onset alcoholics. Arch Gen Psychiatry
53:211–216, 1996.
36. Hesselbrock MN, Hesselbrock VM: Relationships of family history, antisocial personality
disorder and personality traits in young men at risk for alcoholism. J Stud Alcohol 53:619–
625, 1992.
37. Cloninger CR, Przybeck TR, Svrakic DM, Wetzel R: The Temperament and Character Inventory:
A Guide to Its Development and Use. St. Louis, Washington University Center for Psychobiol-
ogy and Personality, 1994.
38. Svrakic DM, Whitehead C, Przybeck TR, Cloninger CR: Differential diagnosis of personality
disorders by the seven-factor model of temperament and character. Arch Gen Psychiatry
50:991–999, 1993.
39. Svrakic N, Svrakic DM, Cloninger CR: A general quantitative model of personality develop-
ment: Fundamentals of a self-organizing psychobiological complex. Dev Psychopathol, in
press.
40. Cadoret RJ, O’Gorman TW, Troughton E, Heywood E: Alcoholism and antisocial person-
ality: Interrelationships, genetic and environmental factors. Arch Gen Psychiatry 42:161–167,
1985.
41. Cadoret RJ, Troughton E, O’Gorman TW: Genetic and environmental factors in alcohol
abuse and antisocial personality. J Stud Alcohol 48:1–8, 1987.
42. Mayfield D, McLeod G, Hall P: The CAGE questionnaire: Validation of a new alcoholism
screening instrument. Am J Psychiatry 131:1121–1123, 1974.
43. McCord W, McCord J: Origins of Alcoholism. Stanford, CA, Stanford University Press, 1960.
44. Donovan JE, Jessor R, Jessor L: Problem drinking in adolescence and young adulthood: A
follow-up study. J Stud Alcohol 44:109–137, 1983.
45. Barnes GM, Welte JW: Patterns and predictors of alcohol use among 7–12th grade students
in New York State. J Stud Alcohol 47:53–62, 1986.
46. Anderson T, Bergman LR, Magnusson D: Patterns of adjustment problems and alcohol abuse
in early adulthood: A prospective longitudinal study. Dev Psychopathol 1:119–131, 1989.
47. Robins LN, McEvoy L: Conduct problems as predictors of substance abuse, in Robins LN,
Rutter M (eds): Straight and Devious Pathways from Childhood to Adulthood. New York, Cam-
bridge University Press, 1990, pp 182–204.
48. van Kammen WB, Loeber R, Stouthammer-Loeber M: Substance use and its relationship to
conduct problems and delinquency in young boys. J Youth Adolesc 20:399–413, 1991.
49. Boyle MH, Offord DR, Racine YA, et al: Predicting substance use in late adolescence: Results
from the Ontario Child Health Study follow-up. Am J Psychiatry 149:761–767, 1992.
50. Pulkkinen L: Self-control and continuity from childhood to late adolescence, in Baltes PB,
Brim OG Jr (eds): Life-span Development and Behavior. San Diego, Academic Press, 1982, vol 4,
pp 63–105.
51. Mehlman P, Higley J, Faucher J, et al: Low 5-HIAA concentrations and severe aggression
and impaired impulse control in nonhuman primates. Am J Psychiatry 151:1485–1491, 1994.
52. Virkkunen M, Eggert M, Rawlings R, Linnoila M: A prospective follow-up study of alcoholic

violent offenders and fire setters. Arch Gen Psychiatry 53:523–529, 1996.
53. Kruesi MJP, Hibbs ED, Zahn TP, et al: A 2-year prospective follow-up study of children and
adolescents with disruptive behavior disorders: Prediction by cerebrospinal fluid 5-hy-
droxyindoleacetic acid, homovanillic acid and autonomic measures? Arch Gen Psychiatry
49:429–435, 1992.
54. Virkkunen M, Rawlings R, Tokola R, et al: CSF biochemistries, glucose metabolism, and
diurnal activity rhythms in violent offenders, impulsive fire setters, and healthy volunteers.
55. Virkkunen M: Urinary free cortisol secretion in habitually violent offenders. Acta Psychiatr
Scand 72:40–44, 1985.
56. Bergman B, Brismar B: Hormone levels and personality traits in abusive and suicidal male
alcoholics. Alcohol Clin Exp Res 18:311–316, 1994.
57. King RJ, Scheuer JW, Curtis D, Zarkone VP: Plasma cortisol correlates of impulsivity and
substance abuse. Person lndivid Diff 3:287–291, 1990.
58. Roy A, DeJong J, Lamparski D, et al: Mental disorders among alcoholics: Relationship to age
of onset and cerebrospinal fluid neuropeptides. Arch Gen Psychiatry 48:423–427, 1991.
59. Virkkunen M, Linnoila M: Serotonin in early onset, male alcoholics with violent behavior.
Ann Med 22:327–331, 1990.
60. Limson R, Goldman D, Roy A, et al: Personality and cerebrospinal fluid monoamine metab-
olites in alcoholics and controls. Arch Gen Psychiatry 48:437–441, 1991.
61. Linnoila M, Virkkunen M, Scheinin M, et al: Low CSF 5-HIAA concentration differentiates
impulsive from nonimpulsive violent behavior. Life Sci 33:2609–2614, 1983.
62. Ballenger J, Goodwin F, Major L, Brown G: Alcohol and central serotonin metabolism in
man. Arch Gen Psychiatry 36:224–227, 1979.
63. Banki C: Factors influencing monoamine metabolites and tryptophan in patients with alco-
hol dependence. J Neural Transm 50:98–101, 1981.
64. Takahashi S, Yamane H, Kondo H, Tani M: CRF monoamine metabolites in alcoholism, a
comparative study with depression. Folia Psychiat Neurol Japan 28:347–354, 1974.
65. Benkelfat C, Murphy DL, Hill JL, et al: Ethanollike properties of the serotonergic partial
agonist m-chlorophenylpiperazine in chronic alcoholic patients. Arch Gen Psychiatry 48:383,
1991.
66. Krystal JH, Webb E, Cooney N, et al: Specificity of ethanollike effects elicited by serotonergic
and noradrenergic mechanisms. Arch Gen Psychiatry 51:898–911, 1994.
67. George DT, Benkelfat C, Hill JL, et al: A comparison of behavioral and biochemical re-
sponses to mCPP in subtypes of alcoholics. Am J Psychiatry, 1996, in press.
68. Moss HB, Yao JK, Panzak GL: Serotonergic responsivity and behavioral dimensions in
antisocial personality disorder with substance abuse. Biol Psychiatry 28:325–328, 1990.
69. Fernstrom JD, Wurtman RJ: Brain serotonin content: Increase following ingestion of car-
bohydrate diet. Science 174:1023–1025, 1971.
70. Fernstrom JD, Wurtman RJ: Brain serotonin content: Physiological regulation by plasma
neural amino acids. Science 178:414–416, 1972.
71. Buydens-Branchey L, Branchey MH, Noumair D, Lieber CS: Age of alcoholism onset: Rela-
tionship to susceptibility to serotonin precursor availability. Arch Gen Psychiatry 46:231–236,
1989.
72. Branchey L, Branchey M, Zucker D, et al: Association between low plasma tryptophan and
blackouts in male alcoholic patients. Alcohol Clin Exp Res 9:393–395, 1985.
73. Virkkunen M, Narvanen S: Plasma insulin, tryptophan and serotonin levels during the
glucose tolerance test among habitually violent and impulsive offenders. Neuropsychobiology
17:19–23, 1987.
74. Nielsen DA, Goldman D, Virkkunen M, et al: Suicidality and 5-hydroxyindoleacetic acid
concentration associated with a tryptophan hydroxylase polymorphism. Arch Gen Psychiatry
51:34–38, 1994.
75. Fowler CJ, Oreland L: Human platelet monoamine oxidase: Some biochemical findings, in
Kamijo K, Usdin K, Nagatsu T (eds): Monoamine Oxidase: Basic and Clinical Frontiers. Amster-
dam, Excerpta Medica, 1981, pp 28–39.
76. Fowler CJ, Tipton KF, MacKay AVP, Youdim MBH: Human platelet monoamine oxidase: A
useful enzyme in the study of psychiatric disorders. Neuroscience 7:1577–1594, 1982.
77. Young WF, Laws ER, Sharbrough PW, Weinshilboum RM: Human monoamine oxidase:
Lack of brain and platelet correlation. Arch Gen Psychiatry 43:604–609, 1986.
78. Oreland L: The activity of human brain and thrombocyte monoamine oxidase (MAO) in
relation to various psychiatric disorders: I. MAO activity in some disease states, in Singer T,
von Korff R, Murphy D (eds): Monoamine Oxidase: Structure, Function and Altered Functions.
New York, Academic Press, 1979, pp 379–387.
79. Oreland L, Shaskan EG: Monoamine oxidase activity as a biological marker. Trends Phar-
macol Sci 4:339–341, 1983.
80. Wiberg A, Gottfries CG, Oreland L: Low platelet MAO activity in human alcoholics. Med
Biol 55:181–186, 1977.
81. Faraj BA, Lenton JD, Kutner M, et al: Prevalence of low monoamine oxidase function in
alcoholism. Alcohol Clin Exp Res 11:464–467, 1987.
82. Faraj BA, Davis DC, Camp VM, et al: Platelet monoamine oxidase activity in alcoholics,
alcoholics with drug dependence, and cocaine addicts. Alcohol Clin Exp Res 18:1114–1120,
1994.
83. Pandey GN, Fawcett J, Gibbons R, et al: Platelet monoamine oxidase in alcoholism. Biol
Psychiatry 24:15–24, 1988.
84. Mukassa H, Nakamura J, Yamada S, et al: Platelet monoamine oxidase activity and person-
ality traits in alcoholics and metamphetamine dependents. Drug Alcohol Depend 26:251–254,
1990.
85. Yates WR, Wilcox J, Knudson R, et al: The effect of gender and subtype of platelet MAO in
alcoholism. J Stud Alcohol 51:463–467, 1990.
86. Anthenelli RM, Smith TL, Craig CE, et al: Platelet monoamine oxidase levels in subgroups of
alcoholics: Diagnostic, temporal and clinical correlates. Biol Psychiatry 38:361–368, 1995.
87. Sherf F, Hallman J, Oreland L: Low platelet gamma-aminobutyrate aminotransferase and
monoamine oxidase activities in chronic alcoholic patients. Alcohol Clin Exp Res 16:1014–
1020, 1992.
88. Devor EJ, Cloninger CR, Kwan SW, Abell CW: A genetic familial study of monoamine
oxidase B activity and concentration in alcoholics. Alcohol Clin Exp Res 17:263-267, 1993.
89. von Knorring L, Oreland L, von Knorring AL: Personality traits and platelet MAO activity in
alcoholic and drug abusing teenage boys. Acta Psychiatr Scand 75:307–314, 1987.
90. von Knorring AL, Hallman J, von Knorring L, Oreland L: Platelet MAO oxidase activity in
type 1 and 2 alcoholism. Alcohol Alcohol 26:409–416, 1991.
91. Lidberg L, Modin I, Oreland L, et al: Low platelet monoamine oxidase activity and psycho-
pathy. Psychiatr Res 16:339–343, 1985.
92. Yehuda R, Southwick SM, Edell WS, Giller EL Jr: Low platelet monoamine oxidase activity
in borderline personality disorder. Psychiatry Res 30:265–273, 1989.
93. von Knorring L, Oreland L, Winblad B: Personality traits related to monoamine oxidase
(MAO) in platelets. Psychiatry Res 12:11–26, 1984.
94. Rutter M: Concluding remarks in Rutter M (ed): Genetics of Criminal and Antisocial Behavior.
Chichester, Wiley, 1996, 265–271.
95. Gabel S, Stadler J, Bjorn J, et al: Homovanillic acid and monoamine oxidase in sons of
substance-abusing fathers: Relationships to conduct disorder. J Stud Alcohol 56:135–139,
1995.
96. Stoff DM, Friedman E, Pollock L, et al: Elevated platelet MAO is related to impulsivity in
disruptive behavior disorders. J Am Acad Child Adolesc Psychiatry 28:754–760, 1989.
97. Shekim WO, Hodges K, Horwitz E, et al: Psychoeducational and impulsivity correlates of
platelet MAO in normal children. Psychiatry Res 11:99–106, 1984.
98. Bowden CL, Deutsch CK, Swanson JM: Plasma dopamine beta-hydroxylase and platelet
monoamine oxidase in attention deficit disorder and conduct disorder. J Am Acad Child
Adolesc Psychiatry 27:171–174, 1988.
99. Coccaro EF, Siever LJ, Klar H, et al: Serotonergic studies of personality disorder: Correlates
with behavioral aggression and impulsivity. Arch Gen Psychiatry 46:587–589, 1989.
100. Halperin JM, Sharma V, Siever LJ, et al: Serotonergic function in aggressive and nonaggres-
sive boys with attention deficit hyperactive disorder. Am J Psychiatry 151:243–248, 1994.
101. Gabel S, Stadler J, Bjorn J, et al: Monoamine oxidase and homovanillic acid in male youth
with predisposition to substance abuse. Alcohol Clin Exp Res 18:1137–1142, 1994.
102. American Psychiatric Association, Committee on Nomenclature and Statistics: Diagnostic
and Statistical Manual of Mental Disorders, 4th ed. Washington, DC, American Psychiatric
Press, 1994.
103. Meltzer HY, Arora RC: Platelet serotonin studies in affective disorders: Evidence for a
serotonergic abnormality? in Sandler M, Coppen A, Harnett S (eds): 5-Hydroxytryptamine in
Psychiatry: A Spectrum of Ideas. New York, Oxford University Press, 1991, pp 50–89.
104. Rausch JL, Monteiro MG, Schuckit MA: Platelet serotonin uptake in men with family histo-
ries of alcoholism. Neuropsychopharmacology 4:83–86, 1991.
105. Ernouf D, Compagnon P, Lothion P, et al: Platelet 3H 5HT uptake in descendants from
alcoholic patients: A potential risk factor for alcohol dependence? Life Sci 52:989–995, 1993.
106. Boismare F, Lhuintre JP, Daoust M, et al: Platelet affinity for serotonin is increased in
alcoholics and former alcoholics. Alcohol Alcohol 22:155–159, 1987.
107. Daoust M, Lhuintre JP, Ernouf D, et al: Ethanol intake and 3H-serotonin uptake. II: A study
in alcoholic patients using platelets 3H-paroxetine binding. Life Sci 48:1977–1983, 1991.
108. Neiman J, Bewing H, Malmgren R: Platelet uptake of serotonin (5HT) during ethanol with-
drawal in male alcoholics. Thromb Res 46:803–809, 1987.
109. Kent TA, Campbell JL, Pazdernik TL, et al: Blood platelet uptake of serotonin in men
alcoholics. J Stud Alcohol 46:357–359, 1985.
110. Brown CS, Kent TA, Bryant SG, et al: Blood platelet uptake of serotonin in episodic aggres-
sion. Psychiatry Res 27:5–12, 1989.
111. Lappalainen J, Dean M, Charbonneau L, et al: Mapping of the serotonin 5-HT1DB autorecep-
tor gene on chromosome 6 using a coding region polymorphism. Am J Med Genet 60:157–
161, 1995.
112. Lappalainen J, Zhang L, Dean M, et al: Identification, expression and pharmacology of a
cys23-ser23 substitution in the human 5HT2c receptor gene. Genomics 27:274–279, 1995.
113. Ozaki N, Lappalainen J, Dean M, et al: Mapping of the serotonin 5HT1Da autoreceptor gene
(5HTR1D) on chromosome 1 using a silent polymorphism in the coding region. Am Med
Genet 60:162–164, 1995.
114. Nielsen D, Nakhai B, Schuebel K, et al: Tryptophan hydroxylase and the 5HT1a receptor:
Behavioral and gene expression studies, in Proceedings of the 33rd Annual Meeting of the
American College of Neuropsychopharmacology. San Juan, ACNP, 1994, p. 28.
115. Lappalainen J, Virkkunen M, Dean M, et al: Identification of a cys23-ser23 substitution in the
5HT2c receptor gene and allelic association to violent behavior and alcoholism. Manuscript
in preparation.
116. Vanyukov MM, Moss HB, Yu LM, et al: Preliminary evidence for an association of di-
nucleotide repeat polymorphism at the MAOA gene with early onset alcoholism/substance
abuse. Am J Med Genet 60:122–126, 1995.
117. Tiihonen J, Kuikka J, Bergstrom K, et al: Altered striatal dopamine reuptake site densities in
habitually violent and nonviolent alcoholics. Nature Med 7:664–657, 1995.
9
A Nonhuman Primate Model

of Excessive Alcohol Intake
Personality and Neurobiological Parallels
of Type I- and Type II-Like Alcoholism
J. Dee Higley and Markku Linnoila
Abstract. Developmental, biochemical, and behavioral concomitants of voluntary excessive al-

cohol consumption were investigated using a nonhuman primate model. Studies were designed
to investigate potential neurobiological and behavioral parallels of Cloninger’s subtypes of type I
and type II alcoholism in nonhuman primates. The studies have shown that a subpopulation of
primates chronically consume intoxicating amounts of alcohol. Subjects that chronically consume
intoxicating amounts of alcohol often exhibit neurobiological and behavioral features that were
predicted by Cloninger’s model for subtypes of alcoholism among humans. Investigations
showed that behavior patterns and biological indices that characterize high anxiety, whether
constitutionally or stress induced, were correlated with high rates of alcohol consumption, con-
sistent with predictions for type I alcoholism. Early untoward rearing experiences that increased
anxiety increased the probability that subjects would chronically drink alcohol to intoxication.
Investigations of type II-like alcohol consumption patterns focused on subjects with low central
nervous system (CNS) serotonin functioning [as measured by reduced cerebrospinal fluid (CSF)
concentrations of the serotonin metabolite 5-hydroxyindoleacetic acid (5-HIAA)]. CSF 5-HIAA in
infancy was shown to be a relatively stable neurobiological trait across development into adult-
hood. An individual CSF 5-HIAA concentration in infancy was shown to be a consequence of
paternal and maternal genetic influences. Early parental neglect reduced CSF 5-HIAA concentra-
tions. Low CSF 5-HIAA and CNS norepinephrine functioning were shown to predict excessive
alcohol consumption in adolescence. Behaviorally, subjects with low CSF 5-HIAA demonstrated
J. Dee Higley • Laboratory of Clinical Studies, Primate Unit, National Institute on Alcohol
Abuse and Alcoholism, Poolesville, Maryland 20837. Markku Linnoila • Laboratory of Clini-
cal Studies, Division of Intramural Clinical and Biological Research, National Institute on Alcohol
Abuse and Alcoholism, Bethesda, Maryland 20892-1256.
191
impaired impulse control, which resulted in excessive and inappropriate aggression, infrequent
and inept social behaviors, low social status, social isolation and expulsion from social groups at
an early age, and high rates of early mortality. With some exceptions, these findings were
consistent with predictions from Cloninger’s type II model of excessive alcohol consumption
among men who exhibit impaired impulse control and violent and antisocial behaviors.
The time has come for a study of inebriety from a

medical stand-point, and when it is treated as a special
disease its curability will be found equal to any other
disease.
—Franklin D. Clum, MD, 1892
1. Introduction
During the past two decades, remarkable progress has been made in the
study of alcohol abuse and alcoholism. As the field has become increasingly
refined, more accurate classification and typology of patients have become
possible. With the advances of classification, there has been the recognition
that alcohol abuse and alcoholism are heterogeneous conditions with multiple
etiological components.1–3 Biological and psychological research has identi-
fied relatively homogeneous subgroups of alcohol-abusing and dependent
patients, with each group exhibiting somewhat different psychobiological
traits. Cloninger’s neurogenetic, tridimensional theory of personality struc-
ture is among the most influential of the current psychobiological models.4
Cloninger identified two subtypes of alcoholism. The first, labeled type I,
is characterized by high levels of traitlike anticipatory anxiety. Individuals
with type I alcoholism are postulated to consume alcohol primarily for its
antianxiety properties.4,5 Excessive anxiety and the resulting increased alco-
hol consumption patterns are hypothesized to result from an interaction be-
tween untoward rearing experiences and genetic background. Type II alco-
holism, on the other hand, is distinguished by impaired impulse control
resulting in excessive alcohol consumption. Initial consumption is thought to
be primarily motivated by the euphorogenic effects of alcohol, followed by
high rates of alcohol consumption that result from loss of control once alcohol
consumption begins. A cluster of behaviors related to impaired impulse con-
trol characterizes individuals with type II alcoholism. Essentially, Cloninger
portrays the behaviors of type II alcoholics as physically aggressive, risk tak-
ing, and having difficulties functioning socially4,6 (see ref. 7, for example).
Each of the two types of alcoholism is proposed to have a different neuro-
genetic background, with anxiety-mediated type I alcoholism based primarily
on a CNS norepinephrine excess, and impulse-mediated type II alcoholism
based primarily on a CNS serotonin deficit.4
Because the systematic study of alcoholism is a relatively new discipline,
9 • A Nonhuman Primate Model 193
until recently our understanding of the syndrome has been somewhat sparse.
Animal models of human diseases are limited by the current understanding
of the pathophysiology of the disease in question. To model a disease, its
symptoms, causal, and precipitating factors must be defined in humans with
such precision that they can be reproduced in another species.8–10 An animal
model is only useful to the extent that the essential features of the disease to
be modeled are clearly defined and potential causal mechanisms are delin-
eated. This is particularly difficult for syndromes, which typically have nu-
merous symptoms along with multiple potential causal and precipitating fac-
tors. Advances in defining the necessary criteria and symptomatology for
alcohol abuse and alcoholism have allowed researchers to develop animal
models of the syndromes and to investigate the basic mechanisms of their
components.11–15
2. Why Study Nonhuman Primates?

Recently, nonhuman primates have been used to model some features of
alcohol abuse and alcoholism.16–20 Nonhuman primates are our closest phy-
logenetic relatives, and as a result share a large percentage of their DNA with
humans. These similarities in DNA makeup yield physiological, neuro-
anatomical, and behavioral similarities, if not actual homologies, that allow
researchers to extrapolate their results to the human condition more readily
than with results obtained from less closely related animal species. Many of
the precursors and consequences of alcohol problems are social in nature.
Like humans, the typical primate is a social being, living in a complex society
with multiple social stressors and social responsibilities. Because of these
similarities, nonhuman primates are particularly suited as research subjects
for modeling the antecedents and concomitants of human alcohol problems.
3. Difficulties Producing a Nonhuman Primate Model

of Alcohol Abuse
Given these advantages for using nonhuman primates to study alcohol
abuse and alcoholism, it is surprising that the use of nonhuman primates to
model alcohol abuse and dependence is a relatively new phenomenon. There
are indications that this is a result of both theoretical and practical problems.
For alcohol abuse and alcoholism, only recently have the descriptive data and
our understanding of the pathophysiology achieved an adequate level of
sophistication to allow the development of standardized diagnostic sys-
tems.21–24 With these recent advances, nonhuman primate models relevant to
alcohol abuse have begun to be developed.8,18,25
An independent limiting factor that has impeded the development of
nonhuman primate models of alcohol abuse has been the difficulty in produc-
ing high levels of voluntary consumption of alcohol among normal nonhu-

man primates. Until recently, it was widely believed that nonhuman primates
would not consume alcohol (e.g., refs. 16, 26, 27). Recent studies have
shown, however, that at least for some species of nonhuman primates, when
the alcohol solution is palatable and freely available, some (but not all) sub-
jects will readily consume it in quantities that produce pharmacological ef-
fects. In many cases individual subjects will consume sufficient quantities of
alcohol to produce blood levels exceeding the limits of legal intoxication for
most states of the United States, resulting in stupor and at times unconscious-
ness.18,28 Such recent observations have prompted investigators to use non-
human primates in research on alcohol abuse and alcoholism.
4. Human and Nonhuman Primate Parallels in Alcohol Use

There are a number of demographic and epidemiological descriptions of
alcohol consumption patterns in human society. Most humans find the initial
taste of alcohol aversive, and, as a result, in human society alcohol is rarely
consumed in its pure state. Instead, alcohol is consumed in solutions with
low concentrations and often the taste is disguised using colas, fruit juices,
and other flavorings. Despite its aversive taste properties, most people who
try alcohol persist either because of expectations concerning its effects, social
pressure, or because they find they enjoy its pharmacological effects.29–31
Eventually, most individuals develop a pattern of modest consumption, with
minimal alcohol-related problems.32,33 Similarly, in at least some species of
nonhuman primates, when the solution is palatable and the concentration of
alcohol is under 15–20%, most subjects will consume alcohol at rates produc-
ing pharmacological effects, but only about 10–20% of normally reared sub-
jects will freely consume palatable alcohol solutions at rates that consistently
produce blood alcohol levels greater than the legal level of intoxication for
most states.19,25,28 Because only a few individual subjects consume alcohol at
high rates, a problem that researchers face when using nonhuman primates is
that they must have access to a large population of monkeys to identify a pool
of subjects to study excessive alcohol consumption.
5. Methodology
Over the course of the past 4 years, more than 70 rhesus monkeys from
our laboratory have been allowed to consume alcohol while living socially in
their home cages. To perform this research, the monkeys are provided unfet-
tered access to a palatable water–alcohol solution (8.5% v/v, flavored to taste
with aspartame and colored with food coloring) for 1 hr each day while they
are in their home environment. To precipitate initial consumption of alcohol
at rates that produce pharmacological effects, subjects are first trained to
drink a colored aspartame-water vehicle by hanging a bottle or a burette on
the side of the cage and connecting it to a drinking spout. When all of the
subjects are drinking the sweetened solution, the color is changed, and suffi-
cient alcohol is added to make an 8.5% v/v alcohol solution. Over a 30- to 10-
day period, most subjects consume alcohol at rates that produce phar-
macological effects. Over an additional 5–15 days, individual response rates
are established and the data collection phase of the experiment begins.
Subjects are not food- or water-deprived throughout their exposure, and
they have alcohol available 5 days a week for 1 hr each day. To assure that the
subjects are not drinking the alcohol solution just for its sweetened taste,
subjects are provided simultaneous access to the sweetened vehicle, tap wa-
ter, and the alcohol solution. In all of our studies to date, the first 10 days of
the experimental paradigms have served as baseline periods and are identical
in methodology except for the social arrangement of the home cage where the
subjects have access to alcohol. The subjects are provided alcohol in their
home cages, which include housing in pairs (both single and mixed gender
pairs), social group housing in a pen containing 7 to 10 other monkeys, and
occasionally in single cages. Those subjects living in social groups are main-
tained under stable social conditions, having lived in their social groups for at
least a year prior to the beginning of the study. Because rhesus monkeys
maintain a social dominance hierarchy that probably affects free access to the
alcohol solution,18,28,34 social groups receiving alcohol are subdivided accord-
ing to relative dominance rankings. The cage is physically divided into small-
er subsections, and subjects receiving alcohol are alternatively placed into
high or low social dominance portions of the cage. Multiple drinking spouts
are present in each subdivision of the cage. When subjects are tested in pairs,
an alcohol dispenser is provided for each monkey: typically, one in the front
and one in the back of the cage. Our paradigm has been designed to provide
all subjects with sufficient time and opportunity to drink alcohol during each
session; indeed, during the majority of the second half-hour-long alcohol
session, the drinking spouts are not occupied.
6. Biological and Behavioral Measures

All behavioral and biochemical observations are obtained when the sub-
jects are alcohol-free. In our studies, repeated cisternal CSF samples are col-
lected from each of the subjects to measure CNS monoamine functioning.
Simultaneous blood samples are drawn to quantify stress hormones. Behav-
ioral observations are recorded during home-cage baseline conditions and
during a stressful period such as social separation by trained observers who
are blind to the subjects’ monoamine status and specific hypotheses of the
study. All cisternal CSF and blood samples are obtained between 1300 and
1500. All behavioral data are collected in the morning between 0900 and 1130,
before alcohol is dispensed, using a standardized behavior scoring system
described in detail in the original articles.35
7. Studies of Interindividual Differences
Studies by us and others have shown that although the rates of consump-
tion vary between individual monkeys, once rates of consumption stabilize,
average interindividual differences in alcohol consumption are markedly sta-
ble over time,18 showing that the underlying motivation to consume alcohol
may be traitlike. This suggests the possibility that within nonhuman primate
societies, the 10–20% of individual nonhuman primates that consume alcohol
at high rates may be homologous to the 10–20% of humans who at some
period of their life abuse alcohol.36 Researchers studying alcohol consump-
tion in nonhuman primates have made significant progress in developing a
nonhuman primate model of alcohol abuse by focusing on those subjects that
show high rates of alcohol consumption.
8. Studies Investigating Features of Type I Alcoholism

The tranquilizer of greatest value since the early
history of man, and which may never become
outdated, is alcohol, when administered in
moderation,
—Nathan Masor37
Nonhuman primates have been used with increasing frequency to investi-

gate features of type I, or anxiety-related alcoholism. Some of the first re-
searchers to investigate alcohol consumption in nonhuman primates indicated
that under certain conditions stress may increase alcohol consumption.16,38,39
Subsequent systematic studies using adolescent rhesus macaques have shown
that social stress increases alcohol consumption, and that during nonstressful
periods interindividual rates of alcohol consumption are positively correlated
with interindividual traitlike anxiety and fearfulness.18 For example, in these
studies, interindividual differences in alcohol consumption were positive cor-
related with anxietylike behaviors such as self-directed orality and clasping of
the body. Moreover, biological measures of stress and anxiety such as plasma
corticotropin are also positively correlated with alcohol consumption.18
Research has further delineated the conditions and the underlying psy-
chobiology related to stress-induced excessive alcohol consumption. This re-
search has shown that alcohol consumption may attenuate the response to a
stressful event. One measure that is quite sensitive to stress is social play. It
typically declines following a number of different stressors.40–43 In one study,
social play declined to an almost nonexistent level following the exposure to a
stressor. However, when a moderate dose of alcohol was given to the
stressed monkeys, social play increased to and even exceeded the baseline
levels.44 Others have also found that a modest-to-moderate dosage of alcohol
reduces adverse consequences of stress, resulting in increased social play and
other positive social behaviors.45 As levels of intoxication increase, however,
play decreases below baseline levels.38
Play is not the only behavior reduced by stress that returns to baseline
levels following alcohol consumption. Ervin and colleagues28 found that in
vervet monkeys there was not a uniform increase in play following the con-
sumption of alcohol; rather, the idiosyncratic behavioral patterns of individu-
al subjects that were suppressed by stress were the same behaviors that in-
creased following alcohol consumption. Further evidence of alcohol’s stress-
reducing properties was seen when subjects underwent the stress of a social
separation. During periods of social separation, rates of despair and anxiety-
like behaviors were attenuated by low doses of alcohol.44 This finding is con-
sistent with a body of literature from rodent studies showing the similarities
between benzodiazepines and alcohol in attenuating anxiety,46 and suggests
that at low to moderate doses, alcohol possesses anxiolytic effects. It is note-
worthy, however, that while small doses of alcohol may attenuate stress, high
doses of alcohol were more likely to exacerbate anxiety and depressionlike be-
haviors.44 Interestingly, while it is widely held that alcohol use is related to or
may exacerbate depressive symptoms in humans, in their relatively exhaus-
tive review of laboratory studies that assessed the activating and depressing
effects of alcohol, Tucker and colleagues47 noted that at blood levels below
10%, alcohol typically acts as an antidepressant. While high doses are seldom
used in the laboratory, the few studies that have shown increased depressive
symptoms following alcohol consumption have utilized high doses.47 This
again suggests the utility of the primate model to investigate alcohol effects,
particularly in cases where ethical concerns, such as administering high doses
of alcohol to volunteers, make human studies impossible to perform.
Probably because alcohol has stress-reducing properties, nonhuman pri-
mates typically consume more alcohol in stressful settings than in otherwise
identical but less stressful settings.18,19,38 Such stress-induced alcohol con-
sumption may be substantial. For example, following a social separation
stressor, adolescent subjects reared in normal settings double their rates of
alcohol consumption, often to levels that produce blood alcohol concentra-
tions in excess of 100 mg/dl, the limit of legal intoxication in most of the United
States.18,19 Nevertheless, some subjects’ rates of alcohol consumption seem
relatively unaffected by increased stress; still others seem particularly prone
to increase their consumption during even minimal stress. Studies of nonhu-
man primates indicate that these differences in how stress affects alcohol
consumption may be, at least in part, due to early experiences that serve to
decrease the threshold for experiencing anxiety. One of the best illustrations
of this comes from studies of rhesus macaques reared under what is termed a
peer-only or peer-rearing condition. These subjects are reared from birth with
constant access to other same-aged peers but without mothers or any adults.
In the absence of adult guidance, these peer-reared monkeys develop trait-
like, chronic anxiety.35,48,49 From infancy and into adolescence, they exhibit
chronic activation of the hypothalamic-pituitary-adrenal axis.18,50–52 They
are more likely to show fearfulness and anxietylike behavior in the face of a
challenge than mother-reared controls.35,48,49 While other factors probably
play a role, at least in part as a result of this predisposition to anxiety and fear,
Figure 1. An illustration of the effects of early rearing experiences and social separa-
tion on alcohol consumption (n = 22, F(3.60) = 5.02, P = 0.007). Each bar represents
the average and SD of alcohol consumption in grams per kilogram of body weight for
each group over experimental conditions. Solid bar represents peer-reared and open
bar represents mother-reared subjects. The preseparation baseline period is the aver-
age of 10 days of home-cage consumption. The average consumption of the four
separations is divided into an overall acute phase (mean of first day of each of the four
separations) and an overall chronic phase (mean of remaining 3 days of each separa-
tion). The postseparation recovery phase is the average of 10 days of alcohol consump-
tion following the social separations. A significant difference between the peer-reared
subjects and the mother-reared subjects within the same period, with the peer-reared
subjects showing an increased consumption (P < 0.05). @, A significant increase in
alcohol consumption for the mother-reared subjects during social separation relative
to home-cage consumption (P < 0.05). The apparent reduction in alcohol consumption
by the peer-reared monkeys during the chronic phase of the social separations is not
statistically significant (T = 1.56, P > 0.10). (From Higley et al.18)
peer-reared subjects are more likely than mother-reared controls to consume

alcohol at rates that produce intoxication when alcohol is freely available in
the home cage.18 On the other hand, when a stressor such as a social separa-
tion is applied, the mother-reared subjects, who normally consume only lim-
ited amounts of alcohol, increase their levels of consumption to levels that
equal those of the peer-reared monkeys18 (see Fig. 1).
Interestingly, within each of the rearing groups there are wide interin-
dividual differences in alcohol consumption rates. A few of the mother-reared
subjects consume alcohol at rates similar to the peer-reared subjects; on the
other hand, some of the peer-reared subjects seem relatively unaffected by the
early rearing experience and consume alcohol at rates similar to the mother-
reared subjects. These within-group interindividual differences in alcohol con-
sumption are maintained across settings and time.18 Clues suggesting the
possible genesis of these within-group individual differences come from mea-
suring the anxietylike behaviors and biological markers indicative of stress
responsiveness and anxiety prior to alcohol exposure. Interindividual differ-
ences in both anxietylike behaviors and biological markers of anxiety correlate
positively with interindividual differences in alcohol consumption. For exam-
ple, there is a positive correlation between individual alcohol consumption rate
and the severity of prestress anxietylike behaviors. These anxietylike behaviors
are also positively correlated with plasma cortisol.18 As might be predicted,
plasma cortisol is positively correlated with alcohol consumption rate.18 These
findings suggest that the intrinsic predisposition to experience anxiety when
stressed may be related to differences in alcohol consumption rates. Kraemer et
al.20 also found evidence for Cloninger’s prediction of high CNS nor-
epinephrine functioning among rhesus monkeys exhibiting type I-like exces-
sive alcohol consumption. According to these authors, nonhuman primates
that show large increases in CSF norepinephrine concentrations following an
alcohol bolus are more likely to show a high rate of stress-induced alcohol
consumption relative to those that exhibit only minimal CSF norepinephrine
increases.20 These results suggest possible reasons why under apparently
identical stressful conditions some subjects increase their alcohol consump-
tion, while others appear to be relatively invulnerable to the same stressor.
These findings indicate that early experiences that result in increased
levels of anxietylike behaviors can have a major impact on alcohol consump-
tion. Studies of human alcoholics have shown that for type I alcoholism early
developmental experiences are important factors in determining alcohol abuse
patterns.4 The results suggest that early rearing experiences, which predis-
pose monkeys to increased fear-related behaviors, produce excessive alcohol
consumption under normal living conditions. Furthermore, a major stressful
challenge such as social separation increases alcohol consumption to levels
producing intoxication even in monkeys not particularly vulnerable to stress.
9. Serotonin and Type II Alcoholism

At the moment, serotonin is the hottest item (in this
author’s opinion) in the search for a biochemical
explanation for what appears to be a genetically
determined transmission of alcoholism.
—Donald W. Goodwin53
Recently, nonhuman primates have been used to model features of type

II alcoholism. These studies used group-living adolescent and adult rhesus
monkeys to investigate high alcohol consumption in subjects with reduced
CNS serotonin functioning.54–57 As noted earlier, a distinguishing neuro-

biological feature of type II alcoholism is impaired impulse control, which is
believed to result in part from impaired CNS serotonin functioriing.7 The
underlying theory of Cloninger’s model of type II alcoholism postulates that
initial alcohol consumption among individuals with the vulnerability is pri-
marily motivated by the euphorogenic effects of alcohol; but because these
individuals are unable to curb their impulses, once consumption begins, loss
of controlled drinking results. In his original formulation and other publica-
tions, Cloninger discusses an overall behavioral style that characterizes indi-
viduals with type II alcohol problems as having a pattern of behaviors that are
related to impaired impulse control, such as physically aggressive behaviors,
antisocial traits, excessive risk taking, and difficulties in social relation-
ships.4,6 Unlike the risk for expressing type I alcoholism, the risk for expres-
sing type II alcoholism appears to be primarily genetically transmitted and
relatively unaffected by early experiences.
10. Reduced CNS Serotonin Functioning as a Long-Term

Enduring Trait
A principal neurobiological feature of type II alcoholism is a CNS seroto-

nin deficit.4 This is typically measured by assaying CSF for concentrations of
the major serotonin metabolite, 5-hydroxyindoleacetic acid (5-HIAA). Several
nonhuman primate studies have suggested that interindividual differences in
CSF 5-HIAA concentrations are traitlike (i.e. , highly stable across time and
experimental settings). For example, when 14 CSF samples were obtained
over a 1-year period from the same adult female subjects, interindividual
stability was high, with the average correlation coefficient in excess of r =
.51.58 When Raleigh and colleagues obtained repeated CSF samples from
adult males of a closely related species of an Old World primate, they also
found a high degree of intraindividual stability in CSF 5-HIAA concentrations
among adult males.59,60 This high degree of intraindividual stability is not
limited to the laboratory setting, where environmental changes are closely
controlled. When adolescent male macaques that live in a free-ranging forest
were trapped once in 1991 and again in 1992, and a CSF sample was obtained
during each of the two captures, the between-year average correlation in CSF
5-HIAA was r = 0.54.61
Although absolute concentrations of CSF 5-HIAA vary with age and
situations,52,62 interindividual differences stabilize starting early in life, dur-
ing infancy. For example, in one recently completed study, CSF 5-HIAA
concentrations obtained from infant monkeys on day 14 correlated with sam-
ples obtained on days 30, 60, 90, 120, and 150.62 This interindividual stability
is also present across situations and settings. In another study of 6-month-old
infant rhesus monkeys, home-cage baseline CSF 5-HIAA concentrations were
strongly correlated with CSF 5-HIAA concentrations obtained under condi-
tions of social separation, when samples were taken 2 and 4 weeks apart.52,63
Kraemer et al.64 also found traitlike stability in infant rhesus monkey across
repeated sampling during the first year of life, particularly among infants
reared by their mothers. These early interindividual differences that stabilize
in infancy also appear to be stable across longer periods, with the mean of
CSF monoamine samples taken in infancy (6 months of age) predicting mean
concentrations in middle childhood, a year later.52 A recent study of these
same subjects showed this traitlike response of the serotonin system may
endure into adulthood. CSF 5-HIAA samples obtained when the subjects
were 6 months old were positively correlated with CSF 5-HIAA samples
obtained 5 years later, when the subjects were adults.56
11. Impaired CNS Serotonin Functioning and High

Alcohol Consumption
Numerous studies have shown that men with low CSF 5-HIAA concen-
trations exhibit evidence of impaired impulse control such as increased fire
setting65 and increased violent criminal recidivism.66 A large number of ani-
mal studies show excessive or high rates of alcohol consumption among
animals with reduced central serotonin functioning (e.g., see ref. 67 for a
recent review). Among humans, clinical studies show evidence of reduced
central serotonin functioning in subjects at risk for or who exhibit alcohol
abuse and alcoholism (e.g., see ref. 68 for a recent review). For example,
when they are compared to healthy volunteers, young alcoholic men and
women have low CSF 5-HIAA concentrations, even during periods of absti-
nence.69–71 Depressed patients with first-degree alcoholic relatives have sig-
nificantly lower CSF 5-HIAA and 3-methoxy-4-hydroxyphenylgycol (MHPG)
concentrations than depressed patients without alcoholic relatives.72 Like hu-
mans, nonhuman primates with low CSF 5-HIAA concentrations are more
likely to exhibit behaviors characteristic of impaired impulse control such as
spontaneous, long leaps at dangerous heights and repeated jumping into
baited traps where they are captured.61,73
Because they are impaired in controlling their impulses, we postulated
that subjects with low CSF 5-HIAA would exhibit high rates of alcohol con-
sumption. This hypothesis is consistent with one of the postulates of Clonin-
ger's tridimensional model of alcoholism that type II alcoholism is mediated
by central serotonin and norepinephrine deficits.4,74 In what is to our
knowledge the first study of this hypothesis in nonhuman primates, we
found that alcohol consumption was related to reduced CNS serotonergic and
noradrenergic functioning.18,56 High rates of alcohol consumption during the
stressful conditions of a social separation were correlated with low CSF
5-HIAA56 and MHPG18,56 concentrations obtained during the social separa-
tion stress. MHPG obtained during social separation was also negatively cor-
related with alcohol consumption during nonstress conditions.56 To the de-
gree that nonhuman primate findings can be extrapolated to the human con-
dition, our findings suggest that serotonin-deficit-associated excessive alcohol
consumption may be particularly associated with stressful conditions; re-
duced norepinephrine, on the other hand, was correlated with high alcohol
consumption under both stressful and nonstressful conditions. This suggests
that the level of stress should be taken under consideration when obtaining
CSF 5-HIAA to use as a biological marker to predict alcohol consumption.
Additional evidence that serotonin is involved in excessive alcohol con-
sumption comes from investigations that use serotonin-function-enhancing
drugs to treat alcohol consumption. While primarily norepinephrine-affecting
antidepressant treatments have met with little success in treating alcohol
abuse, some recent studies using the antidepressants that are highly selective
for serotonin have shown promise as adjunctive pharmacological treatments
for maintenance of abstinence (see, for example, ref. 68). In a recent, un-
published study using rhesus monkeys, we investigated treatment of exces-
sive alcohol consumption with the serotonin reuptake inhibitor sertraline.
Baseline alcohol consumption patterns were established in adolescent mon-
keys. They were then treated for alcohol consumption with the serotonin
reuptake inhibitor sertraline. Subjects who were modest alcohol consumers,
who typically consumed alcohol at rates that would not produce intoxicating
blood alcohol levels, were unaffected by the treatment. On the other hand,
subjects that consumed alcohol on a daily basis to the point of visible intoxica-
tion (i.e., acquired blood alcohol levels in excess of 100 mg/dl) reduced their
rates of consumption to match the modest alcohol consumers. This reduction
in alcohol consumption occurred only after chronic but not acute sertraline
treatment. The treatment effect was not a result of a loss of appetite, since
food consumption was unaltered by the sertraline treatment.75 It is also of
note that our findings are consistent with a large number of animal studies
showing high rates of alcohol consumption and an increased alcohol prefer-
ence in subjects with reduced CNS serotonin functioning.67
12. Serotonin and Violence

The trait nature of the serotonin dysfunction could be
an advantage for the potential detection of a
propensity for impulsive violent behavior and perhaps
for the prediction of dangerousness.
—Jan Volavka76
A number of studies have suggested that impaired CNS serotonin func-

tioning underlies excessive, unprovoked aggression. For example, many men
with low CSF 5-HIAA concentrations exhibit increased unplanned aggression
and violence.65,77–86 Paralleling studies in humans, recent studies using non-
human primates have demonstrated that low CSF 5-HIAA concentrations are
correlated with increased rates of wounding, unprovoked and unrestrained
CSF Quartile
Figure 2. The percentage of subjects dead or alive 4 years after a CSF sample was
obtained to quantify CSF 5-HIAA concentrations. Subjects are grouped in quartiles
according to CSF 5-HIAA concentrations. Units of 5-HIAA concentrations are in
picomoles per milliliter. The figure shows a monotonic increase in death rates with
lower quartiles of CSF 5-HIAA. None of the subjects from the highest quartile of CSF
5-HIAA died over the course of the study. (From Higley et al.135)
violence, and inappropriate aggression.25,58,61,73,87,88 This relationship be-

tween impaired CNS serotonin functioning and violence appears to be partic-
ularly strong in nonhuman primate subjects with less competent social skills,
characterized by low social ranking.60
The data suggest, however, that low CSF 5-HIAA is not correlated with
overall levels of aggression; rather, it is only spontaneous, impulsive aggres-
sion that tends to escalate to physically damaging conflicts that shows a
negative correlation with CSF 5-HIAA concentrations.60,73,88 An illustration of
this comes from a study that investigated the separate roles that central tes-
tosterone and serotonin functioning play in controlling aggression and impul-
sivity. 61 CSF testosterone and 5-HIAA and aggressive behaviors were mea-
sured twice in the same free-ranging adolescent male nonhuman primates
with the samples taken 1 year apart. CSF free testosterone concentrations
were positively correlated with aggressiveness and with behaviors indicative
of high social dominance but not with behaviors indicative of impulsivity.
CSF 5-HIAA concentrations were negatively correlated with impulsive leaps

and with violent aggression that escalated into assaults, wounds, and pro-
longed chase sequences (escalated aggression), but CSF 5-HIAA was not
correlated with overall rates of aggression. Dimensional analyses showed that
subjects with low CSF 5-HIAA exhibited high rates of most forms of aggres-
sion; nevertheless, rates of aggression were augmented even further if a
subject with low CSF 5-HIAA also had a high CSF testosterone concentration
(see Fig. 2). Measures of social dominance were highest among males with
both high CSF testosterone and high CSF 5-HIAA concentrations.61
13. Reduced CNS Serotonin Functioning

and Antisociallike Behavior
In humans, type II alcoholism is associated with various antisocial traits
and other problems of impaired social functioning, such as reduced social
affiliation and less competent social skills.74 Indeed, some of the most fre-
quent correlates of type II alcoholism are antisocial personality traits.6,89,90 For
example, there is evidence that impaired social functioning beginning early in
life is predictive of excessive alcohol use in adolescence.90 Adolescents who
rate themselves as socially isolated are more likely to abuse alcohol.91 Young
and elderly men with few friends are more likely to abuse alcohol,92 and adult
children of alcoholics rate themselves as more socially isolated and with few
friends.93 Indeed, in one study the best predictor of successful alcoholism
treatment outcome for type II alcoholics was the degree of impaired social
functioning, with less social impairment predicting a better outcome.94
Nonhuman primates are close to ideal subjects to investigate the relation-
ship between CNS serotonin responses and social functioning. They are in-
herently social, living in complex settings with numerous partners and fluid
social interchanges. As in human society, the rules for such interchanges are
orderly, structured by status and emotional and physical needs. Like hu-
mans, individual primates vary in their sociality, with some subjects highly
gregarious and others solitary. Chamove, Eysenck, and Harlow95 identified
sociality as one of three personality traits in rhesus monkeys. Sociality in
monkeys is a trait that is similar to the personality trait extroversion, that
Eysenck’s identified in humans.96 Sociality is an enduring trait in mon-
keys97,98 as well as in humans.99 Several studies among nonhuman primates
have demonstrated that sociality is positively correlated with CNS serotonin
function. For example, across repeated studies of captive vervet monkeys
(Cercopithecus aethiops, a highly social Old World primate species that is closely
related to macaques), Raleigh and colleagues100–103 found that enhancing
serotonin functioning by administering the serotonin precursor tryptophan,
the uptake inhibitor fluoxetine, or the serotonin agonist quipazine increased
positive social behaviors such as approaching and grooming other monkeys.
When investigators reduced serotonin functioning by administering the tryp-
tophan hydroxylase inhibitor para-chlorophenylananine (PCPA) to the mon-

keys, it produced opposite effects. Instead, the monkeys withdrew from and
avoided social proximity and affiliative social interactions.101,103,104
More recently, research with our collaborators demonstrated that natu-
rally occurring reduced serotonin functioning is correlated with reduced so-
ciality. For example, in a sample of free-ranging adolescent male monkeys,
subjects with low CSF 5-HIAA concentrations exhibited reduced levels of four
measures of sociality: time spent grooming other monkeys (a measure of
affection in nonhuman primates); time spent in close proximity to other group
members; time spent in general affiliative social behaviors; and mean number
of companions within a 5-m radius (<5 m).88 Low rates of positive social
interactions are also associated with low CSF 5-HIAA concentrations among
juveniles of both sexes and adult females.58
Several studies have also shown that diminished central serotonin activ-
ity is correlated with diminished social competence. For example, Kruesi and
colleagues105 found that in behaviorally disturbed and obsessive–compulsive
children, independent of psychiatric classification, reduced social competence
was correlated with low CSF 5-HIAA concentrations. Similarly, studies of
both aggressive humans78 and nonhuman primates98,106,107 have shown that
individuals high in social deviancy, or who are rated as low in competent
social behaviors,106 have relatively low CSF 5-HIAA concentrations.
Type II alcoholism in humans is closely linked to antisocial behavior.
Given that nonhuman primates with infrequent and less competent social
interactions often possess low CSF 5-HIAA concentrations, one might predict
that reduced sociality would be related to high rates of alcohol consumption.
These postulates were tested by measuring the rate and complexity of social
behaviors prior to alcohol exposure and correlating them with alcohol
consumption. The complexity of social behaviors used during interchanges
varies positively with the degree of social competence. For example, infants
engage in frequent ventral clinging with caregivers, a behavior seldom seen
between adults. With increasing social complexity these infantlike affiliative
social behaviors change, and by adulthood, when subjects spend time in close
proximity to each other, they groom and huddle together105,108 but seldom if
ever cling ventrally to each other. We predicted that ostracized subjects, as
indexed by minimal time in social affiliation, would consume more alcohol
than subjects spending a lot of time in social affiliation. Moreover, immature,
less competent social behaviors would correlate positively with high rates of
alcohol consumption. Consistent with our predictions, infrequent social inter-
actions were predictive of high alcohol consumption; and when the subjects
that consumed high volumes of alcohol interacted socially, they used less
mature social behaviors such as infantlike ventral clinging, a behavior that is
characteristic of infants and seldom seen between adults.56
One measure of social competence among nonhuman primates is social
dominance ranking. Social dominance among nonhuman primates is gener-
ally measured by observing who has primary access to prized or limited
resources. As male monkeys’ tenure in their new troops is prolonged, compe-

tence in forming and maintaining social relationships increases social domi-
nance. Because size and weight play little role in acquiring social domi-
nance,109,110 it is clear that physical prowess alone is not sufficient to gain
success in social competition. Indeed, rarely is the highest-ranking male of a
troop the most aggressive.110,111 Instead, social dominance is acquired and
maintained through the formation of affiliative bonds with other troop mem-
bers who then support the dominant male in hostile and challenging social
encounters.110–113 Building coalitions and maintaining social support is cru-
cial to acquire and maintain a high social dominance rank.114–118 One of
the most replicated findings among nonhuman primates with low CSF
5-HIAA56,58,60,102,105 or pharmacologically lowered CNS serotonin function-
ing60,102,118,119 is low social dominance ranking.
14. Low CSF 5-HIAA as a Risk Factor for Social Ostracism

and Early Mortality
Nonhuman primates that exhibit excessive aggressiveness in a naturalis-
tic environment are typically ostracized and often forcibly expelled from their
social group.110,120 Given these findings, it is not surprising that there is a
strong association between the length of a male’s tenure in a troop and the
number of relationships the male is able to form and maintain, especially with
the adult females of the troop, and the social dominance ranking he attains.121
These findings appear to have important consequences for the adolescent
male macaques migrating between social groups.
One of the most critical life-history events among feral-living male ma-
caques is emigration from the natal group. This is a universal event for male
macaques that the species utilize to prevent inbreeding. Migration to a new
troop is a dangerous and highly stressful event for young males. Wounding
and death is not uncommon.122–129 With the exception of human encroach-
ment, for a number of different primates the most frequent cause of adoles-
cent and young adult male death is aggression by other monkeys or intra-
species resources competition.130–133 While some males migrate early in
adolescence, when their growth acceleration is just beginning,88 other sub-
jects are able to delay migration until they are fully mature and more socially
sophisticated.134 An additional benefit of delaying migration is that older
brothers and other familiar conspecifics may have migrated to nearly troops,
facilitating an immigrant’s integration into a new social group.121,135 While
males with high CSF 5-HIAA concentrations remain in their troops of origin
until they are older than this normative age of migration,134 males with low
CSF 5-HIAA typically migrate from their social group of origin when they are
younger and often not fully mature.88
Based on the above observations, we hypothesized that subjects with low
CSF 5-HIAA concentrations would show disproportionally high rates of mor-
tality during the period that most males migrate from their social groups of
origin. When CSF 5-HIAA samples were obtained from 49 free-ranging,

2-year-old prepubertal male rhesus monkeys, low CSF 5-HIAA concentra-
tions were predictive of which subjects died, with 46% of the subjects with
low CSF 5-HIAA concentrations dead or presumed dead 4 years later.136
Indeed, 91% of the dead subjects came from the two lowest quartiles of CSF
5-HIAA concentrations. Direct observations of aggressive behavior showed
that subjects that were dead or missing initiated escalated aggression, a mea-
sure of unrestrained aggression that has a high probability of trauma or inju-
ry, at a higher rate than subjects that were known to be alive.136 The cause of
death could be ascertained for 6 of the 11 dead subjects. The four subjects that
were known to die as a consequence of violence all came from the lowest
quartile of CSF 5-HIAA concentrations and had been rated as more aggressive
during their initial capture.136 None of the subjects from the highest CSF
5-HIAA concentration quartile were dead or missing.
15. CSF 5-HIAA, Frontal Cortex Serotonin Functioning,

and Impulsive Behaviors
While it is clear that CNS serotonin has crucial effects on impulse control,
aggression, overall social functioning, and ultimately on the rate of early
mortality, it is less clear which areas of the brain are involved in these behav-
iors and what the measurement of CSF 5-HIAA concentrations represents. In
primates but not in rodents, the dorsal raphe nucleus is the origin of the
major ascending serotonin pathway from the brain stem (the major source of
CNS serotonin) to the frontal and prefrontal cortices.137 Moreover, the dorsal
prefrontal cortex may be the only cortical area that projects directly back to the
dorsal raphe nucleus.138 Stanley et al.139 found a high positive correlation
between frontal cortex tissue and lumbar CSF 5-HIAA concentrations post-
mortem in humans. Other human studies have also found strong positive
correlations between frontal cortex 5-HIAA and cisternal CSF 5-HIAA concen-
trations.140,141 In a PET study on nonhuman primates, designed to measure
the relationship between CSF 5-HIAA concentrations and regional glucose
utilization, the strongest correlation between CSF 5-HIAA concentrations and
glucose utilization was with the orbitofrontal cortex region.142 Moreover, rat-
ings for aggressivity were correlated with both CSF 5-HIAA concentrations
and glucose utilization in the orbitofrontal cortex.142 Two recent exhaustive
reviews of serotonin receptor functioning in suicide victims concluded that
most studies show low serotonin-2 receptor density in frontal cortex of sui-
cide victims.143,144 Furthermore, individuals with frontal brain injury exhibit
deficits in impulse control,145,146 increased episodes of violence,147 and low
CSF 5-HIAA concentrations.148 These and other systematic studies of nonhu-
man primates and humans with frontal lobe injury show the importance of
the frontal and prefrontal cortices for the regulation of impulses and mainte-
nance of goal-oriented behavior.149,150 Taken together, these studies suggest
that cisternal CSF 5-HIAA concentrations correlate with frontal cortex func-
tioning and the possible importance of frontal cortex serotonin in impulse

control.
16. Etiology of Low CSF 5-HIAA

16.1. Genetic Influences
The nonhuman primate model has made important contributions toward
understanding genetic contributions to serotonin functioning. Despite the po-
tential importance for understanding mechanisms underlying differences in
behavior and mental disorders, only one study of humans utilizing a small
number of monozygotic and dizygotic twins has investigated genetic contri-
butions to CSF monoamine metabolite concentrations. In that study, CSF
5-HIAA concentrations were not found to be heritable. 151 Environmental and
genetic contributions to CSF monoamine metabolite concentrations were in-
vestigated in a number of nonhuman primate infants with different rearing
backgrounds. To study genetic contributions to CSF monoamine concentra-
tions, 55 young rhesus monkeys were reared apart from their 10 fathers to
perform a paternal half-sibling analysis. To study maternal genetic contribu-
tions, 23 infants were reared with their mothers, 23 infants were removed
from their mothers at birth and fostered to unrelated lactating females, and 24
infants were removed from their mothers at birth and reared in peer-only
groups. When the monkeys reached 6 months of age, CSF was obtained via cis-
ternal puncture prior to and during a series of social separations. When the re-
sults were statistically pooled according to the biological father, CSF 5-HIAA
showed significant heritable effects (h2 > 0.5) for both sons and daughters. In
addition, there were substantial maternal genetic influences on the young off-
springs’ 5-HIAA (h2 > 0.5). Although they did not study maternal contribu-
tions, the finding of a paternal genetic contribution to CSF 5-HIAA concentra-
tions has recently been replicated in a study by Clarke and colleagues.152
These findings suggest that a significant portion of the variance in the turn-
over rate of CNS serotonin is determined by genetic mechanisms. While
somewhat speculative, these findings may suggest the mode of genetic trans-
mission of type II alcoholism. It predicts that low serotonin functioning is ge-
netically transmitted, which leads to deficits of impulse control and ultimately
to increased risk for excessive alcohol consumption once alcohol consumption
begins. This pattern, however, does not fit Cloninger’s predictions and find-
ings replicating his original research, since those studies have shown that
type II is transmitted from fathers to sons153 and our studies show that low
serotonin functioning is genetically transmitted by both mothers and fathers.
16.2. Environmental Influences

Primate societies are explicitly structured to assure that the infant ac-
quires and practices its social skills in a relatively safe, protected environ-
ment. Among most Old World monkey societies, neonate monkeys initially
develop their social skills within the watchful tutelage of their biological
mother. Mothers are especially important social agents through which infant
and juvenile monkeys develop the capacity to properly inhibit and express
emotions, including aggression.49,108,154–156 Infants and young monkeys de-
prived of opportunities to interact with their mothers are likely as adolescents
and adults to show diminished affiliative social behaviors necessary for main-
taining social bonds and relationships,157 and in initial interactions with peers
show less frequent and less skilled aggression to maintain social domi-
nance.158 Later in development, as young monkeys’ motor and cognitive
capacities mature, peers become central in developing and practicing social
skills. Through social play, interactions with peers have a crucial role for
acquiring knowledge regarding the proper settings and intensity for exhibit-
ing aggression. Monkeys deprived of adult role models and opportunities to
practice social behaviors with peers are likely to express aggression at inap-
propriate targets or settings and to demonstrate deviant social responses and
social relationships.159–162
Adults not only affect the acquisition and development of observed
behavior, they play a crucial role in the organization and proper development
of the CNS. For example, a number of studies using nonhuman primates
have shown that prior experiences affect serotonin functioning during infan-
cy and childhood.52,63,64 These studies have shown that adult influences,
particularly maternal input, is critical to govern the development of the CNS
serotonin system. In the absence of adult influence, serotonin functioning is
impaired. For example, one rearing condition that has been widely studied in
monkeys is peer-rearing. These subjects are removed from their mothers at
birth and reared with other age-matched infants. When CSF 5-HIAA was
obtained from neonatal peer- and mother-reared monkeys on days 14 and 30,
60, 90, 120, and 150, parentally neglected peer-reared subjects exhibited lower
CSF 5-HIAA than mother-reared subjects.62 One study with a limited sample
size suggested that the effect of early rearing experiences on CSF 5-HIAA may
disappear by adolescence.51 In a study with a larger sample size, in which
peer- and mother-reared subjects were longitudinally studied from infancy
into adulthood, peer-reared subjects exhibited lower CSF 5-HIAA concentra-
tions than mother-reared subjects both in infancy and adulthood.56
Behaviorally, peer-reared subjects exhibit a number of deficits. Even in
the absence of threatening stimuli, juvenile-aged peer-reared monkeys are
highly fearful, and in the face of a prolonged stressor such as social separation
they are more likely to exhibit behaviors characteristic of despair.108 There is
evidence that these behavioral differences persist into early adulthood. In a
recent study, we investigated 22 young peer- and mother-reared adult rhesus
monkeys. The peer-reared monkeys were more likely than the mother-reared
subjects to show regressive, infantlike behaviors such as self-orality and self-
clasping. In addition, they had higher blood plasma concentrations of ad-
renocorticotropin and cortisol.18
Paralleling the findings of peer-rearing in nonhuman primates, prospec-

tive and retrospective studies show that a rearing history of parental neglect is
very frequent during early development among aggressive children diag-
nosed as having conduct disorder163,164 and among delinquent adolescents
and adults with personality disorders.90,165–169 It is also common among chil-
dren who have poor relations with their peers and who exhibit excessive
aggression.90,170,171 Other studies show that children neglected by their par-
ents are likely to use and abuse tobacco, alcohol, and other drugs.90,172–174
Peer-rearing in nonhuman primates has been described as a model of parental
neglect.105,116 As we noted earlier, the peer-reared monkeys were also more
likely to abuse alcohol than the mother-reared monkeys.18
The behaviors of peer-reared subjects parallel the behaviors of normally
reared subjects with low CSF 5-HIAA concentrations. They exhibit inap-
propriate and excessive aggression.105,116 For example, as adolescents, they
were placed in a room with an infant and required to “baby-sit.” All of the
mother-reared monkeys demonstrated high levels of caregiving to the infant.
The peer-reared monkeys, on the other hand, avoided contact with the infant
and were more likely to threaten and show aggression to the infant.105 Ag-
gression to an infant is a behavior that is virtually never seen in normally
reared subjects. This inappropriate aggression extended to other settings as
well. While severe aggression hardly ever occurs in rhesus monkeys that have
lived together for prolonged periods, peer-reared subjects were more likely to
exhibit severe aggression to their cage-mates with whom they had lived for
most of their lives.105
There are indications that these differences in aggressiveness between
peer- and mother-reared monkeys persist beyond adolescence. In a colony-
wide assessment of inappropriate maternal behavior, adult female monkeys
who were peer-reared were more likely than adult female mother-reared
monkeys to neglect and abuse their own infants.175 In addition, we have
recently acquired new evidence regarding long-term effects of early peer rear-
ing. Although in our colony adult mother-reared monkeys outnumber peer-
reared monkeys by 3 to 1, over a 2-year period, of 12 adult males producing
injuries to a female requiring veterinary treatment, 10 were peer-reared (sta-
tistically significant at P < 0.05 using a Chi-square test). Of the seven females
who had to be removed from stable social groups because of fight wounds or
self-aggression requiring veterinary treatment, six were peer-reared (statis-
tically significant at P < 0.05 using a Chi-square test).105 This finding was
recently replicated with a new cohort.176
Like the subjects with naturally occurring low CSF 5-HIAA concentra-
tions, peer-reared subjects are socially inept. They are more likely as juveniles
and young adolescents to achieve low social dominance ranking.176 They are
less likely to exhibit adultlike social huddling and more likely to engage in in-
fantlike ventral clinging, even as adults.61,136 One of the most important find-
ings from these studies is that early parental absence has long-term effects on
central serotonin and perhaps norepinephrine functions, which in turn are
associated with excessive alcohol consumption, impaired social functioning,
and disruptive social behaviors. The effects of rearing on alcohol consump-

tion and neurophysiological functioning parallel the interindividual negative
correlation between CSF 5-HIAA concentrations and alcohol consumption,
with the high-alcohol-consuming peer-reared subjects18 exhibiting low CSF
5-HIAA concentrations. This effect of early experience on type II-like behav-
iors and alcohol abuse is not consistent with Cloninger’s predictions of no
effect of developmental environment on type II alcoholism. It is consistent
with recent studies showing that early experiences shape and account for
antisocial behavior among adolescents and young adults, however.177 It is
also noteworthy that the peer-reared subjects exhibit high levels of fear and
anxietylike behaviors, which is predicted by type I alcoholism. This suggests
that, at least in this monkey model, the psychobiology of the two types of
alcoholism may overlap in some subjects.
17. Summary and Conclusions

Within the general rhesus monkey population, a phenotype for high
rates of alcohol consumption exists. Subjects with this phenotype chronically
consume alcohol at rates producing intoxication. Our findings show that
underlying etiological mechanisms and biobehavioral correlates parallel many
of the predictions of Cloninger’s neurogenetic model of alcoholism. Behavior
patterns and biological indices that characterize high anxiety, whether consti-
tutionally or stress induced, were correlated with high rates of alcohol con-
sumption, consistent with predictions of type I alcoholism. Reduced CNS
serotonin functioning early in life, as measured by low CSF 5-HIAA concen-
tration, is a risk factor for low serotonin functioning later in life. Subjects with
low CSF 5-HIAA concentrations have impaired impulse control, resulting in
frequent violence, infrequent and less competent social behaviors, and low
social status. They are shunned as companions and forced to leave their social
groups at an early age and are likely to suffer from early mortality, often as an
adjunct consequence to violent behavior. CSF 5-HIAA concentrations are
genetically influenced, and parental neglect during early childhood may con-
tribute to reduced central serotonin functioning during adolescence and early
adulthood. Reduced central serotonin and norepinephrine functioning, de-
creased social affiliation, and less competent social functioning are risk factors
that may contribute to excessive alcohol consumption in adulthood. Early
rearing experiences that reduce serotonin functioning appear to exaggerate an
inherited predisposition for alcohol consumption. These findings suggest the
potential utility of this nonhuman primate model for understanding the neu-
robiology of controlled alcohol consumption and competent social function-
ing on one hand and excessive alcohol consumption, disruptive social behav-
iors, and excessive aggression on the other hand.
ACKNOWLEDGMENTS. The authors would like to thank Alan Dodson, Mar-

ibeth Champoux, Mariken Hasert, Sue Higley, Ted King, Karen Lucas, Pat-
rick Mehlman, Stephen Suomi, and Kristin Zajicek, who assisted in various
aspects of the studies.
References
1. Cloninger CR: Etiologic factors in substance abuse: An adoption study perspective. NlDA
Res Monogr 89:52–72, 1988.
2. Gilligan SB, Reich T, Cloninger CR: Etiologic heterogeneity in alcoholism. Genet Epidemiol
4:395–414, 1987.
3. Nurnberger JI, Gershon ES: Genetics of affective disorders, in Post RM, Ballenger JC (eds):
Neurobiology of Mood Disorders. Baltimore, Williams & Wilkins, 1984, pp 76–101.
5. Cloninger CR: Anxiety and theories of emotion, in Noyes R Jr, Roth M, Burrows GD (eds):
Handbook of Anxiety, vol. 2: Classification, Etiological Factors and Associated Disturbances. New
York, Elsevier Science Publishers, 1988, pp 1-29.
6. Cloninger CR: A unified biosocial theory of personality and its role in the development of
anxiety states. Psychiatr Dev 3:167–226, 1986.
7. Linnoila M, Virkkunen M, George T, et al: Serotonin, violent behavior and alcohol. Experien-
tia 71:155–163, 1994.
8. McKinney WT: Models of Mental Disorders: A New Comparative Psychiatry. New York, Plenum
Medical, 1988.
9. McClearn GE: Animal models in alcohol research. Alcohol Clin Exp Res 12:537–576, 1988.
10. Samson HH, Li T-K: Models of ethanol self-administration: Introduction. Alcohol Clin Exp
Res 12:571–572, 1988.
11. Cicero TJ: Animal models of alcoholism? in Eriksson K, Sinclair JD, Kiianmaa K (eds): Animal
Models in Alcohol Research. New York, Academic Press, 1980, pp 99–117.
12. Crabbe JC, Belknap JK, Buck KJ: Genetic animal models of alcohol and drug abuse. Science
264:1715–1723, 1994.
13. Li T-K, Lumeng L, Doolittle DP: Selective breeding for alcohol preference and associated
responses. Behav Genet 23:163–170, 1993.
14. Rogers AE, Fox JG, Whitney K, et al: Acute and chronic effects of ethanol in nonhuman
primates, in Hayes KC (ed): Primates in Nutritional Research. New York, Academic Press,
1979, pp 249–289.
15. Winger G: Animal models for understanding alcohol as a reinforcer, in Cox WM (ed): Why
People Drink. New York, Gardner Press, 1990, pp 9-36.
16. Crowley TJ, Weisbard C, Hydinger-MacDonald MJ: Progress toward initiating and main-
taining high-dose alcohol drinking in monkey social groups. J Stud Alcohol 44:569–590, 1983.
17. Crowley TJ, Andrews AE: Alcoholic-like drinking in simian social groups. Psychopharmacol-
ogy 92:196–205, 1987.
18. Higley JD, Hasert MF, Suomi SJ, Linnoila M: Nonhuman primate model of alcohol abuse:
Effects of early experience, personality, and stress on alcohol consumption. Proc Nat Acad Sci
USA 88:7261–7265, 1991.
19. Kraemer GW, McKinney WT: Social separation increases alcohol consumption in rhesus
monkeys. Psychopharmacology 86:182–189, 1985.
20. Kraemer GW, Lake CR, Ebert MH, McKinney WT: Effect of alcohol on cerebrospinal fluid
norepinephrine in rhesus monkeys. Psychopharmacology 85:444–448, 1985.
21. Alcoholism NIAAA: Diagnosis and Assessment of Alcohol Use Disorders. Seventh Special Report
to the US Congress on Alcohol and Health. 1990, pp 181-208. Rockville, MD, US Depart-
ment of Health and Human Services.
22. Association AP: American Psychiatric Association. Washington, DC, American Psychiatric
Association, 1987.
23. American Psychiatric Association: Diagnostic and Statistical Manual of Mental Disorders, 4th
ed. Washington, DC, Author, 1994.
24. Cottler LB: Comparing DSM-III-R and ICD-10 substance use disorders. Addiction 88:689–
696, 1993.
25. Higley JD, Suomi SJ, Linnoila M: Progress toward developing a nonhuman primate model
of alcohol abuse and alcoholism. Soc Sci Med in press.
26. Mello NK, Mendelson JH: The effects of drinking to avoid shock on alcohol intake in
primates, in Roach MK, McIsaac WM, Creaven PJ (eds): Biological Aspects of Alcohol. Austin,
University of Texas Press, 1971, pp 313–332.
27. Meisch RA, Henningfield JE, Thompson T: Establishment of ethanol as a reinforcer for
rhesus monkeys via the oral route: Initial results. Adv Exp Med Biol 59:323–342, 1975.
28. Ervin FR, Palmour RM, Young SN, et al: Voluntary consumption of beverage alcohol by
vervet monkeys: Population screening, descriptive behavior and biochemical measures.
Pharmacol Biochem Behav 36:367–373, 1990.
29. Gustafson R: Is the strength and the desirability of alcohol-related expectancies positively
related? A test with an adult Swedish sample. Drug Alcohol Depend 28:145–150, 1991.
30. Gustafson R: The development of alcohol-related expectancies from the age of 12 to the age
of 15 for two Swedish adolescent samples. Alcohol Clin Exp Res 16:700–704, 1992.
31. Gustafson R: Alcohol-related expected effects and the desirability of these effects for Swed-
ish college students measured with the Alcohol Expectancy Questionnaire (AEQ). Alcohol
Alcohol 28:469–475, 1993.
32. Andersson T, Magnusson D: Drinking habits and alcohol abuse among young men: A
prospective longitudinal study. J Stud Alcohol 49:245–252, 1988.
33. Grant BF, Harford TC, Grigson B: Stability of alcohol consumption among youth: A national
longitudinal survey. J Stud Alcohol 49:253–260, 1988.
34. Jones RJ, Bowden DM: Social factors influencing alcohol consumption: A primate model, in
Eriksson K, Sinclair JD, Kiianmaa K (eds): Animal Models in Alcohol Research. New York,
Academic Press, 1980, pp 185–189.
35. Higley JD, Hopkins WD, Thompson WW, et al: Peers as primary attachment sources in
yearling rhesus monkeys (Macaca mulatta). Dev Psychol 28:1163–1171, 1992.
36. Robins, LN: Lifetime prevalence of specific psychiatric disorders in three sites. Arch Gen
Psychiatry 41:949–58, 1984.
37. Gerald MC: Alcohol and alcoholism, in Gerald MC (ed): Pharmacology: An Introduction to
Drugs. Englewood Cliffs, NJ, Prentice-Hall, 1981, p 221.
38. Cadell TE, Cressman R Group social tension as a determinant of alcohol consumption in
Macaca mulatta, in Goldsmith EI, Moor-Jankowski J (eds): Medical Primatology. New York, S.
Karger, 1972, pp 250–259.
39. Mello NK, Mendelson JH: Factors affecting alcohol consumption in primates. Psychosom Med
28:529–550, 1966.
40. Higley JD: Continuity of Social Separation Behaviors in Rhesus Monkeys from Infancy to Adoles-
cence. Unpublished doctoral dissertation, University of Wisconsin, Madison, 1985.
41. Smith PK: Does play matter? Functional and evolutionary aspects of animal and human
play. Behav Brain Sci 5:139–184, 1982.
42. Suomi SJ: Peers, play and primary prevention in primates, in Kent M, Rolf J (eds): Primary
Prevention of Psychopathology: Social Competence in Children. Hanover, NH, Press of New
England, 1979, pp 127–149.
43. Zimmerman RR, Strobe DA, Steere P, Geist CR: Behavior and malnutrition in the rhesus
monkey, in Rosenblum LA (ed): Primate Behavior. New York, Academic Press, 1975, pp 241–
265.
44. Kraemer GW, Daria HL, Moran EC, McKinney WT: Effects of alcohol on the despair to peer
separation in rhesus monkeys. Psychopharmacology 73:307–310, 1981.
45. Crowley TJ: Substance abuse research in monkey social groups. Prog Clin Biol Res 131:255–
275, 1983.
46. Gray JA: The Psychology of Fear and Stress. New York, Cambridge University Press, 1987.
47. Tucker JA, Vuchinich RE, Sobell MB: Alcohol’s effects on human emotions: A review of the
stimulation/depression hypothesis. Int J Addict 17:155–180, 1982.
48. Chamove AS, Rosenblum LA, Harlow HF: Monkeys (Macaca mulatta) raised with only peers.
A pilot study. Anim Behav 21:316–325, 1973.
49. Harlow HF: Age-mate or peer affectional system. Adv Study Behav 2:333–383, 1969.
50. Champoux M, Coe CL, Schanberg S, et al: Hormonal effects of early rearing conditions in
the infant rhesus monkey. Am J Primatol 19:111–117, 1989.
51. Higley JD, Suomi SJ, Linnoila M: CSF monoamine metabolite concentrations vary according
to age, rearing, and sex, and are influenced by the stressor of social separation in rhesus
monkeys. Psychopharmacology 103:551–556, 1991.
52. Higley JD, Suomi SJ, Linnoila M: A longitudinal assessment of CSF monoamine metabolite
and plasma cortisol concentrations in young rhesus monkeys. Biol Psychiatry 32:127–145,
1992.
53. Goodwin DW: Alcoholism: The Facts. New York, Oxford University Press, 1994.
54. Higley JD, Hasert MF, Dodson A, et al: Diminished central nervous system serotonin func-
tioning as a predictor of excessive alcohol consumption: The role of early experiences. Am J
Primatol 33:214, 1994.
55. King ST, Higley JD, Dodson A, et al: Alcohol consumption in rhesus monkeys: Effects
of genetics, gender, early rearing experience, and social setting. Am J Primatol 30:323,
1993.
56. Higley JD, Suomi SJ, Linnoila M: A nonhuman primate model of Type II excessive alcohol
consumption? (Part 1): Low CSF 5-HIAA concentrations and diminished social competence
correlate with excessive alcohol consumption. Alcohol Clin Exp Res 20:629–642, 1996.
57. Higley JD, Suomi SJ, Linnoila M: A nonhuman primate model of Type II excessive alcohol
consumption? (Part 2): Diminished social competence and excessive aggression correlates
with low CSF 5-HIAA concentrations. Alcohol Clin Exp Res 20:543–650, 1996.
58. Higley JD, King ST, Hasert MF, et al: Stability of interindividual differences in serotonin
function and its relationship to aggressive wounding and competent social behavior in
rhesus macaque females. Neuropsychopharmacology 14:67–76, 1996.
59. Raleigh MJ, Brammer GL, McGuire MT, et al: Individual differences in basal cisternal cere-
brospinal fluid 5-HIAA and HVA in monkeys. The effects of gender, age, physical charac-
teristics, and matrilineal influences. Neuropsychophurmacology 7:295–304, 1992.
60. Raleigh MJ, McGuire MT: Serotonin, aggression, and violence in vervet monkeys, in Mas-
ters RD, McGuire MT (eds): The Neurotransmitter Revolution. Carbondale, Southern Illinois
University Press, 1994, pp 129–145.
61. Higley JD, Mehlman PT, Poland RE, et al: A nonhuman primate model of violence and
assertiveness: CSF 5-HIAA and CSF testosterone correlate with different types of aggressive
behaviors. Biol Psychiatry, in press.
62. Shannon C, Champoux M, Higley JD, et al: Interindividual differences in neonatal serotonin
functioning: Stability of interindividual differences and behavioral correlates. Am J Primatol
36:155, 1995.
63. Higley JD, Thompson WT, Champoux M, et al: Paternal and maternal genetic and environ-
mental contributions to CSF monoamine metabolite concentrations in rhesus monkeys (Mac-
aca mulutta). Arch Gen Psychiatry 50:615–623, 1993.
64. Kraemer GW, Ebert MH, Schmidt DE, McKinney WT: A longitudinal study of the effect of
different social rearing conditions on cerebrospinal fluid norepinephrine and biogenic
amine metabolites in rhesus monkeys. Neuropsychopharmacology 2:175–189, 1989.
65. Virkkunen M, Nuutila A, Goodwin FK, Linnoila M. Cerebrospinal fluid monoamine metab-
olite levels in male arsonists. Arch Gen Psychiatry 44:241–217, 1987.
66. Virkkunen M, De Jong J, Bartko J, et al: Relationship of psychobiological variables to recidiv-
ism in violent offenders and impulsive fire setters. A follow-up study. Arch Gen Psychiatry
46:600–603, 1989.
67. LeMarquand D, Pihl RO, Benkelfat C: Serotonin and alcohol intake, abuse, and depen-
dence: Findings of animal studies. Biol Psychiatry 36:395–421, 1994.
68. LeMarquand D, Pihl RO, Benkelfat C: Serotonin and alcohol intake, abuse, and depen-
dence: Clinical evidence. Biol Psychiatry 36:326–337, 1994.
69. Ballenger JC, Goodwin FK, Major LF, Brown GL: Alcohol and central serotonin metabolism
in man. Arch Gen Psychiatry 36:224–227, 1979.
70. Banki CM: Factors influencing monoamine metabolites and tryptophan in patients with
alcohol dependence. J Neural Transm 50:89–101, 1981.
71. Borg S, Kvande H, Liljeberg P, et al: 5-Hydroxyindoleacetic acid in cerebrospinal fluid in
alcoholic patients under different clinical conditions. Alcohol 2:415–418, 1985.
72. Rosenthal NE, Davenport Y, Cowdry RW, et al: Monoamine metabolites in cerebrospinal
fluid of depressive subgroups. Psychiatry Res 2:113–119, 1980.
73. Mehlman PT, Higley JD, Faucher I, et al: Low CSF 5-HIAA concentrations and severe
aggression and impaired impulse control in nonhuman primates. Am J Psychiatry 151:1485–
1491, 1994.
74. Cloninger CR: A unified biosocial theory of personality and its role in the development of
anxiety states: A reply to commentaries. Psychiatr Dev 6:83–120, 1988.
75. Higley JD, Hasert MF, Dodson A, et al: Treatment of excessive alcohol consumption using
the serotonin reuptake inhibitor sertraline using a nonhuman primate model of alcohol
abuse. Paper presented at Research Society on Alcoholism, San Diego, CA, June 13–18,
1992.
76. Volavka J: Neurobiology of Violence. Washington, DC, American Psychiatric Press, 1995.
77. Brown GL, Goodwin FK, Ballenger JC, et al: Aggression in humans correlates with cere-
brospinal fluid amine metabolites. Psychiatry Res 1:131–139, 1979.
78. Brown GL, Ebert MH, Goyer PF, et al: Aggression, suicide, and serotonin: Relationships to
CSF amine metabolites. Am J Psychiatry 139:741–746, 1982.
79. Lidberg L, Tuck JR, Asberg M, et al: Homicide, suicide and CSF 5-HIAA. Acta Psychiatr Scand
71:230–236, 1985.
80. Limson R, Goldman D, Roy A, et al: Personality and cerebrospinal fluid monoamine metab-
olites in alcoholics and controls. Arch Gen Psychiatry 48:437–441, 1991.
81. Linnoila M, Virkkunen M, Scheinin M, et al: Low cerebrospinal fluid 5-hydroxyindoleacetic
acid concentration differentiates impulsive from nonimpulsive violent behavior. Life Sci
33:2609–2614, 1983.
82. Linnoila M, DeJong J, Virkkunen M: Family history of alcoholism in violent offenders and
impulsive fire setters. Arch Gen Psychiatry 46:613–616, 1989.
83. Linnoila M: Monoamines, glucose metabolism, and impulse control, in van Praag HM,
Plutchik R, Apter A (eds): Violence and Suicidality: Perspectives in Clinical and Psychobiological
Research. New York, Brunner/Mazel, 1990, pp 218–244.
84. Roy A, Adinoff B, Linnoila M: Acting out hostility in normal volunteers: Negative correla-
tion with levels of 5HIAA in cerebrospinal fluid. Psychiatry Res 24:187–194, 1988.
85. Virkkunen M, Rawlings R, Tokola R, et al: CSF biochemistries, glucose metabolism, and
diurnal activity rhythms in alcoholic, violent offenders, fire setters, and healthy volunteers.
86. Virkkunen M, Kallio E, Rawlings R, et al: Personality profiles and state aggressiveness in
Finnish alcoholic, violent offenders, fire setters, and healthy volunteers. Arch Gen Psychiatry
51:28–33, 1994.
87. Higley JD, Mehlman P, Taub D, et al: Cerebrospinal fluid monoamine and adrenal correlates
of aggression in free-ranging rhesus monkeys. Arch Gen Psychiatry 49:436–441, 1992.
88. Mehlman P, Higley JD, Faucher I, et al: Correlation of CSF 5-HIAA concentrations with
sociality and the timing of emigration in free-ranging primates. Am J Psychiatry 152:907–913,
1995.
89. Helzer JE, Pryzbeck TR: The co-occurrence of alcoholism with other psychiatric disorders in
the general population and its impact on treatment. J Stud Alcohol 49:219–224, 1988.
90. Windle M: A longitudinal study of antisocial behaviors in early adolescence as predictors of
late adolescent substance use: Gender and ethnic group differences. J Abnorm Psychol 99:86–
91, 1990.
91. Andersson T, Magnusson D: Patterns of adjustment problems and alcohol abuse in early
childhood: A prospective longitudinal study. Dev Psychopathol 1:119–131, 1989.
92. Windle M, Miller-Tutzauer C, Barnes GM, Welte J: Adolescent perceptions of help-seeking
resources for substance abuse. Child Dev 62:179–89, 1991.
93. Meyers AR, Hingson R, Mucatel M, Goldman E: Social and psychologic correlates of prob-
lem drinking in old age. J Am Geriatr Soc 30:452–456, 1982.
94. Domenico D, Windle M: Intrapersonal and interpersonal functioning among middle-aged
female adult children of alcoholics. J Consult Clin Psychol 61:659–666, 1993.
95. Litt MD, Babor TF, DelBoca FK, et al: Types of alcoholics, II. Application of an empirically
derived typology to treatment matching. Arch Gen Psychiatry 49:609–614, 1992.
96. Chamove AS, Eysenck HJ, Harlow HF: Personality in monkeys: Factor analysis of rhesus
social behavior. Q J Exp Psychol 24:496–504, 1972.
97. Eysenck SB, Eysenck HJ: A comparative study of criminals and matched controls on three
dimensions of personality. Br J Soc Clin Psychol 10:362–366, 1971.
98. McGuire MT, Raleigh MJ, Pollack DB: Personality factors in vervet monkeys: The effects of
sex, age, social status, and group composition. Am J Primatol 33:1–13, 1994.
99. Stevenson-Hinde J, Simpson MJA: Subjective assessment of rhesus monkeys over four
successive years. Primates 21:66–82, 1980.
100. Buss AH, Plomin R: Temperament: Early Developing Personality Traits. Hillsdale, NJ, Erlbaum,
1984.
101. Raleigh MJ, Brammer GL, Yuwiler A, et al: Serotonergic influences on the social behavior of
vervet monkeys (Cercopithecus aethiops sabaeus). Exp Neurol 68:322–334, 1980.
102. Raleigh MJ, Brammer GL, McGuire MT: Male dominance, serotonergic systems, and the
behavioral and physiological effects of drugs in vervet monkeys (Cercopithecus aethiops
sabaeus), in Miczek KA (ed): Ethopharmacology: Primate Models of Neuropsychiatric Disorders.
New York, Alan R. Liss, 1983, pp 185–197.
103. Raleigh MJ, Brammer GL, McGuire MT, Yuwiler A: Dominant social status facilitates the
behavioral effects of serotonergic agonists. Brain Res 348:274–282, 1985.
104. Raleigh MJ, McGuire MT: Social influences on endocrine function in male vervet monkeys,
in Ziegler TE, Bercovitch FB (eds): Socioendocrinology of Primate Reproduction. New York,
Wiley-Liss, 1990, pp 95–111.
105. Higley JD, Linnoila M, Suomi SJ: Ethological contributions: Experiential and genetic contri-
butions to the expression and inhibition of aggression in primates, in Hersen M, Ammer-
man RT, Sisson L (eds): Handbook of Aggressive and Destructive Behavior in Psychiatric Patients.
New York, Plenum Press, 1994, pp 17-32.
106. Kruesi MJ, Rapoport JL, Hamburger S, et al: Cerebrospinal fluid monoamine metabolites,
aggression, and impulsivity in disruptive behavior disorders of children and adolescents.
107. Raleigh MJ, McGuire MT, Brammer GL: Subjective assessment of behavioral style: Links to
overt behavior and physiology in vervet monkeys. Am J Primatol 18:161–162, 1989.
108. Higley JD, Suomi SJ: Temperamental reactivity in non-human primates, in Kohnstamm GA,
Bates JE, Rothbart MK (eds): Temperament in Childhood. New York, John Wiley & Sons, 1989,
pp 153–167.
109. McGuire MT, Raleigh MJ: Serotonin-behavior interactions in vervet monkeys. Psychophar-
macol Bull 21:458–463, 1985.
110. Smuts BB: Gender, aggression and influence, in Smuts BB, Cheney DL, Seyfarth RM, et al
(eds): Primate Societies. Chicago, University of Chicago Press, 1987, pp 400–412.
111. Raleigh MJ, Steklis HD: Effect of orbitofrontal and temporal neocortical lesion on the affilia-
tive behavior of vervet monkeys (Cercopithecus aethiops sabaeus). Exp Neurol 73:378–379, 1981.
112. Packer C, Pusey AE: Female aggression and male membership in troops of Japanese ma-
caques and olive baboons. Folia Primatol 31:212–218, 1979.
113. Walters JR, Seyfarth RM: Conflict and cooperation, in Smuts BB, Cheney DL, Seyfarth RM,
et al (eds): Primate Societies. Chicago, University of Chicago Press, 1987, pp 306–317.
114. Chapais B: Why do male and female rhesus monkeys affiliate during the birth season? in
Rawlins RG, Kessler M (eds): The Cayo Santiago Macaques. Chicago, SUNY Press, 1986,
pp 173–200.
115. Chapais B: Rank maintenance in female Japanese macaques: Experimental evidence for
social dependency. Behaviour 102:41–59, 1988.
116. Higley JD, Suomi SJ: Reactivity and social competence affect individual differences to severe
stress in children: Investigations using nonhuman primates, in Pfeffer CR (ed): Intense Stress
and Mental Disturbance in Children. Washington, DC, American Psychiatric Press, 1996, pp 1–69.
117. Raleigh MJ, McGuire MT: Animal analogues of ostracism: Biological mechanisms and social
consequences. Ethol Sociobiol 7:53–66, 1986.
118. Raleigh MJ, McGuire MT, Brammer GL: Serotonergic mechanisms promote dominance
acquisition in adult male vervet monkeys. Brain Res 559:181–190, 1991.
119. Raleigh MJ, Brammer GL, Ritvo ER: Effects of chronic fenfluramine on blood serotonin,
cerebrospinal fluid metabolites, and behavior in monkeys. Psychopharmacology 90:503–508,
1986.
120. Smuts BB: Sex and Friendship in Baboons. New York, Aldine, 1985.
121. Drickamer LC, Vessey SH: Group changing in free-ranging male rhesus monkeys. Primates
14:249–254, 1973.
122. Altmann J: Baboon Mothers and Infants. Cambridge, MA, Harvard University Press, 1980.
123. Dittus WPJ: The social regulation of population density and age-sex distribution in the toque
monkey. Behaviour 63:281–322, 1977.
124. Gartlan JS: Adaptive aspects of social structure in Erythrocebus patas, in Kondo S, Kawai M,
Ehara A, Kawamura S (eds): Proceedings from the Symposia of the Fifth Congress of the Interna-
tional Primatological Society. Tokyo, Japan Science Press, 1975, pp 161–171.
125. Henzi SP, Lucas JW: Observations on the inter-troop movement of adult vervet monkeys
(Cercopithecus aethiops). Folia Primatol 33:220–235, 1980.
126. Ohsawa H: Takeover of a harem and the subsequent promiscuity in patas monkeys, in
Ehara A, Kimura T, Takenaka O, Iwamoto M (eds): Primatology Today. New York, Elsevier,
1991, pp 221–224.
127. Packer C: Inter-troop transfer and inbreeding avoidance in Papio anubis. Anim Behav 27:1–36,
1979.
128. Pusey AE, Packer C: Dispersal and philopatry, in Smuts BB, Cheney DL, Seyfarth RM, et al
(eds): Primate Societies. Chicago, University of Chicago Press, 1987, pp 250–266.
129. van Noordwijk MA, van Schaik CP: Male migration and rank acquisition in wild long-tailed
macaques (Macaca fascicularis). Anim Behav 33:849–861, 1985.
130. Brain C: Deaths in a desert baboon troop. Int J Primatol 13:593–599, 1992.
131. Dittus WP: The socioecological basis for the conservation of the toque monkey (Macaca
sinica) of Sri Lanka (Ceylon), in Rainier III PoM, Bourne GH (eds): Primate Conservation. New
York, Academic Press, 1977, pp 238–265.
132. Dittus WPJ: The social regulation of primate populations, in Lindburg D (ed): The Macaques:
Studies in Ecology, Behavior and Evolution. New York, Van Nostrand Reinhold, 1980, pp 131–
157.
133. Southwick CH, Siddiqi MF: Population dynamics of rhesus monkeys in Northern India, in
Bourne GH, Rainier III PoM (eds): Primate Conservation. New York, Academic Press, 1977,
pp 339–362.
134. Kaplan JR, Fontenot MB, Berard J, et al: Delayed dispersal and elevated monoamine activity
in free-ranging rhesus monkeys. Am J Primatol 35:229–234, 1995.
135. Vessey SH, Meikle DB, Drickamer LC: Demographic and descriptive studies at La Parguera,
Puerto Rico, P R Health Sci J 8:121–127, 1989.
136. Higley JD, Mehlman PT, Taub DT, et al: Excessive mortality in young male nonhuman
primates with low CSF 5-HIAA concentrations. Arch Gen Psychiatry 53:537–543, 1996.
137. Jacobs BL, Azmitia EC: Structure and function of the brain serotonin system. Physiol Rev
72:165–229, 1992.
138. Arnsten AF, Goldman-Rakic PS: Selective prefrontal cortical projections to the region of the
locus coeruleus and raphe nuclei in the rhesus monkey. Brain Res 306:9–18, 1984.
139. Stanley M, Traskman-Bendz L, Dorovini-Zis K: Correlations between aminergic metabolites
simultaneously obtained from human CSF and brain. Life Sci 37:1279–1286, 1985.
140. Knott P, Haroutunian V, Bierer L, et al: Correlations post-mortem between ventricular CSF
and cortical tissue concentrations of MHPG, 5-HIAA, and HVA in Alzheimer’s disease. Biol
Psychiatry 25:112A, 1989.
141. Wester P, Bergstrom U, Eriksson A, et al: Ventricular cerebrospinal fluid monoamine trans-
mitter and metabolite concentrations reflect human brain neurochemistry in autopsy cases. J
Neurochem 54:1148–1156, 1990.
142. Doudet D, Hommer D, Higley JD, et al: Cerebral glucose metabolism, CSF 5-HIAA, and
aggressive behavior in rhesus monkeys. Am J Psychiatry 152:1782–1787, 1995.
143. Coccaro EF: Impulsive aggression and central serotonergic system function in humans: An
example of a dimensional brain–behavior relationship. Int Clin Psychopharmacol 7:3–12, 1992.
144. Molcho A, Stanley B, Stanley M: Biological studies and markers in suicide and attempted
suicide. Int Clin Psychopharmacol 6:77–92, 1991.
145. Jarvie HF: Frontal lobe wounds causing disinhibition: A study of six cases. J Neurol Neuro-
surg Psychiatry 17:14–32, 1954.
146. Miller LA: Impulsivity, risk-taking, and the ability to synthesize fragmented information
after frontal lobectomy. Neuropsychologia 30:69–79, 1992.
147. Heinrichs RW: Frontal cerebral lesions and violent incidents in chronic neuropsychiatric
patients. Biol Psychiatry 25:174–178, 1989.
148. van Woerkom TC, Teelken AW, Minderhous JM: Difference in neurotransmitter metabo-
lism in frontotemporal-lobe contusion and diffuse cerebral contusion. Lancet vol. 1, 812–813,
1977.
149. Goldman-Rakic PS: Motor control function of the prefrontal cortex. Ciba Found Symp
132:187–200, 1987.
150. Miller SE: Serotonin, carbohydrates, and atypical depression. Pharmacol Toxicol 1:61–71,
1992.
151. Oxenstierna G, Edman G, Iselius L, Oreland L, Ross SB, Sedvall G: Concentrations of
monoamine metabolites in the cerebrospinal fluid of twins and unrelated individuals—A
genetic study. J Psychiat Res 20:19–29, 1986.
152. Clarke AS: Long-term effects of prenatal stress on HPA axis activity in juvenile rhesus
monkeys. Biol Psychiatry. 1995.
153. Cloninger CR, Bohman M, Sigvardsson S, von Knorring AL: Psychopathology in adopted-
out children of alcoholics. The Stockholm Adoption Study. Recent Dev Alcohol 3:37–51, 1985.
154. Bernstein IS, Ehardt CL: Modification of aggression through socialization and the special
case of adult and adolescent male rhesus monkeys (Macaca mulatta). Am J Primatol 10:213–
227, 1986.
155. Harlow HF, Harlow MK: The affectional systems, in Schrier AM, Harlow HF, Stollinitz F
(eds): Behavior of Nonhuman Primates. New York, Academic Press, 1965, pp 287–334.
156. Higley JD, Suomi SJ: Parental behavior in non-human primates, in Sluckin W (ed): Parental
Behavior in Animals and Humans. Oxford, UK, Blackwell Press, 1986, pp 152–207.
157. Coelho AM, Bramblett CA: Early rearing experiences and the performance of affinitive and
approach behavior in infant and juvenile baboons. Primates 25:218–224, 1984.
158. Coelho AMJ, Bramblett CA: Effects of rearing on aggression and subordination in Papio
monkeys. Am J Primatol 1:401–412, 1981.
159. Capitanio JP: Behavioral pathology, in Mitchell G, Erwin J (eds): Comparative Primate Biology:
Behavior, Conservation and Ecology. New York, Alan R. Liss, 1986, pp 411–454.
160. Mitchell G: Abnormal behavior in primates, in Rosenblum RA (ed): Primate Behavior: Develop-
ments in Field and Laboratory Research. New York, Academic Press, 1970, pp 195–249.
161. Suomi SJ: Abnormal behavior and primate models of psychopathology, in Fobes JL, King JE
(eds): Primate Behavior. New York, Academic Press, 1982, pp 171–215.
162. Suomi SJ: The development of social competence by rhesus monkeys. Ann 1st Super Sanita
18:193–202, 1982.
163. Dishion TJ, Loeber R: Adolescent marijuana and alcohol use: The role of parents and peers
revisited. Am J Drug Alcohol Abuse 11:11–25, 1985.
164. Offord DR, Boyle MC, Racine YA: The epidemiology of antisocial behavior in childhood and
adolescence, in Pepler DJ, Rubin KH (eds): The Development and Treatment of Childhood Aggres-
sion. Hillsdale, New Jersey, Lawrence Erlbaum Associates, Publishers; 1991, pp 31–54.
165. Forgatch MS: The clinical science vortex: A developing theory of antisocial behavior, in
Pepler DJ, Rubin KH (eds): The Development and Treatment of Childhood Aggression. Hillsdale,
NJ, Lawrence Erlbaum, 1991, pp 291–316.
166. Mulder RT, Joyce PR, Cloninger CR: Temperament and early environment influence comor-
bidity and personality disorders in major depression. Compr Psychiatry 35:225–233, 1994.
167. Patterson GR, Stouthamer-Loeber M: The correlation of family management practices and
delinquency. Child Dev 55:1299–1307, 1984.
168. Patterson GR, DeBaryshe BD, Ramsey E: A developmental perspective on antisocial behav-
ior. Am Psychol 44:329–335, 1989.
169. Patterson GR, Capaldi D, Bank L: An early starter model for predicting delinquency, in
Pepler DJ, Rubin KH (eds): The Development and Treatment of Childhood Aggression. Hillsdale,
NJ, Lawrence Erlbaum, 1991, pp 139–168.
170. Eron LD, Huesmann LR, Zelli A: The role of parental variables in the learning of aggression,
in Pepler DJ, Rubin KH (eds): The Development and Treatment of Childhood Aggression. Hills-
dale, NJ, Lawrence Erlbaum, 1991, pp 169–188.
171. Farrington DP: Offending from 10 to 25 years, in Von Desen KT, Mednick SA (eds): Prospec-
tive studies of crime and delinquency. Boston, Kluver-Nijhoff, 1983, pp 17–37.
172. Steinberg L, Lamborn SD, Darling N, Mounts NS, Dornbusch SM: Over-time changes in
adjustment and competence among adolescents from authoritative, authoritarian, indul-
gent, and neglectful families. Child Dev 65:754–770, 1994.
173. Werner EE: Resilient offspring of alcoholics: A longitudinal study from birth to age 18. J Stud
Alcohol 47:34–40, 1986.
174. Windle M: The difficult temperament in adolescence: Associations with substance use,
family support, and problem behaviors. J Clin Psychol 47:310–315, 1991.
175. Suomi SJ, Ripp C: A history of mother-less mother monkey mothering at the University of
Wisconsin Primate Laboratory, in Reite M, Caine N (eds): Child Abuse: The Nonhuman Primate
Data. New York, Alan R. Liss, 1983, pp 50–78.
176. Lucas K, Krusinger J, Suomi SJ, Linnoila M: Serotonin functioning, early rearing, and severe
aggression are correlated with acquisition of social dominance rank. Am J Primatol 36:140,
1995.
177. Cadoret RJ, Yates WR, Troughton E, Woodworth G, Stewart MA: Genetic-environmental
interaction in the genesis of aggressivity and conduct disorders. Arch Gen Psychiatry 52:919–
924, 1995.
III
Psychology
Alfonso Paredes, Section Editor

Overview
Alfonso Paredes
In spite of popular assumptions about the close association between alcohol

use and human violence, support for a causal relationship is lacking. Epide-
miological and correlational data usually offered to document the association
does not validate causal explanations. It is therefore necessary to review
research specifically designed to elucidate the mechanisms responsible for
this relationship. The chapters included in this section present three research
perspectives. In Chapter 10, Brad J. Bushman describes laboratory experi-
ments conducted with humans, including the methods of alcohol administra-
tion and the measurements used to assess aggressive behavior. The chapter
illustrates how various assumptions have been operationalized for testing.
The merit of explanations of alcohol-related aggression is examined including
the role attributed to physiological disinhibition and personal expectancies. In
this regard, an effect has been attributed to the person’s belief of having
drunk alcohol, contrasting it with the physiological effects of the substance.
Indirect cognitive, emotional, and physiological factors also have been sug-
gested. The author applied meta-analytic procedures to draw conclusions
from the literature. In his opinion, alcohol “causes” aggression, but such
effect cannot be attributed solely to pharmacological mechanisms. Expectancy
effects must be considered. Furthermore, alcohol indirectly produces changes
within the person that increases the probability of expressing this behavior.
Bushman concludes that experimental manipulations that facilitate aggres-
sion, such as provocations, frustrations, and aggressive cues, have a stronger
effect on intoxicated subjects than on their sober counterparts.
In Chapter 11, Mark W. Lipsey and associates provide a broad review of
the literature. The findings from animal models are briefly examined, remind-
Alfonso Paredes • Laboratory for the Study of Addictions and UCLA Drug Abuse Research
Center, West Los Angeles Veterans Administration, Los Angeles, California 90073.
223
224 III • Psychology
ing us at the same time of the necessity of using human subjects in the
experiments. The authors contend that it may not be justified to draw homol-
ogies between “neural circuits” and the physiological activities that mediate
aggressive behavior in animals and those responsible in humans. These stud-
ies, however, may help to identify the neurochemical systems in the brain that
mediate the effects of alcohol. Lipsey et al. emphasize the need of applying
experimental approaches to address the possibility that observed differences in
the expression of violence are related to factors other than alcohol consump-
tion. Violent behavior is a relatively rare event; after all, most individuals who
drink do not become violent. We are therefore dealing with low probabilities
difficult to detect and compare. Nevertheless, it is important to identify the
causal mechanisms by which alcohol consumption might influence these prob-
abilities. There are many processes operating that may account for the effects of
alcohol on violence. These include cognitive impairment, disinhibition of vio-
lent impulses, or the expectancy effects mentioned earlier. Interactions leading
to violence involve alcohol as well as personality and situational variables.
Research with human subjects within laboratory conditions is necessary in
spite of methodological challenges and the contrived nature of laboratory
drinking situations. Within such context it is possible to manipulate alcohol
consumption as well as other relevant variables. Consumption is difficult to
manipulate under naturalistic conditions. Human studies should help to deter-
mine if aggressive behavior can be caused by manipulating consumption under
defined circumstances. If research specifically designed to reveal causal rela-
tionships does not demonstrate their existence, the plausibility of a causal
associations under natural conditions would be less likely.
Lipsey et al. devote considerable attention to the methodological prob-
lems common in studies of alcohol-violence interactions. There are sampling
problems when groups that characteristically exhibit violent behavior are in-
vestigated, as is the case with criminal system populations. Under these
circumstances, comparisons with subjects with similar characteristics and pat-
terns of alcohol consumption, but who do not exhibit violent behavior, cannot
be done. The authors note that many investigations do not control for con-
founding variables like sociodemographic characteristics, personality disor-
der, other drug use, and early exposure to violence. Greater attention to these
variables therefore is encouraged.
A better conceptualization of the issues involved and operational defini-
tions of the key variables are necessary. In spite of the methodological defi-
ciencies of the studies in the literature, the authors note that the relevant
research does not yield consistently or predominantly negative or null results.
The alcohol–violence assumptions ventured are generally consistent with a
causal role of the neurophysiological action of alcohol. Factors beyond the
pharmacological effects of the substance, however, play a role.
Maria Elena Khalsa-Denison and associates, in Chapter 12, begin by
drawing a parallel between the views from the literature regarding the asso-
ciation between alcohol and violence and those between cocaine and vio-
lence. This brief review is followed by a study that examines characteristics of
III • Overview 225
the violent behavior exhibited by cocaine-dependent patients. The authors

point out that a significant proportion of cocaine-dependent individuals are
also users of alcohol. Intake of alcohol in a large proportion of these cases is
severe enough to justify a diagnosis of alcohol dependence. Alcohol is con-
sumed by these individuals concurrently or alternatively. It is therefore im-
portant to consider the consequences of the combined intake, contrasting it
with the use of only one drug. Applying a natural history methodology the
cocaine use career of cocaine-dependent individuals was parceled out in seg-
ments of time defined by the use of the drug in combination or not with
alcohol. The social behavior exhibited during these periods, including vio-
lence, was also recorded. According to the findings of the study, violent
behavior was more frequent during periods of alcohol and/or cocaine use. On
the other hand, this behavior was less likely to be manifested during absti-
nent periods. Alcohol intake periods were more likely to be accompanied by
violent behavior whether or not cocaine was used concurrently. Violent be-
havior was not normative for this group; the subset of patients who exhibited
this type of behavior was also more likely to have manifested this behavior
prior to the onset of their addiction. This would suggest that individual fac-
tors other than the pharmacological effects of the drug were at play. Given the
limitations of a study such as this, the conclusions are tentative and highlight
a need for further research on this issue.
A greater effort comparing the mechanisms underlying violent behavior
in users of alcohol and other drugs is needed. Such investigations may reveal
interesting contrasts. In the future it may be possible to construct a matrix
with the various drugs along one dimension and a sequence of neurobiologi-
cal and psychosocial factors along another dimension. Within this matrix, the
relative contribution of these factors could be presented. In the case of some
drugs, pharmacological factors would play a predominant role, while cogni-
tive, disinhibition, and personality or social–environmental variables may be
of greater salience in others. With certain drugs, social systemic effects, such
as violence resulting from acts aimed to obtain a supply of the drug or control
over the distribution, may be more prominent.1
The nature of the association between alcohol and violence and the
mechanisms involved has not been settled empirically as yet. This is perhaps
reflected in Lipsey's statement indicating that only for some persons and/or
under some circumstances a causal relationship exists between alcohol use
and violent behavior. Precision in the definition of the assumptions regarding
the mechanisms that play a causal role in the genesis and expression of
violent behavior and application of appropriate research designs remain a
major challenge for investigators.
Reference
1. Goldstein J: Cocaine and crime in the United States. Presentation for the United Nations
Interregional Criminological Research Institute International Symposium on Cocaine, March
1991, Rome, Italy.
10
Effects of Alcohol
on Human Aggression
Validity of Proposed Explanations
Brad J. Bushman
Abstract. In the present review, meta-analytic procedures were used to test the validity of three
explanations of alcohol-related aggression: physiological disinhibition, expectancy, and indirect
cause. According to the physiological disinhibition explanation, alcohol increases aggression
directly by anesthetizing the center of the brain that normally inhibits aggressive responding.
According to the expectancy explanation, alcohol increases aggression because people expect it
to. According to the indirect cause explanation, alcohol increases aggression by causing changes
within the person that increase the probability of aggression (e.g., by reducing intellectual func-
tioning). The results from the review were inconsistent with the physiological disinhibition and
expectancy explanations, but were consistent with the indirect cause explanation. Experimental
manipulations that increased aggression (e.g., provocations, frustrations, aggressive cues) had a
stronger effect on intoxicated participants than on sober participants.
1. Introduction
Violent crime is the issue of greatest concern to Americans today.1–3 There is
probably good reason for this concern, because the US violent crime rate has
been increasing over the past several years (see Fig. 1). One violent crime
occurs ever 16 seconds in the United States.4 Although it is not the only factor
that contributes to violent crime, alcohol intoxication does make a significant
contribution. Numerous correlational studies have found a strong relation
Brad J. Bushman • Department of Psychology, Iowa State University, Ames, Iowa 50011-3180.
227
Year
Figure 1. United States violent crime rate per 100,000 inhabitants, 1971–1992. NOTE:
Data from US Department of Justice.27 According to the FBI, violent crimes are of-
fenses of murder, forcible rape, aggravated assault, and robbery. A regression line
(solid line) and 95% confidence limits (dashed lines) were fit to these data. The least-
squares regression line was –9136.53 + 4.73(year), the coefficient of determination
was r 2 = .79, and the coefficient of correlation was r = .89.
between alcohol intoxication and violent crime. These studies generally find
that over 50% of the assailants were intoxicated at the time the violent crimes
were committed.5–11
Unfortunately, it is difficult to draw causal inferences about the relation
between alcohol and aggression from correlational data. Some of the compli-
cations surrounding correlational studies of alcohol-related aggression are
these: The aggressor may misreport alcohol ingestion as an excuse or to avoid
punishment; alcohol consumption may accompany participation in group
events that could lead to violence; alcohol containers (e.g., bottles, beer
glasses) may be used as weapons; alcoholism may force people into a social
stratum where crime is more probable; some alcoholics involve themselves in
crimes to support their habits; alcohol-related bungling of crimes may in-
crease the probability of capture; and alcohol and violent crime may be re-
sponses to an underlying social malaise. 12 The experimental method avoids
these and many other pitfalls because the researcher controls the occurrence
of events and randomly assigns participants to conditions. Consequently, it is
much easier to draw causal inferences about the effects of alcohol on aggres-
10 • Alcohol and Human Aggression 229
sion from experimental data than from correlational data. It also is much
easier to test explanations of alcohol-related aggression with experimental
studies than with correlational studies. The data base for the present review
was therefore limited to experimental studies of alcohol-related human ag-
gression.13,14
2. Prototypical Methods of Administering Alcohol and Measuring

Aggression in Experimental Studies
Although each experimental study of alcohol-related aggression has
unique components, many experimental studies have common features. The
typical laboratory procedures for administering alcohol and measuring ag-
gression are described in the following sections.
2.1. Administration of Alcohol

To evaluate the effects of alcohol on aggression in humans, most re-
searchers use a placebo design, in which all participants are told that they will
receive alcohol. Only half the participants, however, are actually given alco-
hol; the other half are given a placebo. To enhance the credibility of the
placebo drink, experimenters have poured the beverage from “legitimate”
bottles and have placed a small amount of alcohol on the surface of the drink.
Sometimes an additional group is added to the placebo design, in which
participants are told that they will not receive alcohol and they are not given
alcohol. This group serves as a control because participants receive neither
alcohol nor the expectancy of alcohol. The major problem with comparing
participants in the alcohol group with participants in the placebo or control
groups is that the psychological (e. g., expectancy) and pharmacological ef-
fects of alcohol are confounded in these studies.
The balanced placebo design overcomes this confounding problem. 15 In
the balanced placebo design, half of the participants are told that they will
receive alcohol and half are told that they will not receive alcohol. Within each
of these groups, half of the participants are given alcohol and half are not (see
Fig. 2). To enhance the credibility of the antiplacebo drink, experimenters
have poured the beverage from “legitimate” bottles, diluted the alcohol, used
false Breathalyzer readings, and required participants to complete tasks that
distract them from focusing on the interoceptive signs of intoxication. The
pure pharmacological effects of alcohol on aggression can be determined by
comparing the antiplacebo group with the control group. The pure effects of
alcohol-related expectancies on aggression can be determined by comparing
the placebo group with the control group. Because the balanced placebo
design crosses level of alcohol with level of expectancy, the interaction of
these two factors also can be tested. It is difficult, however, to use the bal-
Expect alcohol
Yes No
Yes
Given
alcohol
No
Figure 2. Balanced placebo design.
anced placebo design with large alcohol doses because participants become
suspicious when they notice the physiological effects of alcohol.
2.2. Measurement of Aggression

Buss16 defined aggression as “a response that delivers noxious stimuli to
another organism” (p. 1). Green17 clarified this definition by adding two
elements: (1) the aggressor delivers the noxious stimuli with the intent to
harm the victim, and (2) the aggressor expects that the noxious stimuli will
have their intended effect. Buss further proposed that acts of human aggres-
sion can be classified using combinations of three dichotomous variables:
physical versus verbal, direct versus indirect, and active versus passive. Al-
though there are eight possible combinations of the three dichotomous vari-
ables proposed by Buss, none of the four “passive” types of aggression are
common in experimental studies of alcohol-related aggression. This chapter
therefore focuses on the four “active” types of human aggression. In physical
aggression the noxious stimuli delivered to the victim are pain and injury,
whereas in verbal aggression the noxious stimuli delivered to the victim are
rejection and threat. In direct aggression the aggressor is easily identified by
the victim, whereas in indirect aggression the aggressor is not easily identi-
fied by the victim. There are two ways in which an aggressive act can be
indirect. First, the victim is not present and the noxious stimuli are delivered
via the negative reactions of others. Second, the victim is not injured or
threatened, but his or her belongings are stolen or damaged.
In the “real world,” most extreme acts of aggression are violent crimes,
which the FBI classifies as murder, forcible rape, aggravated assault, and
robbery. According to Buss’s framework, murder, forcible rape, and aggra-
vated assault are examples of direct physical aggression, whereas robbery is
Table I. “Real World” and Laboratory Examples of the Direct and Indirect Types
of Physical and Verbal Aggression
Type of aggression “Real world” examples Laboratory examples
Physical-Direct Assaulting someone with body Using intense shocks to punish a

parts (e.g., limbs, teeth) or confederate whenever he or
weapons (e.g., clubs, knives, she makes an error on a task.
guns). Using shocks to evaluate a con-
federate’s solution to a problem
Delivering intense shocks to a
confederate on a competitive
reaction-time task.
Physical-Indirect Stealing or damaging someone’s Subtracting money from a con-
property federate in an experiment
Setting a booby trap for someone
Hiring an asassin to kill someone
Verbal-Direct Criticizing, derogating, or cursing Making negative verbal state-
someone ments to a confederate
Making obscene gestures to
someone
Threatening someone
Verbal-Indirect Spreading vicious rumors about Negatively evaluating someone
someone on a questionnaire.
an example of indirect physical aggression. Table I gives real world and labo-
ratory examples of each of the four types of active aggression proposed by
Buss (i.e., direct physical, indirect physical, direct verbal, indirect verbal).
Prototypical procedures for measuring each type of active aggression are giv-
en below.
2.2.1. Direct Physical Aggression. The aggression machine paradigm has

been the primary laboratory procedure used to measure direct physical aggres-
sion, although it is used less frequently now than it was in the past.16 In this
procedure, a participant and a confederate are generally told that the study is
concerned with the effects of alcohol on teaching and learning abilities. Using a
rigged lottery, the real participant is selected to be the teacher and the confeder-
ate is selected to be the learner. The participant presents stimulus materials to
the confederate who attempts to master them. Before the learning task begins,
the participant is sometimes angered by the confederate. When the confeder-
ate makes an incorrect response on a trial, the participant is told to punish him
or her by means of electric shock. By using different buttons, the participant
can control the intensity and duration of shock given to the confederate. The
shocks, for example, may range in intensity from “just perceptible” (e.g.,
button 1) to “excruciatingly painful” (e.g., button 10). In some experiments,
shock duration is controlled by holding down the shock button for the desired
duration. The dependent variables are the intensity and the duration of shock
given to the Confederate. Some researchers have used noxious stimuli other
than electric shocks, such as noise blasts.
The competitive reaction-time paradigm is another method commonly
employed to study the effects of alcohol on direct physical aggression.18 In
this procedure, a participant and a confederate are generally told that the
study is concerned with the effects of alcohol on perceptual motor skills. The
participant competes with the confederate on a reaction-time task in which
the slower-responding person receives electric shock. At the beginning of
each trial, the participant sets the level of shock he or she wants the confeder-
ate to receive if the confederate’s response is slower. At the end of each trial,
the participant is informed of the level of shock the confederate set for him or
her to receive on the trial. The slower-responding person then receives the
indicated intensity of shock. In actuality, the experimenter determines who
wins and loses and the feedback/shocks delivered. Sometimes provocation is
manipulated by increasing the intensity of shock set by the “opponent” across
trials on the reaction-time task. The dependent measure is the intensity of
shock the participant sets for the opponent. Some researchers have used
noise blasts rather than electric shocks as noxious stimuli.
2.2.2. Indirect Physical Aggression. The laboratory paradigms used to mea-

sure direct physical aggression also have been modified to measure indirect
physical aggression. In one study, for example, male college students were
given $2.00 and course credit for their participation.19 Participants were told
to subtract between zero cents (button 0) and nine cents (button 9) from a
confederate whenever he made a mistake on a trial. This paradigm measures
indirect physical aggression because the participant takes the confederate’s
belongings (i.e., his money).
The free-operant paradigm is another method commonly employed to
study the effects of alcohol on indirect physical aggression.20 In this proce-
dure, the participant can press one of two buttons on an apparatus. Pressing
button A results in the accumulation of points exchangeable for money. Press-
ing button B results in the subtraction of points from a fictitious second partic-
ipant. Sometimes provocation is manipulated by subtracting points from the
participant; the point loss is attributed to the fictitious second participant. The
fixed ratios associated with each button also can be manipulated (e.g., a fixed-
ratio of 100 responses might be required for button A, whereas a fixed-ratio of
10 responses might be required for button B).
2.2.3. Direct Verbal Aggression. In the laboratory, direct verbal aggression is

measured by recording a participant’s vocal comments to one or more confed-
erates and counting the frequency of attacks or other negative verbal state-
ments. In one study, for example, male participants were told that they would
be participating in a study of alcohol’s effect on creativity with five other
participants whom they would not know.21 Two of the group members were
confederates. One confederate was a social facilitator who tried to develop

cohesiveness among group members by initiating conversation and telling
jokes. The other confederate tried to antagonize participants by complaining
and insulting their intelligence (e.g., he said that this was the dumbest group
of people he had ever encountered). The experimental sessions were video-
taped and later coded for verbally aggressive statements made by the partici-
pants.
2.2.4. Indirect Verbal Aggression. Indirect measures of verbal aggression are

more common in laboratory experiments than are direct measures of verbal
aggression. Generally, a confederate or experimenter first provokes the partic-
ipant. Rather than confronting the confederate or experimenter face-to-face,
the participant uses a pencil-and-paper measure to evaluate him or her. The
participant is led to believe that negative ratings will harm the confederate or
experimenter in some way. In one study, for example, a male participant was
told to trace a circle as slowly as possible.22 After this task was completed, a
male experimenter burst in the room, introduced himself as the supervisor
who had been observing through a one-way mirror, and contemptuously
stated, “Obviously, you don’t follow instructions. You were supposed to trace
the circle as slowly as possible without stopping but you clearly didn’t do this.
Now I don’t know if we can use your data.” The experimenter paused, then
continued (interrupting the participant if he or she tried to respond), “Do it
over again.” After the experiment, the participant completed an evaluation
form for each member of the lab staff, including the obnoxious experimenter.
The form asked the participant to rate each staff member on 7-point scales as
to whether he or she was effective in performing duties, was a capable em-
ployee, was likeable, made the participant feel comfortable, showed respect
for the participant, and should be rehired. The evaluations were placed in a
sealed envelope and were allegedly sent to the principal investigator to be
used in future hiring decisions.
3. Explanations of Alcohol-Related Aggression
Although several explanations have been proposed to account for alco-

hol-related aggression, most can be placed into one of three categories de-
pending on the role each assigns to alcohol: physiological disinhibition, ex-
pectancy, and indirect cause.
3.1. Physiological Disinhibition

Normally, people have strong inhibitions against behaving aggressively,
because society strongly sanctions such behavior. According to the physi-
ological disinhibition explanation, alcohol increases aggression directly by
anesthetizing the center of the brain that normally inhibits aggressive re-
sponding. Disinhibition theorists argue that alcohol facilitates aggression “not

by ‘stepping on the gas’ but rather by paralyzing the brakes” (p. 40).23
If alcohol directly causes aggression by reducing inhibitions, then partici-
pants in the antiplacebo group should behave more aggressively than partici-
pants in the control group. The antiplacebo versus control comparison pro-
vides the best test of the validity of the physiological disinhibition explanation
of intoxicated aggression because it gives the pure pharmacological effects of
alcohol on aggression (i.e., the effects of alcohol-related expectancies on ag-
gression are removed).
3.2. Expectancy
According to the expectancy explanation, alcohol increases aggression
because people expect it to. Those who behave aggressively while intoxicated
can therefore “blame the bottle” for their actions. According to MacAndrew and
Edgerton,24 violence and other antisocial behaviors occur when alcohol is
consumed because, in many societies, drinking occasions are culturally agreed-
on “time-out” periods when people are not held accountable for their actions.
If alcohol-related expectancies cause aggression, then participants in the
placebo group should behave more aggressively than participants in the con-
trol group. The placebo versus control comparison provides the best test of
the validity of the expectancy explanation of intoxicated aggression because it
gives the pure effects of alcohol-related expectancies on aggression (i.e., the
pharmacological effects of alcohol on aggression are removed).
3.3. Indirect Cause

According to the indirect cause explanation, alcohol increases aggression
by causing certain cognitive, emotional, and physiological changes that in-
crease the probability of aggression. For example, some of the cognitive
changes that accompany alcohol consumption are impaired intellectual func-
tioning, inaccurate assessment of risks, and reduced self-awareness.
If alcohol indirectly causes aggression, then manipulations that increase
aggression in laboratory experiments, such as provocations, frustrations, and
aggressive cues, should have a greater effect on intoxicated participants than
on sober participants. Support for the indirect cause explanation of intoxi-
cated would be provided by a significant Alcohol × Manipulation interaction,
followed up by contrasts that show that the experimental manipulation in-
creased aggression more in participants who were given alcohol than in par-
ticipants who were not given alcohol.
4. Present Review
The primary purpose of the present review was to test the validity of the
physiological disinhibition, expectancy, and indirect cause explanations of
intoxicated aggression. There are two general approaches to reviewing the

literature: the narrative (or qualitative) approach and the meta-analytic (or
quantitative) approach. In the traditional narrative review, the reviewer uses
“mental algebra” to integrate the findings from a collection of studies and
describes the results in a narrative manner. In the meta-analytic review, the
reviewer uses statistical procedures to integrate the findings from a collection
of studies and describes the results using numerical effect size estimates.
Traditional narrative reviews are more likely than meta-analytic reviews to
depend on the subjective judgments, preferences, and biases of the re-
viewer.25 In the present review, meta-analytic procedures were used to test
the validity of the proposed explanations of intoxicated aggression.
5. Method
5.1. Literature Search Procedures

All experimental studies retrieved in previous meta-analytic reviews of
alcohol and aggression by Bushman and his colleagues were included.13,14 To
obtain more recent experimental studies, the PsycLIT computer data base was
searched from January 1992 to March 1995. The terms used to describe aggres-
sion (aggression, agonistic, anger, attack, dominant, fight, hostility, violence)
were the same descriptors used by the International Society for Research on
Aggression in their journal, Aggressive Behavior. The aggression keywords
were paired with three alcohol terms: alcohol, ethanol, and intoxicant. Vari-
ous forms of the keywords also were used (e.g., aggress, aggression, aggres-
sive, aggressor). The search was restricted to studies that used human partici-
pants. The literature review retrieved 60 research reports that included 66
independent samples of participants.*
5.2. Criteria for Relevance

Because the primary purpose of this review was to determine the validity
of causal explanations of intoxicated aggression, two exclusion criteria were
used. First, correlational studies were excluded from the review. Second,
studies that used aggressive state measures were excluded unless they also
used behavioral measures of aggression. An aggressive state is a combination
of thoughts, feelings, behavioral tendencies, and physiological arousal levels
that are elicited by stimuli capable of evoking aggression. Although an ag-
gressive state should heighten the likelihood of aggression, it would not be
classified by most psychologists as aggressive “behavior.”
* Extreme outliers were removed from the data set. In the study by Zeichner and Pihl,26 the noise
intensity standardized mean estimates were 6.64 and 6.81 for the alcohol versus placebo com-
parisons, respectively. In the study by Zeichner and Pihl,27 the noise intensity standardized
mean estimates were 11.63 and 7.00 for the alcohol versus control and alcohol versus placebo
comparisons, respectively.
5.3. Coding Frame

The information listed in the Appendix was extracted from the report of
each study. These data were divided into three categories: participant charac-
teristics, experiment characteristics, and primary study results. (For a detailed
description of the source, participant, and experiment characteristics that
moderate the relation between alcohol intoxication and aggression, see the
meta-analysis by Bushman and Cooper.13,14)
5.4. Meta-Analytic Procedures
The effect size estimate used in this review was the standardized mean
difference, d (see Appendix). The standardized mean difference gives the
number of standard deviation units between the sample means of two groups
(e.g., placebo and control). According to Cohen,28 a “small” d is 0.20, a “medi-
um” d is 0.50, and a “1arge” d is 0.80. The average weighted sample standardized
mean differences, where each standardized mean difference was weighted by
the inverse of its variance, was used to estimate the common population
standardized mean difference, δ. (Average standardized mean differences, 95%
confidence intervals, and moderator tests were calculated using the procedures
described in Hedges and Olkin.29) A 95% confidence interval also was calculated
for δ.29 If studies did not provide enough information to compute an effect size
estimate, but did report the direction or statistical significance of results, vote-
counting procedures were used to obtain an effect size estimate.30 The proce-
dure proposed by Bushman and Wang31 was then used to combine the estimates
based effect size and vote-counting procedures.
One problem that arises in estimating average effect size estimates is
deciding what constitutes an independent hypothesis test. The present review
used a shifting unit of analysis.32 Each statistical test was coded as if it were an
independent event. For example, if a single study contained two measures of
aggression (e.g., intensity of shock given to a confederate and amount of money
subtracted from a confederate), two effect size estimates would be coded. For the
estimate of alcohol’s overall effect on aggression, the two effect size estimates
would be averaged so that the study would contribute only one effect size
estimate. For an analysis in which the effects of alcohol are compared for
different measures of aggression, the study would contribute two effect size
estimates (e.g., direct physical aggression and indirect physical aggression).
Thus, the shifting unit of analysis retains as much data as possible without
violating two greatly the independence assumption that underlies the validity of
meta-analytic procedures.
6. Results
6.1. Sex Differences in Intoxicated Aggression
The results showed that alcohol increased aggression more in men than
in women, x2 (1, k = 65) = 4.64, p < 0.05, where k is the number of indepen-
dent samples of participants.29 The average weighted effect size estimate for
the 59 samples of male participants was 0.50 with 95% confidence interval
[0.41, 0.58]. The average weighted effect size estimate for the six samples of
female participants was 0.13 with 95% confidence interval [–0.20, 0.45]. Be-
cause alcohol increased aggression more in men than in women, subsequent
analyses were based on the results from men only.
6.2. Measurement of Aggression

The type of aggression measure used (i.e., direct physical, indirect physi-
cal, direct verbal, indirect verbal) did not significantly influence the results
reported in this review. Thus, the four types of aggression measures were
pooled for subsequent analyses.
6.3. Validity of Proposed Explanations of Alcohol-Related Aggression

6.3.1. Physiological Disinhibition. If alcohol directly causes aggression by
reducing inhibitions, then the level of aggression should be higher for partici-
pants in the antiplacebo group than for participants in the control group. The
average weighted effect size estimate for the 13 antiplacebo versus control
comparisons was –0.01, with 95% confidence interval [–0.21, 0.19]. Because
the confidence interval contains the value zero, it appears that alcohol does
not directly cause aggression.
6.3.2. Expectancy. If alcohol-related expectancies cause aggression, then the
level of aggression should be higher for participantsin the placebo group than for
participants in the control group. The average weighted effect size estimate for
the 20 placebo versus control comparison was 0.11, with 95% confidence interval
[ –0.06, 0.28]. Because the confidence interval contains the value zero, it appears
that alcohol-related expectancies do not cause aggression.
As can be seen in Fig. 3, the effects of alcohol on aggression cannot be
attributed solely to the pharmacological effects of alcohol nor to the effects of
alcohol-related expectancies. Figure 3 also shows the results from studies in
which the psychological and pharmacological effects of alcohol are con-
founded (i.e., alcohol versus placebo and alcohol versus control comparisons
combined). In the real world, of course, the pharmacological and psychologi-
cal effects of alcohol are confounded. For these confounded studies, the level
of aggression was significantly higher for intoxicated participants than for
sober participants. The average weighted effect size estimate based on 75
comparisons was 0.43 with 95% confidence interval [0.44, 0.62]. As can be
seen in Fig. 3, the pharmacological and psychological effects of alcohol on
aggression are not additive nor are they multiplicative. Thus, alcohol-related
aggression cannot be explained by the independent or joint pharmacological
and psychological effects of alcohol; another explanation is required.
6.3.3. Indirect Cause. If alcohol indirectly causes aggression by producing
internal changes that increase the probability of aggression, then experimen-
Figure 3. Psychological and pharmacological effects of alcohol on human aggression.

NOTE: Capped vertical bars denote 1 standard error. Pharmacology, antiplacebo versus
control comparison. Expectancy, placebo versus control comparison. Expectancy and
pharmacology confounded, alcohol versus placebo and alcohol versus control compar-
isons combined.
tal manipulations that increase aggression should have a greater impact on

intoxicated participants than on sober participants. Eighty-two percent of the
studies included in this review used such manipulations. For this subset of
studies, raters coded whether the Alcohol × Manipulation interaction was
significant or nonsignificant. If the manipulation increased aggression, raters
coded whether the effect was stronger for participants who received alcohol
than for participants who did not receive alcohol.
The results from studies with nonsignificant and significant Alcohol ×
Manipulation interactions are shown in Fig. 4.* For nonsignificant interac-
tions, 36 results were positive (i.e., in the predicted direction) and 14 results
were negative (i.e., in the opposite direction). A sign test showed that the
proportion of positive results was significantly greater than .5, p < 0.05. The
corresponding effect size estimate was g = .22, a value close to Cohen’s28
conventional value for a “large” effect (i.e., g = .25). For significant interac-
tions, 14 results were positive and 2 results were negative. A sign test showed
Alcohol by aggression manipulation interaction

Figure 4. Effects of experimental aggression manipulations on intoxicated and sober
participants. NOTE: Positive effect, aggression-eliciting manipulation was stronger for
participants who received alcohol than for participants who did not receive alcohol.
Negative effect, aggression-eliciting manipulation was stronger for participants who
did not receive alcohol than for participants who received alcohol.
that the proportion of positive results was significantly greater than .5, p <
0.05. The corresponding effect size estimate was g = .38, a value greater than
Cohen’s conventional value for a large effect. These results are entirely consis-
tent with the indirect cause explanation of intoxicated aggression.
7. Conclusions
Does alcohol cause aggression? The results from this review suggest that
it does. In experimental studies, intoxicated participants were more aggres-
* Four studies manipulated variables that decreased aggression (i.e., nonaggressive norms, non-
aggressive cues, pain feedback). The Alcohol × Manipulation interaction was significant for one
of the four studies. All four studies found that aggression reducing manipulations had a weaker
effect on intoxicated participants than on sober participants.
Beverage
Figure 5. United States per capita consumption of selected beverages. NOTE: Data
from US Department of Agriculture.28 Milk includes plain and flavored. Fruit juices
excludes vegetable juices. Alcoholic consumption rates are for the adult population.
sive, on average, than were sober participants. Larger effects might be ob-
tained for higher alcohol doses on human aggression. The effects of alcohol
on aggression are as large as the effects of other independent variables on
aggression (e.g., media violence, anonymity, hot temperatures).33 Alcohol
also influences aggressive behavior as much as it influences other social (e.g.,
risk taking, moral judgment, sexual interest) and nonsocial (e.g., information
processing, self-reported mood, physiological arousal) behaviors.34
Why does alcohol increase aggression? The results from this review sug-
gest that intoxicated aggression cannot be solely attributed to the phar-
macological or expectancy effects of alcohol. Another possibility, however, is
that the null effects for the antiplacebo versus control and placebo versus
control comparisons are due to methodological problems associated with the
antiplacebo and placebo groups. These groups both involve deception. Partic-
ipants in the antiplacebo group do not expect to receive alcohol and might
become suspicious when they taste, smell, and notice the physiological ef-
fects of alcohol. Participants in the placebo group expect to receive alcohol and
might become suspicious when they do not experience the physiological ef-
fects of alcohol. When participants in the antiplacebo and placebo groups
realize that the experimenter has attempted to deceive them concerning the
contents of their beverage, they also might become suspicious about other
facets of the experiment and become more aware of their behavior. Because
aggression is not a socially desirable response, participants in the antiplacebo

and placebo groups might inhibit their behavior consciously.
The results from this review are consistent with the idea that alcohol
indirectly causes aggression by producing changes within the person that
increase the probability of aggression. Experimental manipulations that facili-
tate aggression, such as provocations, frustrations, and aggressive cues, were
shown to have a stronger effect on intoxicated participants than on sober
participants. This may explain why “barroom brawls” are so common. Bars
often are crowded, noisy, smoky, and provocative environments.
The relation between alcohol and aggression would be of little concern if
people rarely drank alcohol. But in the United States, alcohol is the beverage
of choice for many people. As can be seen in Fig. 5, the average American
adult drinks more beer than milk each year. What would happen to violent
crime rates in the United States if people drank less alcohol? A study by Cook
and Moore35 reported that if alcohol consumption per capita decreased by just
10%, there would be a corresponding 1% decrease in murders, 6% decrease in
forcible rapes, 6% decrease in aggravated assaults, and 9% decrease in rob-
beries.* Thus, one way to obtain a kinder, gentler society would be to de-
crease alcohol consumption.
8. Appendix
8.1. Participant Characteristics
1. Sex of participants.
2. Number of participants.
8.2. Experiment Characteristics
1. Type of comparison (i.e., antiplacebo versus control, placebo versus

control, alcohol versus placebo, alcohol versus control).
2. Type of aggression (i.e., direct physical, indirect physical, direct ver-
bal, indirect verbal).
3. If the study contained an experimental manipulation that was expected
to influence aggression, was the Alcohol × Experimental manipulation signif-
icant or nonsignificant at the α = 0.10 level? (Due to low statistical power, the
α = 0.10 level was used rather than the α = 0.05 level) What was the direc-
tion of the contrast analysis (i.e., positive, if the manipulation increased ag-
gression, the effect was stronger for participants who received alcohol than
for participants who did not receive alcohol; negative, if the manipulation
* Cook and Moore35 calculated these estimates using regression analysis from US violent crime
rates from 1979 to 1988.
increased aggression, the effect was stronger for participants who did not
receive alcohol than for participants who received alcohol; null, the effect was
the same for participants who did and did not receive alcohol)?
8.3. Primary Study Results
1. Direction of effect (i.e., positive, negative, null effect).

2. Significance of effect (i.e., significant, nonsignificant at the α = 0.05
level).
3. Magnitude of effect (i.e., standardized mean differences estimate, d).
The standardized mean difference estimate was defined as d = (M 1 – M 2)/SD,
where M 1 and M 2 are the respective sample means for groups 1 and 2, and SD
is the pooled standard deviation. When means and standard deviations were
not reported, but t tests or F tests with 1 degree of freedom in the numerator
were reported, d was calculated using Friedman's formula.29 If F tests with
multiple degrees of freedom in the numerator were reported, means and
standard deviations were requested from authors.
References
1. Adler J: Kids growing up scared. Newsweek 43–50, 1994, January 10.
2. Lacayo R: Lock 'em up! Time 50–53, 1994, February 7.
3. Shannon E: Crime: Safer streets, yet greater fear. Time 63–65, 1995, January 30.
4. Shannon E: Crime: Safer streets, yet greater fear. Time 63–65, 1995, January 30.
5. Beck AJ: Profile of Jail Inmates, 1989. Washington, DC, Bureau of Justice Statistics, 1991.
6. Beck AJ, Kline SA, Greenfield LA: Survey of Youth in Custody, 1987. Washington, DC, Bureau
of Justice Statistics, 1988.
7. Greenberg SW: Alcohol and crime: A methodological critique of the literature, in Collins JJ
nal Behavior. New York, Guilford, 1981.
8. Innes CA: Drug Use and Crime. Washington, DC, US Department of Justice, 1988.
9. MacDonald JM: The Murderer and His Victim. Springfield, IL, Charles C. Thomas, 1961.
10. Murdoch D, Pihl RO, Ross D: Alcohol and crimes of violence: Present issues. Int J Addict
25:1065–1081, 1990.
11. Pernanen K. Alcohol in Human Violence. New York, Guilford Press, 1991.
12. Brain PF: Multidisciplinary examinations of the “causes” of crime: The case of the link be-
tween alcohol and violence. Alcohol Alcohol 21:237–240, 1986.
13. Bushman BJ: Human aggression while under the influence of alcohol and other drugs: An
integrative research review. Curr Direct Psychol Sci 2:148–152, 1993.
15. Ross S, Krugman AD, Lyerly SB, Clyde DJ: Drugs and placebos: A model design. Psychol Rep
10:383–392, 1962.
16. Buss AH: The Psychology of Aggression. New York, Wiley, 1961.
17. Green RG: Human Aggression. Pacific Grove, CA, Brooks/Cole, 1990.
18. Taylor SP: Aggressive behavior and physiological arousal as a function of provocation and the
tendency to inhibit aggression. J Pers 35:297–310, 1967.
19. Barnett RK: The effects of alcohol, expectancy, provocation, and permission to aggress upon
aggressive behavior. Dissertation Abstracts International 40:4993B (University Microfilms No.
ADG80-09229, 0000), 1979.
20. Cherek DR, Spiga R, Egli M: Effects of response requirement and alcohol on human aggres-
sive responding. J Exp Anal Behav 58:577–587, 1992.
21. Murdoch D, Pihl RO: Alcohol and aggression in a group interaction. Addict Behav 10:97–101,
1985.
22. Rohsenow DJ, Bachorowski J: Effect of alcohol and expectancies on verbal aggression in men
and women. J Abnorm Psychol 93:418–432, 1984.
23. Muehlberger CW: Medicolegal aspects of alcohol intoxication. Mich State Bar J 35:38–42, 1956.
24. MacAndrew C, Edgerton RB: Drunken Comportment: A Social Explanation. Chicago, Aldine,
1969.
25. Cooper H, Rosenthal R: Statistical versus traditional procedures for summarizing research
findings. Psychological Bulletin 87:442–449, 1980.
26. Zeichner A, Pihl RO: Effects of alcohol and behavior contingencies on human aggression.
J Abnorm Psychol 88:153–160, 1979.
27. Zeichner A, Pihl RO: Effects of alcohol and instigator intent on human aggression. J Stud
Alcohol 41:265–276, 1980.
28. Cohen J: Statistical Power Analysis for the Behavioral Sciences, 2nd ed. Hillsdale, NJ, Lawrence
Erlbaum, 1988.
29. Hedges LV, Olkin I: Statistical Methods for Meta-analysis. New York, Academic Press, 1985.
30. Bushman BJ: Vote-counting procedures in meta-analysis, in Cooper H, Hedges LV (eds): The
Handbook of Research Synthesis. New York, Russell Sage Foundation, 1994, pp 193–213.
31. Bushman BJ, Wang MC: A procedure for combining sample standardized mean differences
and vote counts to estimate the population standardized mean difference in fixed effects
models. Psychol Methods 1:66–80.
32. Cooper HM: Integrating research: A Guide for Literature Reviews, 2nd ed. Hillsdale, NJ,
Erlbaum, 1989.
33. Anderson CA, Bushman BJ: External validity of “trivial” experiments: The case of laboratory
aggression. General Psychology Review, in press.
34. Steele CM, Southwick L: Alcohol and social behavior I: The psychology of drunken excess.
J Pers Soc Psychol 48:18–34, 1985.
35. Cook PJ, Moore MJ: Violence reduction through restrictions on alcohol availability. Alcohol
Health Res World 17:151–156, 1993.
36. Maguire K, Pastotre AL (eds): Sourcebook of Criminal Justice Statistics, 1993. US Department of
Justice, Bureau of Justice Statistics. Washington, DC, US Government Printing Office, 1994.
37. US Bureau of the Census: Statistical Abstract of the United States: 1994, 114th ed. Washington,
DC, US Government Printing Office, 1994.
38. Friedman H: Magnitude of an experimental effect and a table for its rapid estimation. Psychol
Bull 70:194–197, 1968.
11
Is There a Causal Relationship

between Alcohol Use and Violence?
A Synthesis of Evidence
Mark W. Lipsey, David B. Wilson, Mark A. Cohen,
and James H. Derzon
Abstract. This chapter reviews the evidence bearing on the question of whether those individu-
als who consume alcohol have an increased probability of subsequent violent behavior. Four
bodies of relevant research are examined: experimental studies with animals, experimental stud-
ies with humans, individual-level correlational studies, and macro-level correlational studies. All
these research approaches provide some evidence of an association between alcohol consump-
tion and violent behavior, but no firm conclusion can be drawn about whether alcohol plays a
causal role in such behavior. Various limitations, deficiencies, and ambiguities of available re-
search that contribute to this state of affairs are discussed.
1. Introduction
Ample evidence demonstrates that alcohol consumption frequently accom-
panies incidents of intentional violence.1–3 Roizen,4 for instance, summarized
a large number of studies showing that offenders were variously estimated to
have consumed alcohol prior to 28–86% of homicides, 24–37% of assaults, 7–
Mark W. Lipsey, David B. Wilson, and James H. Derzon • Vanderbilt Institute for Public Policy
Studies, Vanderbilt University, Nashville, Tennessee 37212. Mark A. Cohen • Owen Gradu-
ate School of Management, and Vanderbilt Institute for Public Policy Studies, Vanderbilt Univer-
sity, Nashville, Tennessee 37212.
245
72% of robberies, 13–60% of sex offenses, and 6–57% of incidents of marital

violence. What makes the alcohol–violence relationship especially interest-
ing, of course, is the possibility that alcohol consumption may have a distinct
causal influence on subsequent violent behavior. For example, if a causal
influence of alcohol abuse on violence can be demonstrated, we not only gain
some important information about the adverse effects of alcohol and the
etiology of violence, but also identify treatment for alcohol abuse as a poten-
tially important component of any strategy for preventing or reducing violent
behavior.
As Roizen4 noted, “alcohol’s presence [in violent incidents] is often con-
sidered presumptive of a causal relationship” (p. 6). In the research commu-
nity, however, it is widely recognized that correlation does not establish
causality, and considerable skepticism has been voiced about the state of
current evidence on the causal link. As Roizen4 further remarks, “more is
written about the possible contributions alcohol might make to violent and
criminal behavior than is written from research that attempts to establish
whether there is an empirical relationship and what that relationship might
be” (p. 6). Similar sentiments have been expressed by Fagan,5 Greenberg6
Pernanen,7 Walfish and Blount,8 and others.
It is the purpose of this chapter to review the current state of empirical
evidence bearing on the question of whether consumption of alcohol in-
creases the likelihood that those individuals who consume it will subse-
quently engage in violent behavior toward others. Our focus is on the psycho-
pharmacological effects of alcohol ingestion upon the behavior of drinkers,
including both their directly aggressive behavior and other behavior that
might be closely related to the likelihood of those persons engaging in vio-
lence, e.g., risk-taking. We do not include, however, effects on accidental
violence (e.g., motor vehicle accidents), violence toward self (e.g., suicide), or
violence incidentally associated with alcohol distribution (e.g., violence in
attempts to obtain money to purchase alcohol). While these are interesting
topics, we wish to limit our focus to the important public issue of interperson-
al, especially criminal, violence and ask if alcohol consumption is one of its
causes.
1.1. The Nature of the Causal Question

The concept of causality is an inherently complex one, especially in the
behavioral sciences,9 and no less so in application to the question of the causal
effects of alcohol on violent behavior. To bring some structure to the issue for
present purposes, we differentiate three aspects of the causal question:
1. Is there an overall causal relationship, i.e., a main effect? This is a
question of whether alcohol consumption results in an increased likelihood
generally that the drinkers will subsequently engage in violent behavior.
Demonstration of a general causal effect requires evidence that, all other
things equal, typical persons who consume alcohol have an incrementally
higher probability of engaging in subsequent violent behavior than if they
11 • Alcohol-Violence Relationship 247
had not consumed the alcohol. A variant is to demonstrate that persons who
consume alcohol have a higher probability of engaging in violence than essen-
tially similar other persons in essentially similar circumstances.
It is the phrases “all other things equal” and “essentially similar” in the
above statements that provide the greatest difficulty in establishing credible
evidence on the causal question. Any comparison that involves ambiguity
about how well these conditions are met leaves open the possibility that any
observed differences in violence stem from factors other than alcohol con-
sumption. We will have more to say about this critical and refractory matter
later. It is also pertinent to note that causality in this context has to do with the
probability of behavior, that is, the relative frequency of occurrence in sets of
persons, incidents, observations, and the like. Since violent behavior in most
circumstances is a relatively rare event, it follows that we are necessarily
dealing chiefly with low probabilities that will be correspondingly difficult to
detect and compare.
2. Is there a causal relationship for some persons and/or some circum-
stances, that is, a moderator or interaction effect? This is a question of wheth-
er it can be demonstrated empirically that alcohol consumption results in an
increased likelihood of violent behavior for persons identifiable by some dis-
tinctive characteristics or for persons in certain identifiable situations, or
both.10 Many researchers believe that causal effects come essentially in the
form of an alcohol × person × situation interaction.7,11–13 That is, alcohol
consumption increases the probability of violent behavior only for some per-
sons in some situations. The challenge here is twofold: to identify the charac-
teristics of those persons and/or those situations that represent high risk and,
for those cases, to demonstrate an increased likelihood of violence when
alcohol is consumed (as in no. 1, above).
3. Is there an identifiable causal mechanism by which alcohol consump-
tion might influence the probability of violent behavior? This is a contingent
question that is meaningful only if a causal relationship of some sort can be
demonstrated, but it is nevertheless very important to the causal question. An
empirical demonstration of a causal link between alcohol and violence would
make a limited, albeit fundamental, contribution to our understanding in the
absence of any indication of how such an effect comes about. A considerable
literature of research and theory identifies many possible causal processes
that may account for an effect of alcohol on violence. These include a variety
of physiological or psychological effects that alcohol ingestion might have on
an individual, e. g. , cognitive impairment, disinhibiting violent impulses or
risk-taking, creating expectancies, serving as a cue for counternormative be-
havior, “deviance disavowal,” and so forth.7,11,14
In this chapter, we focus primarily on the first two questions having to do

with evidence that alcohol consumption increases the probability of subse-
quent violent behavior generally or that it increases it differentially for some
persons in some situations. These questions are fundamental to the study of
the alcohol–violence relationship13 and answers continue to be elusive. As
various researchers have observed, the evidence for a significant causal link
between alcohol use and violent behavior is ambiguous, complex, and prob-
lematic. 11,5–17
1.2. The Complexities of Empirically Demonstrating a Causal Link

If there were a simple and direct causal relationship between alcohol and
violence, virtually everyone who drank (at least over some threshold) would
become violent, a proposition refuted by everyday experience. For instance,
in over 600 hr of observation in pubs and bars, Grayham et al.18 observed no
instances of fights or physical injury. We know, therefore, that at most there
is a rather loose causal coupling between alcohol consumption and violence.
It follows that the relationship of interest is difficult to observe empirically
and is most assuredly complex, as the experience of many researchers attests.
The methodological challenges inherent in studying this complex rela-
tionship are formidable. First, we must distinguish two rather different ver-
sions of the causal hypothesis at issue with different implications for study
design. The version that first comes to mind when one thinks of alcohol
causing violence involves incidents as the unit of analysis. In this case, the
focus is on persons who drink and the likelihood that they will engage in
violent behavior shortly thereafter (an “incident”), within the time frame
defined by the presence of alcohol in psychoactive form in their systems.
Studies of this situation typically examine incidents of a selected sort, e.g.,
crime incidents, and attend to whether alcohol was consumed and whether
violence occurred. This line of study is often referred to as investigation of the
“acute” effects of alcohol on violence.1,4,15 The other version focuses on per-
sons as the unit of analysis with attention to their patterns of alcohol use and
violent behavior over time. The linkage of interest here is whether persons
who chronically use (or abuse) alcohol are more likely to engage in violent
behavior, irrespective of alcohol consumption immediately prior to a particu-
lar instance of violence. This line of study investigates the “chronic ” effects of
alcohol on violence. Clearly, the nature of the causal connection posited, the
mechanisms through which it might occur, and the methods by which it can
be studied differ quite considerably for these two lines of inquiry.
Perhaps the most problematic aspect of studying the effects of alcohol on
violence is the inherent trade-off between internal and external validity. Well-
controlled experimental designs in which participants are randomly assigned
to alcohol and no-alcohol conditions and observations are made of subse-
quent aggression unquestionably have the highest internal validity for testing
the causal inference at issue. However, for compelling practical and ethical
reasons, researchers cannot control the alcohol consumption of people in
everyday life and then monitor their violence levels in natural situations.
Instead, laboratory-drinking situations must be employed and, most espe-
cially, proxy or surrogate measures of violence must be contrived. Thus
strong testing of the causal role of alcohol consumption on violent behavior
occurs only in a form that is quite unrepresentative of the violence in society

that is the major concern of policymakers, law enforcement, and many
citizens.
Stronger external validity, or generalizability, to violence in natural situa-
tions comes from correlational or observational studies of such violence. For
instance, survey studies have frequently been used to examine the association
between reported alcohol use and violent behavior for various population
groups. Assuming reasonable validity to those reports (which cannot always
be assumed), such studies have the advantage of dealing with the types of
drinking and violence that are of concern as social problems. What associa-
tions they find between alcohol use and violent behavior, however, are inher-
ently ambiguous with regard to causality. An alcohol–violence correlation
may result from alcohol causing violent behavior, from violent behavior or its
intentions causing alcohol use (e.g., to build courage, to forget), or from other
factors that influence both drinking and violence (e.g., risk-taking person-
ality, disadvantaged economic circumstances). It is also possible for such
correlations to be spurious, that is, to result from coincidence or artifact rather
than from a meaningful causal process. Alcohol abusers, for instance, may be
more inclined to embellish the extent of their violent behavior in research
interviews than nonusers, creating a spurious association between reported
alcohol use and violence. Determining which of these accounts applies is
difficult and almost never conclusive. Study of the causal role of alcohol in
violence thus presents a perverse quandary: Either strong testing of causal
inference or generalizability to the circumstances of principal interest must be
sacrificed in any empirical research design that attempts to address the issue.
A third area of difficulty in alcohol–violence research has to do with the
definition and operationalization of the constructs selected to represent alco-
hol use and violent behavior, the independent and dependent variables, re-
spectively, in such study. If the focus is on alcohol use in a single instance (the
“acute” case), issues of dosage and timing apply. Alcohol has few discernable
physiological effects in very small doses and in sufficiently large doses ren-
ders drinkers comatose and incapable of violence. There is a wide range in
between, however, over which alcohol may have different effects on the
likelihood of violence. Similar concerns apply to the timing between alcohol
consumption and the point of observation for violence. Assuming some de-
cay curve in alcohol’s effects, any given dose of alcohol might have quite
different effects at different intervals subsequent to consumption.
Where the long-term effects of alcohol use are at issue (the “chronic”
case), there is also great variety in the possible drinking patterns that may be
at issue, often distinguished with phrases such as “social drinking,” “abuse,”
“alcoholism,” “binge drinking,” and the like. In neither the chronic nor the
acute case does the research literature show any consistency in the definition
and operationalization of alcohol use nor has there been much systematic
exploration of the relative importance of dose, timing, or drinking pattern as
dimensions of the alcohol use construct and its variants. As a result, the
alcohol construct appears as something of a “fuzzy set” in research studies

with corresponding ambiguity for the researcher attempting to properly oper-
ationalize the construct or the reviewer attempting to interpret the research
results .
A further complication is that alcohol use is almost inevitably a highly
bundled construct; that is, most observations or measurements incorporate
several distinctly different aspects of drinking packaged together in a single
index. At the narrowest level is the pharmacological aspect of alcohol con-
sumption, e.g., as represented in blood alcohol content. This cannot gener-
ally be manipulated and measured in isolation, however. In most practical
measurement circumstances, drinkers are aware of their alcohol consump-
tion, introducing such psychological aspects as expectancy and cultural
meanings into the picture, variables that may have their own effects indepen-
dent of the pharmacological effects of alcohol. In addition, drinking is gener-
ally embedded in a social context involving a mix of circumstances, locations,
companions, and the like that may also separately influence the likelihood of
violence. It is quite difficult to isolate these different aspects of alcohol use in
any operationalization of the construct or even to attain some consistency in
the mix of aspects from one research study to another. Correspondingly,
when relationships between alcohol use and violent behavior are found, it is
generally difficult to know which of the quite different ingredients of the
alcohol use construct package is responsible. While some attention has been
given to this issue, especially in experimental studies of alcohol consumption,
it continues to confound the study of alcohol and violence.
Turning to the dependent variable in the equation, we find similarly
thorny problems associated with the definition and operationalization of the
violence construct in this research area. The commonsense notion that vio-
lence, at least in its clearest and most extreme version, involves intentional
behaviors that physically harm another person is at the core of most defini-
tions, but for very good reasons operationalizations frequently include more
and sometimes less than this core concept. One problem is with the notion of
intent, which invokes aspects of the internal or cognitive states of individuals
that cannot be easily measured. However, without this element in the defini-
tion, even if it cannot be explicitly verified, accidental harm to another person
falls under the violence construct and it takes on a different meaning in
relation to alcohol effects (given the well-documented motor impairment alco-
hol produces). If intent is viewed as a defining feature, on the other hand,
then it may be of less relevance whether actual physical harm was done.
Thus, many definitions and operationalizations include threats or attempts as
well as completed violence. Another area of variation has to do with the target
of the violence. Some operationalizations include violence against property or
violence against self along with violence against another person. Further
variation comes in around the notion of physical harm—some operationaliza-
tions include verbal abuse, arguments, and the like as violence, albeit perhaps
psychological violence.
Moreover, violence is often measured using multi-item or composite

measures designed to increase the reliability of measurement compared to
single items or to better reflect the multifaceted nature of violent behavior.
Frequently, such composite measures encompass a range of antisocial behav-
iors with physical violence generally viewed as an extreme form of such
behavior. Depending on the mix of items or facets included, however, the
portion of such measures related to the intentional physical harm of another
person can be quite small. In sum, then, a broad range of relatively distinct
and disparate behaviors, individually and in aggregates, are encompassed
under one or another operationalization of the violence construct in alcohol-
violence studies. The design, review, and interpretation of such studies, cor-
respondingly, suffers from ambiguity about the construct at issue and what
meaning to attribute to its diverse operationalizations.
With these issues in mind, we turn to a review of the four distinct bodies
of empirical evidence bearing most directly on the question of the causal
influence of alcohol consumption on the likelihood of violent behavior. Two
of the relevant research paradigms chiefly use controlled experimental de-
sign—anima1 studies and laboratory simulations with humans. Two use ob-
servational or correlational methods—surveys of drinking and violent behav-
iors and community level analyses of the covariation between alcohol
availability and violence.
2. Experimental Approaches
2.1. Animal Studies

Studies of the effects of alcohol on animals necessarily have uncertain
generalizability to human behavior. However, they do offer an opportunity to
explore the possibility that there is a very general pharmacological link be-
tween alcohol ingestion and aggressive behavior, at least in mammals whose
neurophysiology may be similar in important respects to humans. In addi-
tion, of course, animal studies offer possibilities of laboratory control and
experimentation with dosage, stimuli, and aggressive contexts that would be
impossible with human subjects. Experimental studies of the effects of alco-
hol on the aggressive behavior of animals have been itemized and thoroughly
reviewed by Miczek et al.3 and Berry and Smoothy,19 and no attempt will be
made here to do more than summarize some of the conclusions that can be
drawn from their efforts.
Perhaps the most important conclusion to be drawn from these reviews is
that examination of the alcohol–violence link in animals does not yield any
simple picture that might guide the understanding of this issue for human
behavior. To begin with, there are many different forms of animal aggres-
sion—conflict in relation to status hierarchies, defense of territory, male rival
fighting, female defense of young, predatory aggression, antipredator de-
fense, and various forms of aggression under such stresses as isolated hous-
ing, crowding, restricted access to food, and noxious stimulation.20 More-

over, there is evidence that many of these forms of aggression are related to
specific brain mechanisms unlikely to respond the same way to alcohol.3,21
Nor do such brain mechanisms bear any necessary resemblance to those in
humans. As Miczek et al.3 put it, “There is no direct evidence . . . that demon-
strates homology between the neural circuitry and physiologic activity that
mediate aggressive behavior in animals and those responsible for human
violence” (p. 381).
Another complication in the animal studies is evidence of significant
individual differences in the aggressive response to alcohol among members
of the same species in similar circumstances. Miczek, Weerts, and DeBold22
reviewed a variety of studies in which subpopulations of rodents and pri-
mates showed reliable increases in aggressive behavior when given alcohol,
while other individuals exhibited no such effect or even a suppression of
aggression. Among primates, these differences in response seem to be related
to social status (see also ref. 20). Some evidence also links these individual
differences in aggressive response to differences in specific neurotransmitter
systems.22,23 Thus, the animal studies reveal important moderator and inter-
action effects involving alcohol, even though the specific variables involved in
those interactions have not yet been fully identified.
As a result of the variation in types of aggression and neurological mech-
anisms, individual differences in aggressive response to alcohol, and other
possible sources of variation (e.g., methodological differences20), animal
studies of alcohol and aggression have not produced consistent results. The
studies itemized by Berry and Smoothy19 and Miczek et al.3 include a notable
portion that show no effects of alcohol on aggression and some show reduced
aggression with the same alcohol doses that are associated with increases in
other studies. Miczek, DeBold, and VanErp23 described the experimental re-
search with animals on the effects of alcohol on aggression as “largely incon-
clusive” (p. 407), and Brain, Miras, and Berry20 characterized it as “somewhat
disappointing” (p. 140).
Nonetheless, there are some instructive results from the animal studies.
Despite the varied animal subjects, forms of aggression, and research para-
digms, a substantial proportion of these studies do show increased aggressive
behavior relative to control animals with administration of alcohol. Moreover,
where alcohol effects appear, they are at lower and moderate doses; higher
doses generally sedate the animals and depress aggression, as in humans.
Thus, there are causal pathways from alcohol ingestion to increased aggres-
sive behavior, at least in some animals, at some doses, in some circumstances,
and for some forms of aggression. Further, animal studies have produced
evidence that the effects of alcohol are mediated by certain neurochemical
systems in the brain.24 Some of these neurochemical systems, in turn, have
been linked to aggressive behavior, though no specific role for these mecha-
nisms has yet been established in mediating alcohol’s effects on such be-
havior.
Also, some studies in mice, rats, and primates have shown effects of
chronic alcohol consumption on aggressive behavior under stress condi-
tions.3,24 Rats administered three daily doses of alcohol, for instance, attacked
intruders more severely than control animals.25 Relatively few animal studies
have investigated chronic alcohol consumption, however, so evidence on this
matter is limited.
Thus, there are experimentally demonstrable effects of both acute and
chronic alcohol ingestion on aggression among some animal subjects. Such
effects, however, do not appear consistently or universally. Individual differ-
ences among the test subjects and variations in the conditions, methods,
doses, forms of aggression, and the like appear to interact with alcohol in a
complex way that sometimes elevates aggression and sometimes does not.
2.2. Human Studies

A considerable number of experimental studies of the effects of acute
alcohol ingestion on human subjects has been conducted with attention to a
range of outcome variables. For purposes of reviewing the results of the
subset of those studies that examined effects on aggression, we rely on a
meta-analysis we26 have recently completed in order to update similar efforts
by others.27–30 For this meta-analysis, we made a thorough search for pub-
lished and unpublished study reports, retrieved those identified as poten-
tially eligible for inclusion, systematically coded the characteristics of each
eligible study and its statistical findings, and analyzed the resulting database.
Since the studies at issue all employed experimental designs, the metric we
used to record their findings was the standardized mean difference on an
aggression measure between the experimental group receiving alcohol and
the control or placebo group not receiving alcohol. This metric, known as an
effect size or Cohen’s d, is conventional in meta-analysis and indexes the
difference between groups in standard deviation units.31 Thus, in this appli-
cation, an effect size of .50 indicates that the group that was administered
alcohol yielded a mean aggression score that was half a standard deviation
higher than the mean for the control group with which it was compared. For
purposes of computing mean effect sizes and other such statistics, each indi-
vidual effect size was weighted by its inverse variance to give larger weight to
those based on larger samples.32
We coded 52 independent studies from 50 reports that met criteria requir-
ing: (1) human subjects, (2) experimental design in which alcohol consump-
tion was the independent variable and some measure described as aggression
or a proxy for aggression was the dependent variable, and (3) the study was
reported in English and contained sufficient information to permit calculation
of appropriate effect sizes for its key results. (A bibliography of the coded
studies is available from the authors.) The aggression measure in eight of
these studies, however, did not involve any physical stimulus or actual im-
plied potential of physical harm (e.g., monetary penalties, harsh evaluation
ratings). In order to keep a focus on physical aggression, we eliminated these

from the pool and retained the 44 studies using some physical stimulus as an
aggressive response. Of these, one used noxious noise and all the others used
electric shock.
With one exception, all of these studies used one of two laboratory para-
digms: the competitive reaction time paradigm (31 studies) or the teacher-
learner paradigm (12 studies). In the competitive reaction time paradigm, the
subject is led to believe that he or she is competing with another subject in a
series of tasks set up to examine the influence of alcohol on reaction time.
When the subject wins a reaction time trial, he or she delivers a shock to the
opponent at an intensity level selected by the subject immediately prior to
the trial. The intensity of these shocks serves as the dependent measure for
the study. In actuality, the opponent does not exist and the wins and losses
are controlled by the experimenter. The degree of threat or provocation is
manipulated in some of these studies by varying the frequency and/or inten-
sity of the shocks received by the subjects on the trials they lose. Depending
on the researcher’s hypothesis, of course, other features of the paradigm may
be varied as well.
In the teacher–learner paradigm, the subject is informed that the experi-
ment is studying the effects of alcohol on teaching and learning abilities. A
rigged drawing determines that the subject will serve as the teacher in a visual
discrimination or paired association memory task. The learner is actually a
confederate of the experimenter (or a computer) who follows a predetermined
pattern of right and wrong answers. Each time the learner makes an error, the
subject (teacher) can choose to correct the learner via an “aggressive” re-
sponse (a shock in 11 studies, a noxious noise in 1) or a “neutral” response (a
red feedback light). In this paradigm the learner cannot retaliate toward the
subject (ref. 33 is the one exception). The frequency, intensity, or duration of
these shocks (or noise) serves as the primary dependent measure for these
studies.
The one study that departed from these two paradigms34 examined the
influence of alcohol, arousal, and aggressive cues on aggressive behavior by
asking subjects in various conditions to assist the experimenter in “stimulat-
ing” a bogus subject. The subject could choose to stimulate the bogus subject
with either a tone or an electric shock. The duration of the shocks served as
the measure of aggression.
2.2.1. External Validity. The major strength of these studies, of course, is

their ability to manipulate alcohol consumption in a randomized trial under
relatively controlled conditions and, hence, directly investigate the causal
influence of the alcohol manipulation on the response variables. Their great-
est weakness is uncertainty regarding the extent to which the results can be
generalized from the laboratory conditions in which they were obtained to
real instances of alcohol consumption and violence in natural situations. From
the standpoint of ecological validity, we would prefer that there be some
straightforward mapping that could be made between the circumstances of

these experimental paradigms and the natural situations to which we would
like to generalize. Clearly, however, there are a number of notable differences
between these laboratory paradigms and the typical circumstances of natu-
rally occurring drinking and violence that make such a mapping problematic.
Intentionally administered electric shock, for instance, is a relatively rare
violent response in natural situations. While some limited evidence suggests
that it may be an acceptable proxy,35 this issue has not been explored carefully
and the results to date are not especially convincing. Also, while interperson-
al violence is generally face-to-face and intimately personal, the “violence” in
the laboratory paradigms is largely depersonalized: shock is administered to
an unseen subject in another room by pushing a button. Additionally, these
paradigms depend on some artificial constraints on respondents’ options. In
the competitive reaction time paradigm, the respondents are required by the
experimental protocol to administer shock, i.e., to be aggressive. The only
choice allowed is with regard to how aggressive they will be. The teacher–
learner paradigm does generally permit a “neutral” or nonaggressive option,
but the demand characteristics of the experimental situation work against its
exclusive use by a subject who does not desire to be aggressive.
Another constraint on the generalizability of the results of these studies
arises from the subject selections typically made by the experimenters and the
alcohol doses administered to those subjects. Of the 44 eligible studies in our
meta-analysis, 34 used college student subjects exclusively and 4 others used
a mix in which college students were a large proportion. In addition, 41 of
these studies used exclusively male subjects. The latter selection may not be
especially damaging with regard to generalization issues since most naturally
occurring violence is committed by males. It is not the case, however, that
college students commit a large proportion of such violence.
Alcohol doses in these studies were typically reported in terms of millili-
ters per kilogram of body weight, but more than half also reported blood
alcohol concentrations (BAC). By using the high correlation between millili-
ters per kilogram and BAC (r = .82), we were able to estimate mean BAC
levels for the subjects in most of the studies. Some subjects attained BAC
levels that would only be considered in the mild impairment range (13 studies
with an estimated BAC from 0.03 to 0.08). Most of the studies (31), however,
used dose levels around the per se drunk level established in many states
(0.08–0.10), though none exceeded 0.10. Thus, the subjects given alcohol
ranged from moderately intoxicated to legally drunk, but none were in the
range of severe intoxication often found in cases of criminal violence.13
2.2.2. Alcohol-Violence Relationship. Despite uncertainty about the ecolog-

ical validity and subject representativeness of these studies in relation to
naturally occurring drinking and violence, they do provide an opportunity to
determine if ostensibly aggressive behavior can be caused by manipulating
alcohol consumption under at least some circumstances. If some form of an
alcohol–violence causal relationship cannot be shown in laboratory situations

specifically designed to reveal it, the plausibility of such a link under more
complex and varied natural conditions is undermined. If such a link can be
shown under laboratory conditions, then more focused inquiry can be made
into the nature of that linkage and the other circumstances in which it might
occur.
The mean effect size found in our meta-analysis for the major compari-
sons between the aggression levels shown by subjects to whom alcohol was
administered and the control subjects who received no alcohol was 0.54. This
overall mean indicates that subjects to whom the alcohol conditions were
applied in this body of studies scored, on average, about half a standard
deviation higher on the aggression measure than the control subjects who
participated in the same task but did not consume any alcohol. If we use the
10-point shock intensity scale from the competitive reaction time paradigm as
a frame of reference, a half standard deviation effect size indicates that sub-
jects in the alcohol conditions, on average, administered shocks that were
about one scale step higher than subjects in the nonalcohol conditions. Our
effect size result is in agreement with earlier meta-analytic work by Bush-
man,27 Bushman and Cooper,28 Hull and Bond,29 and Steele and South-
wick,30 all of whom reported positive effects of the alcohol conditions on
measures of aggression.
On the face of it, this overall mean effect size can be interpreted as
relatively strong evidence for the causal effect of alcohol consumption on at
least one form of aggressive behavior. Researchers in this area have recog-
nized, however, that the situation is more subtle than that. The manipula-
tions used in these experiments have various aspects, as do the reactions of
subjects to those manipulations. Before drawing a firm conclusion about the
causal role of alcohol, it is necessary to seek some assurance that it was
the alcohol part of these manipulations that was primarily responsible for the
effect on aggression and not some other aspect that was bundled with alcohol
in these manipulations. We therefore used the information coded in our
meta-analysis to examine differences among studies that might shed some
light on the potential for misleading confounds or artifacts in their results.
2.2.3. Potential Artifacts. We first compared the mean effect sizes resulting
from the two different research paradigms that dominate these studies. As
Table I shows, the competitive reaction time paradigm produced significantly
larger effect sizes than the teacher-learner paradigm (statistically significant
between-groups Q). The use of shock-based aggression measures and the
alcohol doses administered were similar for the studies in the two major
paradigms, suggesting that some other difference between them may be re-
sponsible for the difference in results. We therefore further differentiated the
procedures in each paradigm to explore any association between them and
the effect sizes found. We did this on an exploratory basis, recognizing that
the Q statistics already showed homogeneous results within each paradigm
Table I. Effect Size Statistics for Alcohol versus Nonalcohol

Conditions by Experimental Paradigma
Weighted
Experimental paradigm mean ES 95% CIb Qwc n
Competitive reaction time 0.61 0.49, 0.74 30.2 29

Teacher–learner 0.41 0.20, 0.62 13.0 12
Other 0.04 –0.47, 0.55 NA 1
Total 0.54 0.43, 0.64 49.8 42
aTwo studies (Zeichner and Pihl57,58) were removed from the analysis due to their
having extreme outliers (effect sizes greater than 6.0).
bCI, Confidence interval (see Hedges and Olkin32).
cQ , Homogeneity within groups (*p < 0.05, reject homogeneity). The between-groups
w
Q indicates that there is heterogeneity between groups (QB(2) = 6.7, p < 0.05).
(Table I). For such small numbers of studies, however, Q has relatively little
statistical power to reject the hypothesis of homogeneity.
A major variation in the competitive reaction time paradigm had to do
with the shocks the subject received. Twenty-two studies used some version
of the standard format in which all subjects received shocks back from the
“opponent.” In three studies, the subjects received no shocks. In the remain-
ing four studies, half the subjects received the standard format, while the
other half received no shocks. As Table II shows, the effects of the alcohol
condition on aggression were much lower when the subjects were not
shocked. Note that the small number of studies requires that this comparison
be interpreted cautiously but, as Gustafson36 suggested, it may be that alco-
hol–violence effects only appear under provocation when nonaggressive op-
tions are unavailable.
The major variations among the smaller number of studies using the
teacher–learner paradigm were as follows: one used a noxious noise for the
aggressive stimulus instead of shock, three involved a provocation (e.g., an
insulting comment or negative evaluation from the learner or a threat of a
retaliatory shock), and three studies did not offer the nonaggressive response
Table II. Effect Size Statistics for the Competitive Reaction Time
Studies by Whether the “Opponent” Shocked the Subjects
Weighted
Provocation conditiona mean ES 95% CIb Qwc n
Subjects shocked 0.72 0.58, 0.86 21.9 26

Subjects not shocked 0.22 –0.03, 0.48 3.2 7
a
Four studies contributed an effect size to both provocation conditions.
CI, Confidence interval (see Hedges & OIkin32).
b
Qw, Homogeneity within groups (*p < 0.05, reject homogeneity). The between-groups,
c
Q indicates that there is heterogeneity between groups (QB(1) = 11.3, p < 0.05).
option (red feedback light). As Table III shows, these variations appear to
produce rather different mean effect sizes, though, with the small numbers of
studies involved, none of these differences attained statistical significance.
There appear to be especially modest effects for conditions involving a provo-
cation and for conditions involving no nonaggressive response option. Note
that in this regard the results from the teacher–learner paradigm show an
apparent reversal of the results from the competitive reaction time paradigm.
Provocation in the competitive reaction time studies (exclusively shock) was
associated with larger effects; provocation (shock threat and verbal) in the
teacher-learner studies was associated with smaller effects. Similarly the
competitive reaction time studies obtained relatively large effects while inher-
ently allowing no nonaggressive response option, while the teacher-learner
studies found their largest effects when a nonaggressive response option was
allowed.
Unfortunately, not enough studies have been conducted for most of the
paradigm variations to give stable results regarding their different effects. It
does appear, however, that the alcohol effect is rather dependent on the
particular procedures used in the research paradigm. Moreover, the para-
digms seem to yield inconsistent results with regard to the effects associated
with provocation and nonaggressive response options. It also appears that
the alcohol effect is rather modest in many of the variations.
One other procedural aspect of these research paradigms was examined.
Table III. Effect Size Statistics for the Teacher–Learner Studies

by Major Procedural Variationa
Weighted
Procedural variation mean ES 95% CIb Qwc n
Dependent variable
Shock intensity 0.40 0.13, 0.67 8.1 7
Number of shocks 0.63 0.28, 0.99 5.5 4
Shock duration 0.67 0.33, 1.02 1.6 4
Shock composite 0.51 0.12, 0.90 3.7 4
Noise composite 0.13 –0.44, 0.69 NA 1
Threat/provocation
No provocation 0.45 0.22, 0.69 14.9 12
Verbal provocation 0.05 –0.52, 0.62 0.003 2
Shock provocation –0.24 –1.12, 0.64 NA 1
Nonaggressive response options
Yes 0.50 0.27, 0.74 9.3 9
No 0.08 –0.36, 0.52 1.0 3
Four studies had an effect size for shock intensity, number, and duration.
a
CI, Confidence interval (see Hedges and Olkin32).

b
Qw, Homogeneity within groups (*p < 0.05, reject homogeneity). The between-groups Q statistic for the
c
dependent variable was not computed, because the inclusion of the same subjects in multiple distributions
violates the assumptions of this statistic. The between-groups Q for provocation indicates that there is homo-
geneity between groups (QB(2) = 3.6, p > 0.05). The between-groups Q for nonaggressive response options
indicates that there is homogeneity between groups (QB(1) = 2.7, p > 0.05).
11 • Alcohol–Violence Relationship 259
Table IV. Effect Size Statistics for Each Paradigm by Whether

Experimenter Blinding Was Used
Weighted
Breakout by paradigm Mean ES 95% CIa QWb n
Competitive reaction time

B1inding –0.30 –1.10, 0.51 NA 1
No blinding 0.64 0.51, 0.76 25.1 28
Teacher-learner
B1inding 0.11 –0.39, 0.60 1.0 2
No blinding 0.47 0.24, 0.70 10.2 10
a
Qw, Homogeneity within groups (*p < 0.05, reject homogeneity). The between-groups
b
Q statistic for the competitive reaction time breakout indicates that there is hetero-
geneity between groups (QB(1) = 5.1, p < 0.05). The between-groups Q statistic for the
teacher–learner breakout indicates that there is homogeneity between groups (QB(1) =
1.8, p > 0.05).
For this, we determined whether the researcher administering the instruc-

tions and collecting the data from the subjects was kept unaware (“blind”) of
the experimental condition in which the subject had been placed. In these
studies experimenter blinding is important to control for subtle clues the
experimenter may give the subject on how the researcher expects the subject
to behave, e.g., unintentionally cuing subjects in the alcohol condition to
behave aggressively.37 Only three of the studies reported that they employed
this methodological control (Table IV). In each research paradigm, the few
studies using experimenter blinding found notably smaller effects than those
studies that did not employ that control. Unfortunately, too few studies ap-
plied experimenter blinding to reach a firm conclusion here, but these results
leave open the possibility that the apparent alcohol effect is little more than
subjects’ response to experimenter demand.
Last, we looked at an issue that has received a fair amount of attention in
this research literature: alcohol expectancy effects. If the subjects in a placebo
control condition believe they have consumed alcohol, then the difference
between the alcohol and placebo condition should reflect the pharmacological
effects of alcohol, rather than subjects’ expectations about the effects of alco-
hol. The balanced placebo design was developed to explore directly the influ-
ence of expectations on aggressive responding. This design is a 2 × 2 factorial
with four conditions: (1) alcohol (expect alcohol, receive alcohol); (2) placebo
(expect alcohol, receive tonic); (3) control (expect tonic, receive tonic); and (4)
antiplacebo (expect tonic, receive alcohol). This design yields a main effect for
alcohol, a main effect for expectancy, and an interaction between alcohol and
expectancy. Of the six studies that used the balanced placebo design, five
were teacher–learner studies and one was a competitive reaction time study.
Table V shows the mean effect sizes for the various effects separately for each
paradigm.
Table V. Effect Size Statistics for the Balanced Placebo Design

by Experimental Paradigm
Weighted
Factor Mean ES 95% CIa Qwb n
Alcohol factor
Competitive reaction time 0.24 –0.33, 0.81 NA 1
Teacher–learner 0.11 –0.12, 0.33 14.0* 5
Expectancy factor
Competitive reaction time 0.61 0.03, 1.18 NA 1
Teacher–learner 0.19 –0.03, 0.42 12.7* 5
Interaction
Competitive reaction time –0.54 –1.12, 0.03 NA 1
Teacher–learner 0.06 –0.17, 0.28 6.4 5
a
Qw, Homogeneity within groups (*p < 0.05, reject homogeneity). A positive interaction
b
effect size indicates that high expectancy increases the effect of alcohol.
As Table V reveals, the careful attempt of the balanced placebo designs to

separate out the effect of actually consuming alcohol from the effect of think-
ing you have consumed alcohol produces some rather surprising results.
First, the main effect of alcohol is notably small in both the research para-
digms. Also, the main effect for alcohol is smaller than the main effect for
expectancy for all the studies. In other words, the largest part of the subjects’
increased aggressiveness in these studies appears to come from their belief
that they have drunk alcohol, not from the actual effects of the alcohol itself.
2.2.4. Individual Differences. A particularly noteworthy feature of the re-
sults of these experimental studies is the great variability reported for subjects
in their aggressive responses.10,12 Individual difference variables have been
relatively neglected in this area of research, however, so little can be said
about the characteristics of those who respond more or less aggressively un-
der the influence of alcohol. Some modest evidence suggests that predisposi-
tion for aggressiveness and related constructs such as unfriendliness, quick-
ness to anger, and trait hostility may differentiate those who respond more
aggressively to alcohol,10 but other plausible factors have not yet been investi-
gated.
2.2.5. Conclusions about Experimental Studies of Humans. The research

paradigms used in experimental studies of the effects of alcohol consumption
on aggression in human subjects are not good simulations of socially impor-
tant forms of drinking and violence. Moreover, the subjects used in those
studies (mostly male college students) and, to a lesser extent, the alcohol
doses administered do not well represent the populations and doses of great-
est practical interest. Nonetheless, this body of research does demonstrate
that there are some laboratory manipulations of alcohol consumption that
exert causal influence on the aggressiveness of responses that subjects believe

are directed at another person. Unfortunately, the role of alcohol per se in
those manipulations remains uncertain. The measured effects of alcohol on
aggressive responses in these studies appear to vary inconsistently according
to the research paradigm and its particular conditions, diminish when con-
trols for experimenter demand are employed, and largely reflect the belief
that alcohol has been consumed rather than actual consumption. It will re-
main difficult to interpret the results of this body of research until further
studies are done to better disentangle the actual effects of alcohol from the
variety of other aspects of subjects’ perceptions and expectations, experimen-
ter demand characteristics, and procedural particulars with which alcohol
consumption is packaged in these studies.
3. Correlational Approaches
The research that most directly investigates the association between alco-
hol use and socially significant forms of violence is necessarily observational
or correlational. Such studies yield information about the covariation of real-
world alcohol use and violent behavior and generally represent one of two
different levels of analysis. Most frequent are studies with individual persons
as the unit of analysis which investigate the relationship between the varying
degrees of alcohol use by those individuals and the extent of their violent
behavior, e.g., a survey of convicted felons about whether they drank prior to
committing their offenses. A second and less common type of study investi-
gates the alcohol–violence relationship at a more macro-level. These macro or
community studies examine aggregate statistics for regions or political juris-
dictions to investigate covariation between alcohol availability and violence
over time or across different communities. Such a study might, for instance,
examine whether communities with higher levels of alcohol consumption
experienced more criminal violence than communities with lower levels.
Correlation, of course, is not necessarily causality, and the major chal-
lenge faced by correlational studies investigating the causal role of alcohol in
violence is to rule out alternate explanations of any alcohol–violence correla-
tion. Since experimental control of the factors related to such alternate expla-
nations is not possible in these studies, statistical controls must be employed
instead. The essence of statistical control is to introduce one or more control
variables into the analysis of the relationship between the alcohol indepen-
dent variable and the violence dependent variable. The results are then exam-
ined for indications that the control variable is sufficient to explain all or part
of the zero-order alcohol–violence relationship.
The difficulty with statistical controls, other than technical issues associ-
ated with using them correctly, is that they can only be employed with spe-
cific, measured variables. Thus, a researcher must be able to identify in ad-
vance every variable that might account for an alcohol–violence correlation in
his or her sample and research circumstances, measure it, and use it in subse-
quent data analysis. The practical impossibility of such a complete specifica-
tion means that, even in the most favorable circumstances, there is some
uncertainty about whether all of the relevant control variables have been
included. And with that comes uncertainty about whether any remaining
alcohol–violence correlation represents a causal relationship or merely the
confounding influence of some unmeasured control variable.
We turn now to an examination of the major bodies of correlational re-
search and consider how they have handled the difficult problem of assessing
the causal influence of alcohol on violence.
3.1. Individual-Level Studies

The typical individual-level correlational study is a survey of a sample of
persons who are questioned with regard to their alcohol consumption (acute
or chronic) and their involvement in violent behavior. The data relevant to the
alcohol–violence relationship yielded by these studies come in the form of
one or more alcohol variables on which alcohol use varies across the individu-
als in the sample and one or more violence variables whose values similarly
vary across the sample. The key correlation in these studies reflects whether
those individuals with higher alcohol use also exhibit higher levels of violent
behavior and, conversely, whether those with lower alcohol use exhibit lower
levels of violence.
The reason we spell out the nature of the correlation at issue in these
studies so specifically is that there is widespread reporting of simple, but
potentially misleading, co-occurrence statistics. Various studies show, for in-
stance, that alcohol was consumed by large proportions of violent offenders
prior to their offense.4 Statistics about the co-occurrence of alcohol use and
violent behavior, however, do not constitute a correlation between alcohol
and violence, since there is no variation on the violence variable—everyone in
the sample is violent. Put another way, there is no base rate information in
these statistics about the level of alcohol use among those comparable persons
who were not violent. Thus, if alcohol is involved in 50% of violent offenses,
we do not know if it is involved in 10, 50, or 90% of nonviolent offenses by
otherwise comparable persons. Without the “other half” of the data required
for a correlation, no strength of association information can be garnered from
co-occurrence statistics.
To summarize the relatively large body of individual-level correlational
studies on alcohol and violence, we draw once again on our own meta-
analytic work (details in ref. 26). Studies were deemed eligible for this meta-
analysis if (1) the report contained quantitative information on the relation-
ship between a measure of alcohol use and a measure of violent or aggressive
behavior by individual persons; (2) the quantitative information was sufficient
to compute a product-moment correlation between alcohol use and violence,
or multivariate results were presented with statistics describing the alcohol-

violence relationship; and (3) the report was in English. A combination of
bibliographic search strategies was applied to locate and retrieve as many
eligible published and unpublished reports as possible. Using these strate-
gies, 870 documents were located and screened, yielding 129 studies that met
the eligibility criteria for the meta-analysis (a bibliography of these studies is
available from the authors).
Over 100 items of information were coded for each eligible study. This
included effect size information for all alcohol–violence relationships report-
ed in a study (coded as correlation coefficients), all control variable informa-
tion, and various particulars of the study methods, measurements, samples,
settings, and so forth. The coded studies were divided into those that dealt
with acute alcohol use (consumption associated with a single incident) or
chronic alcohol use (a pattern of consumption over some time period). A
second distinction was made between studies involving criminal violence and
those involving domestic violence. For our purposes, domestic violence was
any type of interpersonal violence against family members within the home,
such as spouse or child abuse. Criminal violence was any other type of inter-
personal violence potentially chargeable as a crime, whether defined in a legal
context (e.g., murder, assault, rape) or not (e.g., fighting).
Table VI describes the general characteristics of this body of studies,
categorized according to the type of alcohol–violence relationship. The vast
majority were conducted in the United States, most since 1980. Criminal and
general population samples characterized the criminal-chronic studies, while
most of the criminal-acute studies sampled criminal populations and most of
the domestic-chronic studies involved clinical populations (e.g., marital or
family therapy clients, alcohol or drug treatment clients). The type of sam-
pling varied considerably and relatively few studies were based on true prob-
ability samples. Face-to-face interviews were the most common method of
obtaining data; few studies used physical measures, e.g., blood alcohol con-
centration. Note that only two studies reported on the relationship between
acute alcohol use and domestic violence.
Effect sizes were coded as correlation coefficients for all the alcohol-
violence relationships reported in each study; some studies contributed only a
single effect size, while others contributed several. We randomly selected one
of the relevant correlations within each study for each category of relationship
using the most general and differentiated measures of alcohol consumption
and violence and the largest aggregate sample in each study. The characteris-
tics of each respective alcohol and violence measure represented in these
summary effect sizes are reported in Table VII. Alcohol measures chiefly
represented level of alcohol use or distinctions regarding problem drinking
(e. g., alcoholism). This information was most often collected using question-
naires or standardized instruments for self-report by respondents. Most of
the violence measures reflected physical violence. Like the alcohol measures,
Table VI. Study Descriptors for Correlational Studies

by Type of Relationshipa,b
Criminal Domestic
Chronic Acute Chronic Acute
Variable n (%) n (%) n (%) n (%)
Country /region
United States 53 (77) 18 (58) 33 (89) 2 (100)
Scandinavia 1 (1) 2 (7) 1 (3)
United Kingdom 8 (12) 5 (16)
Canada 5 (7) 4 (13)
Other/cannot tell 2 (3) 2 (7) 3 (8)
Publication year
1950–69 3 (4) 2 (7)
1970–74 6 (9) 1 (50)
1975–79 7 (10) 5 (16) 1 (3)
1980–84 19 (28) 7 (23) 8 (22) 1 (50)
1985–89 20 (29) 10 (32) 15 (41)
1990–94 14 (20) 7 (23) 13 (35)
Population sampled
Criminal 29 (42) 24 (77) 2 (5) 2 (100)
Criminal/psychiatric 8 (12) 4 (13) 1 (3)
Clinical/treatment 8 (12) 2 (7) 25 (68)
General population 24 (35) 1 (3) 9 (24)
Sample size
25–49 3 (4) 5 (14)
50–99 10 (15) 2 (7) 9 (24)
100–199 7 (10) 6 (19) 11 (30)
200–499 17 (25) 7 (23) 7 (19) 1 (50)
500–999 14 (20) 6 (19) 2 (5)
1,000–9,999 15 (22) 5 (16) 3 (8) 1 (50)
10,000+ 3 (4) 4 (13)
Cannot tell 1 (3)
Type of sampling
Convenience 17 (25) 5 (16) 24 (65) 1 (50)
Probability 22 (32) 9 (29) 5 (14)
Quasi-probability 5 (7) 2 (7)
Census 3 (4) 3 (10) 1 (3)
Time-sampling 18 (26) 12 (39) 5 (14) 1 (50)
Cannot tell 4 (6) 2 (5)
Survey method
Face-to-face 35 (51) 14 (45) 21 (57) 1 (50)
Self-report 23 (33) 6 (19) 14 (38)
Physical measure 1 (1) 1 (3)
Archival 8 (12) 9 (29) 2 (5) 1 (50)
Cannot tell 2 (3) 1 (3)
Total 69 31 37 2
a Percentages may not sum to 100 due to rounding error.
b Nine studies provided both criminal-chronic and criminal-acute relationships and one
study provided both a domestic-chronic and criminal-chronic relationship. There were
129 studies and 139 relationships.
these were most often questionnaires or standardized instruments for self-

report, but a large proportion came from archives or official records (e.g.,
arrest data).
We turn now to an overview of the results of this meta-analysis bearing
on the nature and magnitude of the relationship between alcohol use and
violence. The first line of evidence is the summary correlations between alco-
hol and violence reported in these studies. We averaged these across studies,
using a sample-size weighting procedure,31,38 to yield the figures in Table
VIII. The largest overall mean correlation was in the domestic-chronic cate-
gory with a weighted mean of 0.22; the criminal-chronic category had a
weighted mean of 0.15, while the criminal-acute category had a mean of 0.10.
(There were too few studies to calculate a mean for the domestic-acute catego-
ry.) The confidence intervals indicated that all of these mean correlations were
statistically significant. Taken together, these findings show that greater alco-
hol use among individuals is clearly associated with higher levels of violent
behavior.
It is useful to appraise the magnitude of the alcohol–violence correlations
in Table VIII using terms more intuitively meaningful than the product-
moment correlation. Assuming a normal distribution on the violence mea-
sure, a transformation analogous to that called U3 by Cohen39 provides one
useful approach. To apply this transformation, we imagine dividing the ag-
gregate sample represented in each mean effect size into two groups: those
with no or low alcohol use and those with moderate to high use. If we
imagine further that 10% of the low-use group demonstrates violent behavior,
we can then ask what proportion of the high-use group must show violent
behavior to yield a relationship of the same magnitude as the mean correla-
tions in Table VIII.
For the largest weighted mean correlation in Table VIII, r = .22 for the
relationship between chronic alcohol use and domestic violence, the corre-
sponding proportion is .20. That is, if 10% of the low-alcohol-use group en-
gaged in domestic violence, then 20% of the high-alcohol-use group would
also. Thus, the correlation of .22 can be understood as representing a 10–
percentage-point increase in the proportion of the sample engaged in vio-
lence, or a doubling of the likelihood of violence. Similarly, the criminal-
chronic mean correlation of .15 is equivalent to a contrast between a
low-alcohol-use group that is 10% violent and a high-use group that is 17%
violent. The mean correlation of .10 for the criminal-acute relationship, in
turn, corresponds to a contrast between 10% and 14% violent.
We see, therefore, that even though the correlations in Table VIII are
numerically modest, the relationships they index are not trivial. What we do
not know is how much of the difference in violence between high- and low-
alcohol users is really due to the causal influence of alcohol and how much is
due to other factors confounded with alcohol use. We must, therefore, find
ways to probe more deeply into the nature of the correlations in Table VIII
before drawing any conclusions about the causal influence of alcohol con-
sumption on violence. Of particular importance in this regard is the evidence
Table VII. Alcohol and Violence Measure Descriptors for Correlational Studies
by Type of Relationshipa
Criminal Domestic
Chronic Acute Chronic Acute
Measure descriptor n (%) n (%) n (%) n (%)

Alcohol Measure
Construct (Chronic)
Binge use 2 (3) 1 (3)
Problem use 8 (12) 7 (19)
Alcoholism 25 (36) 18 (49)
Level of use 25 (36) 9 (24)
Role impairment 2 (3)
Other 4 (6) 2 (5)
Construct (acute)
Presence/absence 3 (4) 21 (68) 2 (100)
Level of consumption 8 (26)
Impairment 1 (3)
BAC 1 (3)
Type of alcohol measure
Psychometric 11 (16) 20 (54)
Published questionnaire 8 (12) 1 (3) 3 (8)
Researcher questionnaire 36 (52) 18 (58) 9 (24)
Judgment rating 4 (6) 1 (3) 3 (8) 1 (50)
Archival 8 (12) 10 (32) 2 (5) 1 (50)
Other/cannot tell 2 (3) 1 (3)
Source for measure
Self-report 55 (80) 17 (55) 24 (65)
Archive 11 (16) 11 (36) 4 (11) 1 (50)
Observation 1 (3) 7 (19) 1 (50)
Instrument/BAC 1 (1) 1 (3)
Other 2 (3) 1 (3) 2 (5)
Violence measure
Type of violence
Physical 36 (52) 9 (29) 32 (87) 2 (100)
Sexual 4 (6) 5 (16)
Physical and sexual 26 (38) 16 (52) 4 (11)
Type of violence measure
Psychometric 5 (7) 1 (3) 11 (30)
Published questionnaire 6 (9) 5 (14)
Researcher questionnaire 28 (41) 2 (7) 9 (24)
Judgment rating 1 (1) 4 (11) 1 (50)
Archival 27 (39) 27 (87) 6 (16) 1 (50)
Source for measure
Self-report 39 (57) 2 (7) 18 (49)
Archive 28 (41) 28 (90) 7 (19) 1 (50)
Observation 1 (3) 9 (24) 1 (50)
Other 2 (3) 3 (8)
Totals 69 31 37 2
Percentages may not sum to 100 due to rounding error.
a
in Table VIII from the standard deviation and homogeneity test. These results
show that there was substantial variability across the effect size estimates
aggregated into these mean values. This indicates that different studies are
yielding effect size estimates that differ from one another much more than
would be expected on the basis of sampling error alone. We turn, then, to
consideration of some of the differences among studies that may be associ-
ated with the variability in results.
As noted earlier, the main problem in interpreting the mean correlations
between alcohol use and violent behavior is the uncertainty about the con-
founding influence of other variables. In observational studies, groups differ-
ing in level of alcohol use (or level of violence) are likely to also differ on any
of a number of other relevant variables. The research literature in this area
provides limited discussion of the types of variables that are most likely to be
confounded in the alcohol–violence correlation. While there are many vari-
ables that might be candidates, we chose to focus on three rather fundamental
categories of variables that are known to be empirically associated with either
alcohol use or violent behavior, which represent conditions temporally prior
to the violent behavior of interest and which plausibly play a rather direct role
in any differences among persons who drink more or less or who are more or
less violent. These three categories of variables are (1) stable sociodemo-
graphic characteristics that differentiate persons likely to be more or less
involved in alcohol and violence, (2) major risk factors for violence that pre-
cede and may also influence alcohol use, and (c) use of other drugs along with
alcohol that could independently affect the probability of subsequent violent
behavior.
The types of statistical control procedures that may be employed for such
confounding variables include simple ones like sample restrictions (e.g., re-
stricting the study sample only to males), matching (e.g., matching persons
who had or had not committed violent crimes on age and gender), cross-
tabulations (e.g., a cross-tabulation between violence and alcohol use by gen-
Table VIII. Individual Level Cross-Sectional Correlations (r) by Type of Relationshipa
Type of Weighted Unweighted

relationship mean r 95% CIb SD Qwc n
Criminal-chronic 0.15 0.14, 0.16 0.15 470.7* 67

Criminal-acute 0.10 0.08, 0.12 0.16 231.5* 29
Domestic-chronic 0.22 0.20, 0.24 0.16 132.7* 34
aOnly one correlation (r) per study was used for each type of relationship. When multiple correlations were
available, one was randomly selected. Each correlation was weighted by its sample size less three (see Hedges
and Olkin32), with samples sizes greater than 1000 recoded to 1000. An alcohol–violence correlation could not
be obtained for two criminal-chronic and criminal-acute studies, three domestic-chronic studies, and one
domestic-acute study. These studies had multivariate analyses of the alcohol–violence relationship and were
analyzed separately.
bCI, Confidence interval (see Hedges and Olkin32).
cQ = Homogeneity within groups (*p < 0.05, reject homogeneity).

w
der), and more sophisticated multivariate techniques (e.g., multiple regres-

sion with alcohol use and one or more control variables entered to predict
violent behavior). If confounding variables play a role in the alcohol–violence
correlation, we would expect different results from studies attempting more
control of such variables than those with less control.
The frequency with which the simple control procedures of sample re-
striction, matching, and cross-tabulations were applied in the studies repre-
sented in our meta-analysis can be seen in Table IX. The most common types
of controls were restrictions of the study sample by gender (typically re-
stricted to males), age, criminal status (i.e., restriction to convicted offenders,
delinquent youth, etc.), or socioeconomic status (usually restricted to low
ranges). Matching and cross-tabulation were seldom used. Beyond sample
restrictions on the few variables indicated, there was little use of simple
control techniques in this literature.
If there is little application of simple control procedures in the research
reporting on the empirical correlation between alcohol use and violence, we
might hope to find a correspondingly greater use of sophisticated multivariate
analysis to produce more highly controlled estimates. Our search strategy,
however, turned up only 14 criminal and 13 domestic violence studies that
conducted a multivariate analysis with alcohol use and one or more other
Table IX. Frequency of Simple Control Procedures for Confounding Variables

across All Relationship Types (n = 139)
Sample Matching
restriction between groups Cross-tabulation
Variable category n (%) n (%) n (%)
Demographics
Gender 81 (58) 2 (1) 7 (5)
Age 38 (27) 5 (4) 5 (4)
Ethnicity 9 (7) 2 (1) 1 (1)
SES 31 (22) 2 (1) 0 (0)
Income level 1 (1) 0 (0) 0 (0)
Education level 12 (9) 1 (1) 0 (0)
Employment status 2 (1) 0 (0) 1 (1)
Marital status 17 (12) 0 (0) 0 (0)
Risk factors for violence
Prior violence measure 7 (5) 0 (0) 0 (0)
Criminal status 64 (46) 0 (0) 0 (0)
Long-term recidivism 0 (0) 0 (0) 0 (0)
Observed violence in home 0 (0) 0 (0) 0 (0)
Abused as child 0 (0) 0 (0) 0 (0)
Childhood aggression 0 (0) 0 (0) 0 (0)
Alcohol and drug use
Prior alcohol use 8 (6) 0 (0) 0 (0)
Prior drug use 5 (4) 0 (0) 0 (0)
p
Table X. Variables Used in the Criminal Violence Studies That Reported Multivariate Analyses of the Alcohol–Violence Relationship
Risk for Drug

Study Violence measure Alcohol measure Sociodemographics violence use Other
Abram59 Number of arrests for violence Alcoholism 2a 1a 1a 0

Collins and Violent versus nonviolent of- Drinking at time of of- 4a 2 0 0
Schlenger60 fense fenseb
Dermen and Frequency of physical aggression Drinking habitsb 1a 2a 0 2a
George61
Fagan et al 62 Self-reported violent delinquency Drinking problems 0 2a 1a 8a
Fagan et al 63 Domestic versus domestic and Drinking while abusivec 4c 6c 1c 0
11 • Alcohol–Violence Relationship
extradomestic violence
Harrison and Violence in past year Intoxicated monthly in 8 0 2a 0
Gfroerer64 Booked for violence in past year past yearb
Kandel et al 65 Fighting Months used alcohol 0 2a 1 9a
Lawton66 Current violence score Current alcohol score 2a 1 0 0
Past alcohol score
Mann et al 67 Total self-reported violence Alcohol use 2a 4a 1 6a
V
Mason68 Number of commitments and Alcohol influence during 0 3a 1a 1a

convictions for violent offenses most recent crimeb
Miller and Person versus property offending Alcohol prior to crimeb 6a 0 0 0
Welte 69
Rathus et al70 Fist fighting MacAndrew Alcoholism 1a 0 0 0
Carrying a weapon Scaleb
Fighting w/weapon
Using force to steal
Simonds and Person versus property offending Alcohol useb 0 0 12a 0
Kashani71
Watts and Violent delinquency Alcohol useb 0 0 3c 0
Wright72
a
At least one variable was a significant predictor of the violence measure.
b
Significant independent predictor of violence measure.
c
Significance levels not reported.
269
variables as predictors and a measure of violence as the dependent (or pre-

dicted) variable. Table X summarizes the studies reporting a multivariate anal-
ysis in a criminal violence study, while Table XI summarizes the studies re-
porting a multivariate analysis in a domestic violence study. The categories
and numbers of control variables that were reported in the multivariate an-
alyses are noted, along with an indication of whether or not any were found
to be statistically significant.
Two general findings can be observed in the collection of criminal vio-
lence studies summarized in Table X. First, virtually all the studies found
control variables that accounted for a significant portion of the variance in
violence independent of the effect of alcohol. Thus, while persons with high-
er levels of alcohol consumption tended to display more violence than those
with lower levels, there were other differences between the groups that were
also capable of accounting for the differences in violence, e.g., differences in
drug use or prior history of violence. Second, in 6 of these 14 studies, no
statistically significant relationship between alcohol and violence remained
after the influence of the control variables was removed. In contrast to the
criminal studies, the domestic violence studies showed a different pattern of
findings (Table XI). While many control variables did exercise a significant
influence, 11 of the 13 studies that reported significance levels indicated that a
significant alcohol–violence relationship remained after controlling for other
variables.
Perhaps the most important finding shown in Tables X and XI, other than
how few studies there are that make sophisticated attempts to produce a
controlled estimate of the alcohol–violence relationship, is the limited selec-
tion of control variables in those studies. Only 5 of the 27 studies included
control variables from all three of the categories shown (sociodemographics,
risk for violence, and drug use). These findings, therefore, confirm the obser-
vations made by recent reviewers regarding the shortage of sophisticated
information about the nature of the statistical association between alcohol use
and violent behavior (e.g., refs. 4, 7, 40).
In order to integrate the limited information on the results of application
of the simple control techniques shown in Table IX and the few sophisticated
multivariate control studies shown in Tables X and XI, we created a crude
index of the amount of control of potentially confounding factors in each
study. In brief, we assigned points to each study for various degrees and
types of control and added those into a scale ranging from –2 to +7, with
higher values representing more control. We then correlated this “statistical
control index” with the study level effect sizes. The weighted correlation of
this index with effect size across all the studies was –0.20. This shows a
tendency for more controlled studies to yield smaller alcohol-violence cor-
relations. Since the application of well-chosen statistical controls is so infre-
quent in this literature, however, and such a limited range of control variables
is used, this result gives only a weak indication of what effects adequate
control might have. Nonetheless, it gives additional empirical confirmation to
Table XI. Variables Used in the Domestic Violence Studies That Reported Multivariate Analyses of the Alcohol–Violence Relationship
Risk for Drug

Study Violence measure Alcohol measure Sociodemographics violence use Other
Berk et al 73 Severity of female's injuries Male drinking 5a 4 0 5

Male's number of alcohol
priorsb
Chapa et al74 Mexican-American child abusers Use of alcoholb 4a 0 2a 8a
versus nonabusers
Anglo-American child abusers
versus nonabusers
Coleman et al 75 Conjugal violence versus no con- Husband's alcohol useb 1a 2a 1a 0
11 · Alcohol Violence Relationship
jugal violence
Diacatou et al 76 Father's aggressiveness toward Alcohol consumptionb 2a 0 0 0
family members
Fagan et al 63 Most serious past injury to do- Drinking while abusivec 4c 8c 1c 0
mestic partner
Famularo et al 77 Physical child maltreatment Alcohol abuseb 0 0 3a 0
Sexual child maltreatment
Hofeller78 Severity of abuse Man's use of alcoholb 2a 3a 0 2
Violent versus nonviolent mar-
riages
Julian and Physically abusive vs. nonabu- Alcohol usageb 1a 0 0 3a
McKenry79 sive male partner
Kantor and Minor spouse abuse Husband drunkb 3a 0 2a 1a
Straus80 Severe spouse abuse
Leonard and Marital aggression Alcohol Dependence 0 1 0 1
Blane81 Scale scoreb
Leonard and Premarital aggression Husband heavy drinkingb 4a 1a 0 3a
Senchak82
Miller et al 83 Parolee-to-spouse violence Alcohol problems 3 0 2 2
Wahl84 Physically abusive vs. nonabusive Alcohol abuse (SMAST)b 0 2a 0 2a
husbands
a
At least one variable was a significant predictor of the violence measure.
271
b
Significant independent predictor of violence measure.
c
Significance levels not reported.
the view that confounding variables inflate the alcohol–violence correlation in

this literature and contribute to the heterogeneity of results reported.
3.1.1. Individual Differences. One consequence of the limited application of

control variables to analysis of correlational alcohol–violence data is a short-
age of information on possible moderating variables that might differentiate
persons or circumstances in which alcohol and violence were more closely
linked. Much current theorizing emphasizes the idea that alcohol may have a
causal influence on violence only for certain persons and/or situations.5,7,41
The correlational research reviewed above, however, pays scant attention to
exploring individual differences or situational variables that may moderate
the alcohol–violence relationship. A full accounting of the variables that
might differentiate the persons or situations in which alcohol has significant
causal effects would have to include various neurobiological,12,24,42 psycho-
logical,10 psychiatric,1,11,12,43,44 and environmental5,7 factors, among others.
Combing through our meta-analysis data revealed little investigation of
such variables, and hence little that could be gleaned on the issue of moderat-
ing variables. Juvenile samples (under age 21) produced larger alcohol–vio-
lence correlations than older samples. We also found a slight tendency for
criminal–psychiatric samples (e.g., psychiatric wards of criminal institutions)
to show larger alcohol–violence correlations than general criminal samples
(e.g., incarcerated felons) or clinical treatment samples (e.g., alcoholism pa-
tients). Alcohol dose is so poorly described in these studies that no statements
about dose relationships can be made. Relatively few studies even examined
the difference between alcohol consumption and alcohol abuse in characteriz-
ing the alcohol–violence relationship. Other variables may well differentiate
cases of closer alcohol–violence association but are not sufficiently well repre-
sented in this body of studies to be examined.
3.1.2. Conclusions about Individual-Level Correlational Studies. The most

general finding of this research synthesis of individual-level correlational
studies is that the overall statistical association between alcohol use and vio-
lent behavior is positive and of nontrivial magnitude for the relationships of
chronic alcohol use to criminal behavior, acute alcohol use to criminal behav-
ior, and chronic alcohol use to domestic violence, with the latter showing the
largest correlation. Insufficient studies are available to appraise the correlation
for acute alcohol use and domestic violence.
A number of potentially important confounding variables, of course,
could easily inflate the zero-order alcohol–violence correlation. Our attempts
to explore the role of such variables was disappointing. The available research
base simply does not include sufficient control variables to permit analysis.
Even the relatively small number of studies that used multivariate techniques
often represented a very limited selection of control variables. Nonetheless,
their results tended to show a reduction in the alcohol–violence relationship
when other variables were controlled. Our own index of the overall amount
of control in each study showed a similar trend. In the few studies that did
apply sophisticated multivariate techniques to statistically control selected
confounding variables, the alcohol–violence correlation was generally re-
duced, and, in some cases, went to zero. In particular, it appears that well-
selected controls may reduce the residual correlation between alcohol and
criminal violence, although less reduction is apparent in the domestic vio-
lence relationship.
There is thus good reason to believe that the positive correlation between
alcohol and violent behavior represents a relationship that is confounded by
other variables (e.g. , sociodemographics, other drug use, early exposure to
violence, and other variables such as personality disorders). There is also
reason to believe that this correlation varies by situational context and by level
of alcohol consumption. Unfortunately, the body of literature reporting alco-
hol–violence correlations does not adequately examine these potentially con-
founding and moderating variables in a systematic fashion. Despite the large
volume of studies, there is little in this body of research that bears convincing-
ly on the issue of causality in the alcohol–violence relationship.
3.2. Macro-Level Studies

Macro-level or aggregate level studies examine the covariation between
alcohol consumption or availability and violence at the level of some social
grouping (community, city, state, etc.). Such studies make no attempt to link
the violent behavior they examine specifically to the individuals whose con-
sumption of alcohol is implicit in their aggregate measures of alcohol avail-
ability. Thus, they do not directly address the question of whether individu-
als who consume alcohol are more likely to behave violently. It is reasonable
to assume, however, that if alcohol has such effects on drinkers, the results
will show up in the aggregate statistics used in macro-level studies. In addi-
tion, of course, such studies will reflect any other pathways by which alco-
hol might lead to violence, e.g., by making potential victims more vulner-
able, by attracting potential offenders or victims to high risk environments,
and the like.
As with individual-level correlational studies, the challenge to this form
of research is to rule out the possibility that any alcohol–violence covariation
is spurious or is the result of some other factor that independently affects both
alcohol consumption and violence levels in a given region (e.g., poverty
rates).
While a number of studies of the macro-level relationship between alco-
hol availability and traffic fatalities, health indicators, and the like have been
conducted,45,46 relatively few studies have examined violent behavior as a
dependent variable. These generally use one of three approaches: investigat-
ing covariation between alcohol availability and rates of violence (1) across
groups, (2) across time (time series), or (3) some combination of group and
time comparisons.
3.2.1. Across-Group Studies. Study of covariation across social groupings is

illustrated by one of the two studies reported in Parker and Rebhun.47 Parker
analyzed data from 256 US cities on alcohol availability (number of liquor
stores per 1000 people), homicide rates from the Uniform Crime Reports
(UCR), and a set of control variables selected on theoretical grounds and
including median age, median family income, racial composition, population
density, region, migration, a social bonds composite, female labor force par-
ticipation, percent of children with one parent, and volume of retail eating
and drinking establishments. Separate models were fit for data from 1960,
1970, and 1980, with the 1970 and 1980 models including homicide rates from
the 10-year prior data, so that the dependent variable reflected change in
homicide rates over that period.
Parker found a significant relationship between alcohol availability and
homicide in the 1970 model, representing a period of rapid increase in alcohol
consumption, but not in the 1960 or 1980 model. He also showed that many
other factors in his models were related to levels and changes in homicide
rates, particularly racial composition and female labor force participation (a
proxy for amount of work and leisure time spent out of home).
In another illustrative study examining covariation across groups, Dull
and Giacopassi48 analyzed 37 cities in Tennessee with local alcohol control
ordinances of varying strength. They found positive correlations between the
permissiveness of the ordinances and the rates for each of the UCR index
crimes, but these reached statistical significance only in the cases of rape and
robbery (notably, not murder and assault). Moreover, when control variables
were introduced for poverty, age, racial composition, and the like, no signifi-
cant relationships remained between the alcohol variable and any of the UCR
index crimes. Similar results appeared for a second alcohol variable: number
of alcohol outlets per 100,000 population.
Similarly mixed results have emerged in studies analyzing covariation at
the state level for US data. Parker49 found that statewide alcohol consumption
was related to indices of two of five types of homicide and interacted with
other important predictors. Lester50 found that suicide rates were higher in
states with weaker alcohol restrictions, but that homicide rates were higher in
states with the strongest alcohol restrictions. Lester,51 using measures of alco-
hol availability, use, and abuse, however, found these to have no relationship
to statewide homicide rates.
3.2.2. Across-Time Studies. Lenke52 reported that time-series analyses of al-

cohol and violence have been around since at least 1896, when Ferri is said to
have found increased violence following good wine harvests in France. In his
own analyses, Lenke discovered significant covariation between annual per
capita alcohol sales and rates of assault and homicide in Sweden (approx.
1921–1984), rates of assault in Finland (1950–1980) and Norway (1931–1977),

and between alcohol production and assaults in France (1831–1869, 1873–
1913,1919–1958). However, significant relationships were not found between
alcohol production and homicides in France (1865–1913, 1946–1982) or be-
tween per capita alcohol sales and assaults in Denmark (1950–1978). Similarly,
Lester53 reported significant covariation between annual per capita alcohol
consumption and both male and female homicide rates in Australia from 1966
to 1985. When control variables for divorce rates and unemployment were
included in the model, however, only the relationship for males continued to
be significant.
Ensor and Godfrey54 attempted to fit a particularly sophisticated model
to the rates of six crime categories in England and Wales over the years 1960–
1988. They controlled for the probability of being caught and prosecuted and
for prosperity using variables chosen to reflect opportunities for crime and the
opportunity cost of crime in a model designed to represent interactions be-
tween alcohol, crime levels, and the criminal justice system response to
crime. Per capita alcohol consumption was found to be a significant predictor
of rates of violent crime as well as rates of several types of property crimes.
Alcohol consumption was also found to be a predictor of the probability of
being caught for violent crimes, suggesting one way in which alcohol involve-
ment might be overrepresented among apprehended offenders.
3.2.3. Across-Time and -Group Studies. Some macro-level studies provide a

combined analysis of covariation across both social groups and time. Cook
and Moore,55,56 for instance, examined the relationship between state-level
beer taxes and alcohol consumption (based on sales data) and rates of homi-
cide, rape, robbery, and assault (from the UCR). They analyzed annual data
for each of the contiguous 48 states for the years 1979–1988 using dummy
variables for each state to control for unobserved, persistent state-specific
determinants of crime rates. They found that variations in the beer tax were
associated with alcohol consumption, which was lower when taxes were
higher. Most important for present purposes, they found significant covaria-
tion between alcohol consumption and crime rates for rape, assault, and
robbery, but not for murder.
Parker41 criticized the Cook and Moore study for omitting variables such
as poverty, routine activity levels, racial composition, and the like that are
linked to violence rates and were themselves changing over the time period at
issue. Parker and Rebhun,47 in one of their two studies, examined the rela-
tionship of changes in state minimum age of purchase laws for alcohol with
homicide rates over the period 1976–1983. Prior to this period, 29 states low-
ered the minimum age below 21 years and others already had lower mini-
mum ages. By 1988, all of these states had raised the minimum age to 21
years. Parker used a pooled cross-section time series analysis on data for each
of the 50 states plus the District of Columbia for each year from 1976 through
1983, using a dummy code to identify the year in which the drinking age was
raised. Other variables in the model included alcohol consumption (beer sales
in barrels per capita), infant mortality (as a poverty index), an index of in-
equality, racial composition, region, and total state population.
Parker and Rebhun47 applied this model to six different homicide rates:
“primary” when victim and offender knew each other, “nonprimary” when
there was no known relationship between them, with each of these further
divided according to offender age (15–18, 19–20, and 21–24). The results
showed that beer consumption was positively and significantly related to all
categories of homicide except primary, age 21–24 (though the coefficient for
the latter was in the same direction). In addition, change in the minimum age
of purchase law was significantly related to reductions in homicide for the
primary, age 21–24 category, but not to the others (though these coefficients
were in the same direction). Other significant variables across the models
were infant mortality (poverty index), racial composition, region (South vs.
other), and total state population.
3.2.4. Natural Experiments. Studies like Cook and Moore55,56 and Parker and
Rebhun,47 in which relatively sudden events restrict the availability of alco-
hol, constitute a kind of natural experiment. Whereas there are many co-
occurring trends in typical time series that could plausibly account for the
covariation in alcohol consumption and violence (e. g. , unemployment), few
of these trends are likely to show sudden changes that coincide with the
timing of such “external” events as new legislation or other social incidents
that restrict access to alcohol. Such episodes, therefore, provide an especially
good opportunity to examine the macro-level relationship between alcohol
and violence.
Lenke52 summarized an interesting group of natural experiments set in
Sweden: (1) the rationing of alcohol during World War I, (2) the repeal of the
general alcohol restriction system in 1955, (3) the strike at the Alcohol State
Monopoly in 1963, (4) the legalization of sales of medium beer in grocery
stores in 1965, (5) the legalization of sales of strong beer in grocery stores in
some provinces in 1967, and (6) the discontinuation of Saturday open hours at
the Alcohol Monopoly Stores in 1981. While the results were not totally con-
sistent, Lenke summarized as follows:
The general conclusion from the cases described . . . is that changes and
variation in the availability and consumption of alcohol tend to affect
crimes of violence. When availability of alcohol has been reduced or in-
creased, the rates of violent crimes have tended to follow the same direc-
tion. (p. 103)
Similar results were found in several, but not all, analogous natural experi-
ments in Finland and Norway (summarized in refs. 52, 55, 56).
3.2.5. Conclusions about Macro-Level Correlational Studies. The authors of
many of these macro-level studies argue plausibly that their various results
have important policy implications. Demonstration of covariation between
alcohol availability and rates of violence and, especially, evidence that policy
actions such as alcohol taxes and drinking age laws may affect violence rates
are indeed provocative. For purposes of assessing the causal role of alcohol
consumption in violent behavior, however, the macro-level studies available
are less than conclusive. First, not all of these investigations have revealed a
significant covariation between alcohol availability and rates of violence. In
addition, there is uncertainty regarding the appropriate control variables to
include in the models to account for other factors that may co-occur with
greater or lesser levels of alcohol availability, and relatively few of the range of
possibilities have been examined. Also, these studies do not distinguish be-
tween alcohol consumption per se and other co-occurring factors (such as
male gatherings) that might be responsible for any effects found. Nonethe-
less, given the level of aggregation represented in these studies, it is notable
that the alcohol–violence relationship appears with sufficient strength that it
cannot be readily dismissed.
It is unfortunate that few of the macro-level studies have attempted any
probing of their data to better pinpoint the particular persons or circum-
stances most responsible for the overall alcohol–violence relationships discov-
ered. One exception is Lenke,52 who included a liver cirrhosis mortality vari-
able to distinguish rates of heavy, long-term drinking. This variable was
found to be significantly related to homicide rates, but not assaults, in his
Swedish time series. Lenke also cited evidence from several of the natural
experiments he reviewed that indicated that heavy drinkers may have played
a disproportionate role in the changes in the rates of violent crime that were
recorded.
4. Overall Conclusions
Aside from the many specific conclusions drawn above regarding the
research in each of the broad categories reviewed, we believe three general
conclusions are warranted:
1. The research base relevant to the question of the causal role of alcohol
consumption in violent behavior, despite its overall volume, is very unsat-
isfactory. It is permeated by problems of inadequate experimental and statisti-
cal control, questionable generalizability to socially important forms of vio-
lence, limited attention to individual differences and moderator variables,
weak conceptualizations of the issue, and capricious operationalizations of
the key variables. As a result, the causal issue is still cloudy and uncertain.
Some of the difficulties are inherent in the nature of the issue under study, but
much is remediable. The alcohol–violence relationship is not merely an aca-
demic issue, it is one with important social implications that deserves more
systematic, careful, and probing attention from researchers.
2. While granting the inadequacies of available research, it is nonetheless
important to recognize that none of the relevant bodies of research yield a
consistently or even predominately negative or null result on the causal ques-

tion. Each provides substantial evidence of an alcohol–violence associa tion that
is consistent with a causal interpretation and none yet provides evidence
sufficient to rule out such an interpretation. One is tempted to say that where
there is this much smoke, there must be some fire. Such a conclusion would
clearly be premature given the deficiencies of the empirical evidence, but the
possibility highlights the importance of continued research on this issue.
3. While a causal influence of alcohol consumption on violence cannot be
ruled out with present evidence, it seems apparent that there is no broad,
reliable, “main effect” of alcohol on violence, analogous to the easily demon-
strated and almost ubiquitous effects on motor and cognitive functioning that
occur at sufficient doses. If alcohol has any causal effects on violence, they
almost certainly occur only for some persons and/or some circumstances. The
most important research question regarding the alcohol–violence relation-
ship, therefore, is not one of global causal influence. Rather, it is the more
focused question of what individual differences, moderator, and situational
variables characterize circumstances in which alcohol might potentiate violent
behavior. The greatest failure of the research reviewed in this chapter is the
inadequate attention to this question and, as a consequence, the inability to
address the most pressing social issues involving alcohol and violence.
ACKNOWLEDGMENT. Support for the preparation of this chapter has been pro-
vided in part by the Washington Technical Information Group, Inc. and the
Dean’s Fund of the Owen Graduate School of Management, Vanderbilt Uni-
versity. The views stated in this chapter are those of the authors and do not
necessarily represent the positions of Washington Technical Information
Group, Inc.
References
1. Collins JJ: Drinking and violence: An individual offender focus, in Martin SE (ed): Alcohol and
Interpersonal Violence: Fostering Multidisciplinary Perspectives (Research Monograph No. 24).
Washington, DC, National Institute on Alcohol Abuse and Alcoholism, 1993, pp 221–235.
2. Reiss AJ, Roth JA: Alcohol, other psychoactive drugs, and violence, in Reiss AJ, Roth JA (eds):
Understanding and Preventing Violence. Washington, DC, National Academy Press, 1993, pp
182–220.
3. Miczek KA, DeBold JF, Haney M, et al: Alcohol, drugs of abuse, aggression, and violence, in
Reiss AJ, Roth JA (eds): Understanding and Preventing Violence: vol 3, Social Influences. Washing-
ton, DC, National Academy Press, 1994, pp 377–570.
Interpersonal Violence: Fostering Multidisciplinary Perspectives (Research Monograph No. 24).
of violent events, in Martin SE (ed): Alcohol and Interpersonal Violence: Fostering Multidisciplin-
ary Perspectives (Research Monograph No. 24). Washington, DC, National Institute on Alcohol
Abuse and Alcoholism, 1993, pp 161–192.
6. Greenberg SW: Alcohol and crime: A methodological critique of the literature, in Collins JJ
(ed): Drinking and Crime. New York, Guilford Press, 1981, pp 70–109.
7. Pernanen K: Alcohol-related violence: Conceptual models and methodological issues, in

Martin SE (ed): Alcohol and Interpersonal Violence: Fostering Multidisciplinary Perspectives (Re-
search Monograph No. 24). Washington, DC, National Institute on Alcohol Abuse and Alco-
holism, 1993, pp 37–70.
8. Walfish S, Blount WR: Alcohol and crime: Issues and directions for future research. Special
Issue: Alcohol and the criminal justice system. Criminal Justice Behav 16:370–386, 1989.
9. Mackie JL: Causes and conditions. Am Philosoph Q 2:245–264, 1965.
10. Lang AR: Alcohol-related violence: Psychological perspectives, in Martin SE (ed): Alcohol and
Interpersonal Violence: Fostering Multidisciplinary Perspectives (Research Monograph No.
24). Washington, DC, National Institute on Alcohol Abuse and Alcoholism, 1993, pp 121–
147.
11. Fagan J: Intoxication and aggression, in Tonry M, Wilson JJ (eds): Drugs and Crime. Chicago,
University of Chicago Press, 1990, pp 241–320.
12. Lang AR, Sibrel PA: Psychological perspectives on alcohol consumption and interpersonal
aggression: The potential role of individual differences in alcohol-related criminal violence.
Criminal Justice Behav 16:299–324, 1989.
13. Pernanen K: Research approaches in the study of alcohol-related violence. Alcohol Health Res
World 17:101–107, 1993.
14. Collins JJ: Suggested explanatory frameworks to clarify the alcohol useiviolence relationship.
Contemp Drug Problems 15:107–121, 1988.
15. Collins JJ: Alcohol and interpersonal violence: Less than meets the eye, in Weiner NA,
Wolfgang, ME (eds): Pathways to Criminal Violence. Newbury Park, CA, Sage, 1989, pp 49–57.
16. Evans CM: Alcohol and violence: Problems relating to methodology, statistics, and causation,
in Brain PF (ed): Alcohol and Aggression. Dover, NH, Croom Helm, 1986, pp 138–160.
17. McCord J: Considerations of causes in alcohol-related violence, in Martin SE (ed): Alcohol and
Znterpersonal Violence: Fostering Multidisciplinary Perspectives (Research Monograph No. 24).
18. Grayham K, La Rocque L, Yetman R, et al: Aggression and bar-room environments. J Stud
Alcohol 41:277–292, 1980.
and aggression in infra-human animals, in Brain PF (ed): Alcohol and Aggression. Dover, NH,
Croom Helm, 1986, pp 84–137.
20. Brain PF, Miras RL, Berry MS: Diversity of animal models of aggression: Their impact on the
putative alcoholiaggression link. J Stud Alcohol Suppl 11:140–145, 1993.
21. Berry MS, Brain PF: Neurophysiological and endocrinological consequences of alcohol, in
Brain PF (ed): Alcohol and Aggression. Dover, NH, Croom Helm, 1986, pp 19–54.
22. Miczek KA, Weerts EM, DeBold JF: Alcohol, benzodiazepine-GABA, receptor coomplex and
aggression: Ethological analysis of individual differences in rodents and primates. J Stud
Alcohol Suppl 11:170–179, 1993.
23. Miczek KA, DeBold JF, VanErp AM: Neuropharmacological characteristics of individual dif-
ferences in alcohol effects on aggression in rodents and primates. Behav Pharmacol 5:407–421,
1994.
24. Miczek KA, Weerts EM, DeBold JF: Alcohol, aggression, and violence, in Martin SE (ed):
Alcohol and Interpersonal Violence: Fostering Multidisciplinary Perspectives (Research Monograph
No. 24). Washington, DC, National Institute on Alcohol Abuse and Alcoholism, 1993, pp 83–
119.
25. Peterson JT, Pohorecky LA: Effect of chronic ethanol administration on intermale aggression
in rats. Aggress Behav 15:201–215, 1989.
26. Cohen MA, Lipsey MW, Wilson DB, Derzon JH: The Role of Alcohol Consumption in Violent
Behavior: Final Report. Research Report, VIPPS, Vanderbilt University, 1994.
27. Bushman BJ: Human aggression while under the influence of alcohol and other drugs: An
integrative research review. Curr Direct Psychol Sci 2:148–152, 1993.
29. Hull JG, Bond CF: Social and behavioral consequences of alcohol consumption and expectan-
cy: A meta-analysis. Psychol Bull 99:347–360, 1986.
30. Steele CM, Southwick L: Alcohol and social behavior I: The psychology of drunken excess.
31. Cooper H, Hedges LV (eds): The Handbook of Research Synthesis. New York, Russell Sage
Foundation, 1994.
32. Hedges LV, Olkin I: Statistical Methods for Meta-Analysis. New York, Academic Press, 1985.
33. Gustafson R: Threat as a determinant of alcohol-related aggression. Psychol Rep 58:287–297,
1986.
34. Heermans HW: The effect of alcohol, arousal and aggressive cues on human physical aggres-
sion in males (doctoral dissertation, Rutgers University, 1980). Dissertation Abstracts Interna-
tional 41:3630B (University Microfilms No. 81-05011), 1981.
35. Taylor SP: Experimental investigation of alcohol-induced aggression in humans. Alcohol
36. Gustafson R Alcohol and aggression. J Offender Rehab 21:41–80, 1994.
37. Rosenthal R, Rosnow RL (eds): Artifact in Behavioral Research. New York, Academic Press, 1969.
38. HunterJE, Schmidt FL: Methods ofMeta-Analysis: Correcting Error and Bias in Research Findings.
Newbury Park, CA, Sage, 1990.
39. Cohen J: Statistical Power Analysis for the Behavioral Sciences, 2nd ed. Hillsdale, NJ, Lawrence
Erlbaum, 1988.
40. Fagan J: Interactions among drugs, alcohol, and violence. Health Aff 12:65–79, 1993.
41. Parker RN: Rational choice and pooled cross-section time series: Theoretical and meth-
odological pathways to new understanding of the alcohol/violence relationship, in Martin SE
(ed): Alcohol and Interpersonal Violence: FosteringMultidisciplinary Perspectives (Research Mono-
graph No. 24). Washington, DC, National Institute on Alcohol Abuse and Alcoholism, 1993,
pp 213–220.
42. Roth JA: Psychoactive Substances and Violence. Washington, DC, National Institute of Justice
(NCJ Document Reproduction Service No. 145534), 1994.
25:1065–1081, 1990.
44. Swanson JW: Mental disorder, substance abuse, and community violence: An epidemiologi-
cal approach, in Monahan J, Steadman HJ (eds): Violence and Mental Disorder: Developments in
Risk Assessment. Chicago, University of Chicago Press, 1994, pp 101–136.
45. Grossman M, Chaloupka FJ, Saffer H, Laixuthai A: Effects of alcohol price policy on youth: A
summary of economic research. J Res Adolesc 4:347–364, 1994.
46. Gruenewald PJ, Millar MS, Treno AJ: Alcohol availability and the ecology of drinking behav-
ior. Alcohol Health Res World 17:39–45, 1993.
47. Parker RN, Rebhun L: Alcohol and Homicide: A Deadly Combination of Two American Traditions.
Albany, NY, State University of New York Press, 1995.
48. Dull RT, Giacopassi DJ: The impact of local alcohol ordinances on official crime rates: The case
of Tennessee. Justice Q 4:311–323, 1987.
49. Parker RN: Bringing “booze” back in: The relationship between alcohol and homicide. J Res
Crime Delinq 32:3–38, 1995.
50. Lester D: Restricting the availability of alcohol and rates of personal violence (suicide and
homicide). Drug Alcohol Depend 31:215–217, 1993.
51. Lester D: Alcohol availability, alcoholism, and suicide and homicide. Am J Drug Alcohol Abuse
21:147–150, 1995.
52. Lenke L: Alcohol and Criminal Violence: Time Series Analysis in a Comparative Perspective. Stock-
holm, Almqvist & Wiksell, 1990.
53. Lester D: Alcohol consumption and rates of personal violence in Australia. Drug Alcohol
Depend 31:15–17, 1992.
54. Ensor T, Godfrey C: Modelling the interactions between alcohol, crime and the criminal
justice system. Addiction 88:477–487, 1993.
55. Cook PJ, Moore MJ: Economic perspectives on reducing alcohol-related violence, in Martin
SE (ed): Alcohol and Interpersonal Violence: Fostering Multidisciplinary Perspectives (Research
Monograph No. 24). Washington, DC, National Institute on Alcohol Abuse and Alcoholism,
1993, pp 193–212.
56. Cook PJ, Moore MJ: Violence reduction through restrictions on alcohol availability. Alcohol
Journal of Abnormal Psychology 88:153–160, 1979.
58. Zeichner A, Pihl RO: The effects of alcohol and instigator intent on human aggression. J Stud
Alcohol 41:265–276, 1980.
59. Abram KM: Effect of co-occurring disorders on criminal careers: Interaction of antisocial
personality, alcoholism, and drug disorders. International Journal of Law & Psychiatry 12:133–
148, 1989.
60. Collins JJ, Schlenger WE: Acute and chronic effects of alcohol use on violence. Journal of
Studies on Alcohol 49:516–521, 1988.
61. Dermen KH, George WH: Alcohol expectancy and the relationship between drinking and
physical aggression. Journal of Psychology 123:153–161, 1989.
62. Fagan J, Piper ES, Moore M: Violent delinquents and urban youth. Criminology 23:439–466,
1986.
63. Fagan J, Hansen KV, Stewart DK: Violent men or violent husbands? Background factors and
situational correlates of violence toward intimates and strangers. In Finkelhor D, Gelles RJ,
Hotaling GT, Straus MA (eds): The dark side of families: Current family violence research. Beverly
Hills, CA, Sage, 1983, p 49–68..
64. Harrison L, Gfroerer J: The interaction of drug use and criminal behavior: Results from the
National Household Survey on Drug Abuse. Crime and Delinquency 38:422–433, 1986.
65. Kandel D, Simcha-Fagan R, Davies M: Risk factors for delinquency and illicit drug use from
adolescence to young adulthood. Journal of Drug Issues 16:67–90, 1986.
66. Lawton EG: Persons who injure others: A study of criminally assaultive behavior in Philadel-
phia (Doctoral dissertation, Temple University). Dissertation Abstracts International 40:3549A,
1979.
67. Mann F, Friedman CJ, Friedman AS: Characteristics of self-reported violent offenders versus
court identified violent offenders. International Journal of Criminology and Penology 4:69–87,
1976.
68. Mason FH: Demographic and psychological test variables associated with prior offender
violence (Doctoral dissertation, University of South Carolina). Dissertation Abstracts Interna-
tional 47:1790B, 1986.
69. Miller BA, Welte JW: Comparison of incarcerated offenders according to use of alcohol and/or
drugs prior to offense. Criminal Justice and Behavior 13:366–392, 1986.
70. Rathus SA, Fox JA, Ortins JB: The MacAndrew Scale as a measure of substance abuse and
delinquency among adolescents. Journal of Clinical Psychology 36:579–583, 1980.
71. Simonds JF, Kashani J: Specific drug use and violence among delinquent boys. American
Journal of Drug & Alcohol Abuse 7:305–322, 1979.
72. Watts WD, Wright LS: The drug use-violent delinquency link among adolescent Mexican-
Americans. In De La Rosa M, Lambert EY, Gropper B (eds): Drugs and violence: Causes,
correlates, and consequences (Research Monograph Series No. 103). Washington, DC, National
Institute on Drug Abuse, 1990, pp 136–159.
73. Berk RA, Berk SF, Loseke DR, Rauma D: Mutual combat and other family violence myths. In
Finkelhor D, Gelles RJ, Hotaling GT, Straus MA (Eds): The dark side of families: Current family
violence research. Beverly Hills, CA, Sage, 1983, pp 197–212.
74. Chapa D, Smith PL, Rendon FV, Valdez R, Yost M, Cripps T: The relationship between child
abuse and neglect and substance abuse in a predominantly Mexican-American population. In
Lauderdale ML, Anderson RN, Cramer SE (eds): Child abuse and neglect: Issues on innovation
and implementation (Proceedings of the Second Annual National Conference on Child Abuse
and Neglect, vol. 1). Washington, DC, National Center on Child Abuse and Neglect, 1978, pp
116-125.
75. Coleman KH, Weinman ML, Hsi BP: Factors affecting conjugal violence. Journal of Psychology
105:197–202, 1980.
76. Diacatou A, Mamalakis G, Kafatos A, Vlahonikolis J, Bolonaki I: Alcohol, tobacco, and fa-
ther’s aggressive behavior in relation to socioeconomic variables in Cretan low versus medi-
um income families. International Journal of the Addictions 28:293–304, 1993.
77. Famularo R, Kinscherff R, Fenton T: Parental substance abuse and the nature of child mal-
treatment. Child Abuse & Neglect 16:475–483, 1992.
78. Hofeller KH: Social, psychological, and situational factors in wife abuse. Palo Alto, CA, R&E
Research Associates, 1982.
79. Julian TW, McKenry PC: Mediators of male violence toward female intimates. Journal of Family
Violence 8:39–56, 1993.
80. Kantor GK, Straus MA: Substance abuse as a precipitant of wife abuse victimizations. Ameri-
can Journal of Drug & Alcohol Abuse 15:173–189, 1989.
81. Leonard KE, Blane HT: Alcohol and marital aggression in a national sample of young men.
Journal of Interpersonal Violence 7:19–30, 1992.
82. Leonard KE, Senchak M: Alcohol and premarital aggression among newlywed couples.
Journal of Studies on Alcohol 54(Suppl. No. 11):96–108, 1993.
83. Miller BA, Nochajski TH, Leonard KE, Blane HT, Gondoli DM, Bowers PM: Spousal violence
and alcohol/drug problems among parolees and their spouses. Women and Criminal Justice
1:55–72, 1990.
84. Wahl JA: Self-esteem, attitudes toward women, alcohol abuse and history of family violence
in spouse battering males (Doctoral dissertation, Oklahoma State University 1987). Disserta-
tion Abstracts International 19:1961B, 1988.
12
Alcohol and Cocaine Interactions

and Aggressive Behaviors
M. Elena Denison, Alfonso Paredes,
and Jenia Bober Booth
Abstract. This chapter presents (1) a review of several studies on the relationship between
violent/aggressive behavior and the use of cocaine and/or the use of alcohol; and (2) findings
from our study of cocaine-dependent men, illustrating deviant and violent behavior before and
during cocaine addiction careers. As had been found in previous research, use of alcohol and
cocaine seemed to increase the likelihood of the cocaine users in our sample engaging in deviant
or violent behaviors. The extent of deviant or violent behavior, in our sample, during periods of
cocaine use, periods of cocaine–alcohol use, periods of alcohol use only, and periods of absti-
nence for both alcohol and cocaine are discussed. Changes in the nature of the deviant or violent
behaviors prior to and after the onset of cocaine addiction are also described.
1. Introduction
Deviant behaviors including crime occur at high rates among cocaine users,
and violence has long been associated with alcohol use; a better understand-
ing of the behavioral correlates of combined and individual use of alcohol and
cocaine is needed. In this chapter, we outline findings from several studies
that examined the relationships between violent/aggressive behavior and the
use of alcohol and other drugs. This brief overview is followed by a presenta-
tion of results derived from our study of cocaine-dependent men to illustrate
M. Elena Denison, Alfonso Paredes, and Jenia Bober Booth • Laboratory for the Study of
Addictions and UCLA Drug Abuse Research Center, West Los Angeles Veterans Administration,
Los Angeles, California 90073.
283
some of the relationships between violence and use of cocaine and alcohol
that manifest in cocaine-dependent individuals.
2. Alcohol and Violence

Epidemiological investigations have found a close association between
human violence and alcohol use. Drinking of alcohol has been implicated in
domestic violence, homicide, sexual assaults, suicide, and traffic accidents.
Aggressive behavior also seems to be concomitant with the use of drugs such
as cocaine.1 The mechanisms responsible for these putative interactions are
complex. Alcohol and other drugs may act through pharmacological mecha-
nisms that inhibit neurobehavioral systems, which under normal circum-
stances control aggression. For instance, impulsive alcoholic violent offenders
appear to have a deficit of brain serotonin metabolism and low concentration
of 5-hydroxyindoleacetic acid in the cerebrospinal fluid.2 At a behavioral lev-
el, alcohol may facilitate aggressive behavior by reducing awareness and by
impairing intellectual functioning and risk assessments.3
Individuals with certain personality features are more likely to exhibit
heightened destructiveness when under the influence of alcohol. For exam-
ple, adolescents with conduct disorder, in general [as defined in the Diagnos-
tic and Statistical Manual of Mental Disorders, 4th edition (DSM-IV)], con-
sume more alcohol and show increased violence; even those without conduct
disorder demonstrate heightened aggression if they are alcohol users.4
The mechanisms involved in the alcohol–violence relationship have been
investigated experimentally.5,6 These investigators suggest that alcohol ef-
fects are selective and do not disrupt higher cognitive functions “across the
board.” When exposed to low levels of personal threat, inebriated individuals
allocated equal attention to all situational cues, in which case alcohol did not
impair the salience of relevant information; subjects did not respond aggres-
sively. Under high levels of personal threat, however, alcohol-intoxicated
individuals give greater attention to negative or personally threatening cues
and display more intense aggression. Under placebo conditions, subjects do
not respond differentially to positive and negative personal information and,
in general, showed low levels of aggression.5 However, Zeichner et al.6 also
found that the consumption of alcohol had a dampening effect on the physi-
ological effects of situational stress (decrease in heart rate and systolic blood
pressure). Thus, aggressive reactions of hostile men under the influence of
alcohol may be due more to the psychological effects than the physiological
effects of alcohol.
3. Cocaine and Violence

Violence may be strongly influenced by social circumstances. In the case
of cocaine, individuals moved by their compulsion to use the drug may en-
12 • Alcohol–Cocaine Interaction 285
gage in violence to obtain resources for drugs. In addition, violence is often an

outcome of territorial struggles of groups or individuals who try to control
procurement and distribution of illicit drugs.7 In some urban centers where
cocaine use prevalence is high, firearm aggression, including homicide, in-
creased as much as 123% within a 5-year period.1 The association of cocaine
and violent crime as an economically motivated event was suggested by the
authors, but this relationship was apparently found to be weak. Notably,
violence observed among cocaine users was not a function of compulsive
drug seeking, while concomitant use of alcohol was related to more than 70%
of the severe violent events observed in these individuals.
The nature of the interactions between alcohol and violence and between
cocaine and violence differs in important ways. It has been suggested that the
alcohol relationship to violence appears to be primarily pharmacological,
while the cocaine relationship to violence is primarily “social–systemic.”8 In
another study,9 daily use of crack cocaine and the frequency of illicit criminal
activities aimed to obtain a supply of the drug appeared to be interrelated, but
dose–response effects were difficult to demonstrate. The level of violence is
only partially related to the method of cocaine administration. Violence of
greater intensity tended to be associated with freebase crack cocaine smoking
or intravenous administration in contrast with intranasal use, but this rela-
tionship was not strong. Violent actions requiring sustained activity such as
rape, burglary, and armed robbery are not found to be related to the route of
administration of the drug used by the perpetrator.10
Drug-related activities such as trafficking are accompanied by increased
mortality and account for one third to one half of homicide-related deaths among
cocaine users.11 The authors found that drug dealing was associated with other
high-risk behaviors, including nonfatal violence and the probability of being
incarcerated. The authors stated that, in spite of commonly held beliefs, drug
involvement and weapon carrying are not normative among youth. Young men
involved in drug dealing or weapon carrying are a subset of individuals more
likely to participate in a wide variety of delinquent (or high-risk) behaviors. The
subjects in Stanton and Galbraith’s11 study show poor school performance and
greater likelihood of receiving disciplinary dismissals. These men usually
reported poor relationships with their parents. The perception of drug dealing
as the only viable income-generating option has also been found to predict drug
trafficking and violence.12 In contrast, among inner-city male adolescents in
New York City, type of crime including violent crime was not related to drug
dealing but to the use of alcohol and drugs (such as marijuana and intranasal
heroin). Again, though, these youths reported multiple arrests, truancy, and
psychological distress, gave history of sexual molestation as children, sub-
stance-abusing parents, and cocaine/crack-using friends.13
There is no progressive linear relationship between level of drug use and
gun possession, number of guns owned, and routinely carrying guns as
indicator of proneness to violence. On the other hand, individuals who sold
drugs were more likely to own guns.14 In Los Angeles, gang-related homi-
cides have increased from 18 to 43% within the past 15 years, but drug
286 Ill • Psychology
trafficking has not been a major factor in this increase.15–17 The authors sug-
gest that in crack cocaine abusers, crack distribution and drug selling may not
create violent or actively criminal individuals, but rather individuals who
have a tendency toward violence may be attracted to this activity. However,
in the Belenko et al.18 study, crack addicts had significant increases in the
seriousness of the crimes committed, as well as increases in the rate of crimi-
nal activity after initiation of crack use. In addition to this sample of crack
addicts being more likely to be violent overall, the increase for violent arrests
was more than three times larger than the increase among powder cocaine
defendants. Belenko et al.18 found a general acceleration in arrests following
initiation into crack cocaine use, suggesting that the diversity and frequency
of involvement in nondrug crime implies a more generalized pattern of devi-
ance underlying their behavior.
Several investigators have identified two major subpopulations of co-
caine-dependent individuals. One group uses mainly cocaine, while the other
group (of approximately equal size in some samples) uses cocaine and alcohol
severely enough to meet the diagnosis of abuse for both substances.19,20
Users of alcohol and cocaine and users of cocaine only differ along impor-
tant dimensions. In a nationally representative sample of young adults, com-
bined alcohol–cocaine abusers demonstrated higher levels of drug use in the
prior 30 days and lifetime drug use, higher levels of delinquent activity, and
higher rates of unemployment and marital instability. A decreasing order of
problem severity was observed: cocaine and alcohol abusers had the most
severe problems, followed by cocaine-only users and by alcohol-only abusers.
Interestingly, however, the study noted that joint alcohol–cocaine abusers
reported the lowest high school dropout rate and scored the highest on mea-
sures of verbal intelligence.21
4. Drug Use Status and Deviant Behaviors: Results of the Study

Our research offered an opportunity to conduct a preliminary exploration
of some of these interactions. We had available a large database from subjects
who participated in our investigations of the progression of cocaine depen-
dence and factors that influence clinical and treatment outcomes. With these
data, we initiated a preliminary examination of the behaviors of our subjects
during periods of different states of drugs use: cocaine only, alcohol only,
cocaine and alcohol in combination, and abstinence from both alcohol and
cocaine. Retrospective data covering the period lasting an average of approx-
imately 10 years from the first use of cocaine to admission to treatment for
cocaine dependence were used in this investigation.
4.1. Characteristics of the Study Sample

The subjects selected for this study were veterans of military service who
requested treatment for cocaine dependence at the Veterans Affairs Medical
Center in West Los Angeles. The population served by this hospital includes
very few women; thus, the 320 subjects selected for this study were all men.
Participants met DSM-III-R criteria for cocaine dependence and signed a con-
sent form approved by the Human Subjects Subcommittee from the medical
center, An additional 63 men from this population who were asked to partici-
pate in the research projects refused. The main reasons given were lack of
interest, skepticism about assurance of confidentiality, unwillingness or ap-
prehension about disclosing previous drug or criminal history, and lack of
time to devote to research interview.
In regard to the sociodemographic characteristics of this group, at admis-
sion, 30% of the men were married, 25% divorced, 17% separated, 1% were
widowed, and 27% had never married. Eighty-six percent completed at least
twelfth grade. Twenty-eight percent did not work during the 4 weeks preced-
ing entry into treatment, 58% worked full time, and the remaining 14%
worked less than full time. Eighty-two percent had been incarcerated at some
time in their lives and 13% were on probation at the time of the intake
Table I. Demographic Characteristics

of the Sample (n = 320)
Characteristic Percent
Age at interview, yearsa

21–30 21.8
31–40 56.7
>41 21.5
Ethnicity
African American 67.5
Caucasian 24.7
Hispanic 6.9
Other 0.9
Education, years
<12 13.8
12 42.8
>12 43.4
Main occupation at admission
Skilled 15.0
Semiskilled 33.7
Unskilled 9.1
Unemployed 42.2
Ever incarcerated (self-report) 81.8
Age at first arrest, yearsb
<18 12.3
18–24 39.4
>24 48.3
Ever married 72.2
Homeless at admission 10.9
aMean age at interview = 35.5 years old (SD = 6.4).
bMean age at first incarceration = 25.1 years old (SD =
7.2).
interview. Additional information on the sociodemographic profile of the

sample is given in Table I.
4.2. Data Collection Methodology

Specially designed schedules were used at project intake to collect socio-
demographic data, developmental history, and information on social func-
tioning and history of treatment. Another instrument, the Natural History
Interview, was the main source of detailed longitudinal information concern-
ing the patterns of cocaine and other drug use. This instrument is an adapta-
tion of a schedule developed by Nurco and his colleagues22 and is described
in detail elsewhere.19 Briefly, a schematic time chart is prepared during a face-
to-face interview showing important events in the subject’s life, such as initia-
tion of drug use, treatment episodes, dates of marriage, childbirths, geo-
graphic moves, arrests, and periods of incarceration and legal supervision.
These events are used as time-anchoring points in order to facilitate the re-
view of significant aspects of the subject’s drug and social history as auto-
biographical sequences and have been shown to provide one method that
subjects can use to organize and anchor their memories of personal events.23
With the assistance of the subject, a trained interviewer establishes the
date of first cocaine use, records in on a time chart, and proceeds chronologi-
cally from 12 months before first use to the time of the interview. Data are
collected on the use of cocaine, alcohol and other drugs, health status, drug
treatment experience, employment, family adjustment, criminal behavior,
and other social adjustment variables for the initial time period. Subsequent
time periods are demarcated by changes in levels of cocaine use and signifi-
cant life events such as any social adjustment variable, subject’s drug treat-
ment, or legal status. The interviewer proceeds to review each successive time
segment, repeating the systematic process of collecting retrospective longi-
tudinal data covering long periods of the addicts history.24 From these raw
data, monthly rates of various categories of behaviors become available within
each segment. These rates are invariant or constant within each time period
with respect to items such as type of drug use, methods of administration,
work tenure, interpersonal relationships, drug dealing, and property crimes.
4.3. Cocaine Use Patterns

The mean age of the subjects at the time of first cocaine use was 24 years
old (SD = 7 years). The main reasons given for initiating use were to satisfy
curiosity (65%), to get a “good high” (8%), peer acceptance (7%), friends were
using (7%), to enhance sexual pleasure or to procure sex partners (3%), and to
relieve depression (3%). Subjects often gave more than one reason for initiat-
ing cocaine use.
In most cases a friend or acquaintance was instrumental in introducing
the person to cocaine use (76%); in 12% of the cases a parent, sibling, or other
relative played a role in the initiation; for 870, either a dealer, a prostitute, or
other person; and for 4%, a girlfriend or wife influenced first cocaine use. The
first supply of cocaine was obtained free by 84% of the respondents and 16%
reported purchasing the drug. Forty-two percent of the subjects reported that
they were not high on any other drug when they first used cocaine, though
26% of them reported they were high on alcohol the first time they used
cocaine, 22% were high on marijuana, and 10% were high on other drugs.
The favored route of administration at first cocaine use (FCU) was intra-
nasal (75%). Less-preferred routes were smoking crack (11%), smoking cocaine
as primo (marijuana cigarettes spiked with cocaine) or rails (tobacco cigarettes
with cocaine) (both totaling 5%), intravenous (5%), and freebasing (4%).
The average time span from FCU to treatment entry (TE) for the group
was 10 years (standard deviation = 6 years). The proportion of users using
cocaine intranasally decreased from 75% at initiation of the cocaine use career
to 7% at the time of TE and admission to the research program. On the other
hand, crack use increased from 11% at initiation to 64% at the time of TE.
4.4. Other Drug Use

Table II shows the average drug histories for this group for cocaine,
heroin, marijuana, downers, amphetamines, and other drugs. All subjects
Table II. Drug History: Cocaine, Heroin, Other Opiates, Marijuana, Downers,
and Amphetamines (n = 319)
Percent Age first Percent used Age first used

Drug group ever used used regularly regularly
Cocaine 100% 24 ± 7 99% 29 ± 7

Cocaine (intravenous) 25% 27 ± 7 18% 29 ± 7
Heroin (intravenous) 23% 23 ± 7 11% 23 ± 7
Speedball (cocaine and heroin, 12% 28 ± 6 6% 28 ± 7
intravenous)
Other opiates (morphine, 42% 22 ± 6 17% 24 ± 6
opium, codeine, Demerol,
Percodan, etc.)
Marijuana or hashish 98% 16 ± 4 81% 18 ± 4
Downers (reds, rainbows, 56% 20 ± 5 21% 19 ± 5
Quaaludes, etc)
Amphetamines (tablets or 67% 20 ± 5 31% 21 ± 5
pills)/uppers (whites, Dexe-
drine, Ritalin, Preludin)
Hallucinogens (LSD, mes- 65% 20 ± 5 21% 19 ± 4
caline, peyote)
Tranquilizers (Valium, Librium, 35% 24 ± 7 8% 26 ± 8
Miltown, etc.)
PCP (angel dust) 63% 23 ± 7 16% 23 ± 6
Synthetic drugsidesigner drugs 6% 28 ± 7 1% 38 ± 1
(Fentanyl or synthetic H)
Glue, spray cans, gasoline, etc. 25% 14 ± 3 7% 13 ± 3
had used cocaine and the majority were using crack at TE, but only 25% of
them had ever used cocaine intravenously and only 18% had ever done so
regular1y.
While 25% of the subjects had used cocaine intravenously and 23% had
used heroin intravenously, only 12% had ever “speedballed” (injected cocaine
and heroin) and only 6% used this combination intravenously on a regular
basis.
The majority of the subjects had tried marijuana (98%). Of the 81% who
used marijuana regularly, the average age they started regular use was 18
years old (SD = 4). Sixty-seven percent of the subjects had used amphet-
amines and uppers, 31 % had used them regularly. On average, these subjects
started using amphetamines regularly (average age = 21 years old, SD = 5)
before they started regular cocaine use. Fifty-six percent of the subjects had
used downers and 21% had used them regularly. On average, these subjects
also started regular use of downers (average age = 19, SD = 5) before they
started regular cocaine use. Over more than 60% of the subjects had tried
hallucinogens and PCP, but less than one fourth had ever used these drugs
regularly. Less than 20% of all the subjects ever used other drugs regularly.
4.5. Use of Alcohol

Alcohol was consumed by a large proportion of the subjects. Ninety
percent of the subjects reported drinking during their cocaine use careers
(FCU to TE). The average age of first reported alcohol use was 14 years old (SD
= 4). The majority of the subjects (76%) reported first getting drunk between
the ages of 12 and 21. While 38% of the subjects reported that they used no
other drug at the same time as they used cocaine, when subjects reported that
they used at least one other drug at the same time that they used cocaine, 34%
reported that this drug was alcohol. Thirty-seven percent of the subjects re-
ported that they used alcohol after their cocaine use. Fourteen percent report-
ed they used alcohol as a substitute for cocaine. However, when asked wheth-
er their alcohol use was related to their cocaine, 61% of these subjects reported
that, during periods of cocaine use, there was no relation.
Thirty-six percent of sample reported that they had experienced at least
one blackout from drinking, 17% had experienced “the shakes,” and 3% had
had delirium tremens. Thirty-six percent of these subjects reported having
developed a tolerance for alcohol and 37% reported that they had drunk
alcohol to relieve or avoid withdrawal symptoms. Sixty-four percent had
tried, at least once, to cut down or control their alcohol use; only 36% were
successful.
Alcohol also affected these subjects’ lives in other manners. Thirty-two
percent of the sample reported they had lost friends, missed social activities,
or given up hobbies or sports because of alcohol. Thirty-eight percent report-
ed that they continued to use alcohol despite having medical, family, finan-
cial, psychological, or other alcohol-related problems. The majority of the
subjects (74%) reported they drove while drunk and 32% had been arrested
for driving under the influence of alcohol.
4.6. Socially Disruptive or Deviant Behavior during Adolescence

Table III shows the number of subjects who engaged in socially disrup-
tive or deviant behavior before the age of 18, their average ages when they
first and last engaged in these activities, the frequency in which these subjects
engaged in these behaviors, and the time frame of their participating in these
activities in reference to their primary (cocaine) addiction. Note that many of
these variables have large standard deviations (sometimes larger than the
mean), illustrating the higher variability within the sample (see Table III).
These subjects engaged, in at least once before they were 18 years old, in
the following activities: getting suspended and/or expelled from school (56%);
going to school drunk and/or high (50%); and running away from home
(33%). Deviant behaviors include: threatening an adult before age 18 (25%)
and actually hitting an adult before the age of 18 (18%). Thirteen of these
subjects had stolen from school and 11% had damaged school property.
Ninety-seven percent of all the subjects who had been suspended/
expelled from school at least once had done so prior to their cocaine addic-
tion. Only 2% had had this happen to them after their cocaine addiction had
started and only 1% reported having this happen both before and after their
cocaine addiction.
Before the age of 18, as many as 86% of the subjects who had gone to
school drunk and/or high had done so prior to their cocaine addiction. Six
percent had gone to school drunk/high only after their cocaine addiction had
started, while 8% of these subjects had gone to school this way both before
and after the commencement of their cocaine addiction. The first time these
subjects ever went to school drunk/high they were, on average, 16 years old
(SD = 3). Their average age the last time they went to school this way was 18
(SD = 4). Forty-five of the subjects who had gone to school drunk/high report-
ed that they had done so too many times to remember how often, while five
of the subjects estimated they had gone to school this way at least 1000 times.
The other subjects who got drunk and/or high at school reported having done
so an average of 52 times. Even so, the variability of the number of times these
subjects went to school drunk/high was quite large (SD = 108), with the
minimum number of reported times being one time but the maximum being
an estimate of 500 times.
Most of the subjects who had threatened (95%) or had hit an adult (95%)
had done so prior to the beginning of their cocaine addiction. The first time
these subjects ever threatened an adult they were, on average, 15 years old
(SD = 2). They were also, on average, 15 years old (SD = 2) the first time they
ever hit an adult. The last time they threatened to hit an adult they were, on
average, 16 years old (SD = 4), which was also the last time they reported they
had hit an adult when they were under 18 years old. Four subjects reported
292
Table III. Criminal Activity in Cocaine Addicts before the Age of 18
Of the subjects who had

Of the subjects Of the subjects engaged in this criminal
who had who had activity Of all the other
engaged in engaged in subjects who Percentage of the subjects
this criminal this criminal Number of knew or who had engaged in this
activity activity subjects who approximated criminal activity
reported they Number of how often they
Average Average had engaged subjects who engaged in this Both
age when age last in the activity reported they criminal activity before
Criminal activity Percentage first time so many times had engaged Before After and after
prior to the age who ever engaged engaged they couldn't in the activity Number cocaine cocaine cocaine
of 18: whether engaged in in this in this remember 1,000 times of times addiction addiction addiction
arrested or not this activity activity SD activity SD how often or more everdone SD (%) (%) (%)
Expelled/suspended 56% 14 3 15 3 2 0 2 3 97 2 1
from school
Drunk/high in 50% 16 3 18 4 45 5 52 108 86 6 8
school
Ran away 33% 12 3 13 3 1 0 2 3 97 0 0a
Threatened an adult 25% 15 2 16 4 4 0 5 9 95 1 4
(before the age of
18)
Hit an adult (before 18% 15 2 16 4 1 0 4 8 95 2 2a
the age of 18)
Stolen from school 13% 14 3 17 5 1 1 6 17 91 2 7
Damage school 11% 13 3 14 3 1 0 2 2 97 0 3
property
a
Subject(s) did not know if engaged in activity before or after cocaine addiction, or both.
III • Psychology
they had threatened an adult so many times they could not remember how
often. Yet, on average, the other subjects who had threatened an adult had
done so five times (SD = 9). One subject reported that he had hit an adult so
many times he could not remember how many. Still, most of the subjects who
had hit an adult while they were under the age of 18 reported that, on
average, they had done so four times, though the variability of their answers
was a high (SD = 8), with a minimum number of times being one and the
maximum being 50 times.
The majority of subjects who had stolen from (91%) school or damaged
school property (97%) had done so prior to their cocaine addiction. The aver-
age age when subjects first stole from school was 14 years old (SD = 3). Their
average age when they first damaged school property was even younger, 13
years old (SD = 3). While they did stop damaging school property at an earlier
age, 14 years old (SD = 3), the last time these subjects ever stole from school
they were, on average, 17 years old (SD = 5).
4.7. Social Deviance and Violence during Adulthood

A considerable amount of socially deviant behavior was reported by
these subjects in adulthood. Table IV shows the percentage of subjects who
reported engaging in criminal or unlawful activities, their average age when
they first and last committed these acts, the number of times they engaged in
these behaviors, and the percentage of subjects who reported having engaged
in the behaviors either prior to, during, or before and during their cocaine
addiction (see Table IV).
The most frequently reported unlawful activity these subjects engaged in
was driving while under the influence of alcohol and/or drugs (85%). Seven-
ty-two percent of the subjects who reported having driven under the influ-
ence had done so both prior to and during their cocaine addiction; 15%
reported having only driven under the influence before their cocaine addic-
tion; and 13% had only engaged in this behavior after they had become
addicted to cocaine.
Subjects reported that their average age when they first drove under the
influence was 19 years old (SD = 5). Their average age the last time they drove
drunk and/or high was 32 years old (SD = 7). One hundred thirty-seven of
these subjects reported that they had engaged in this behavior so many times
that they could not remember how many; 23 reported they drove under the
influence over 1000 times. Of the other subjects who had driven under the
influence, the average number of times they engaged in this behavior was 69
times, though the variance in responses was still extremely high (SD = 132),
indicating that some of these addicts engaged in this behavior a substantial
number of times.
The pattern of nonviolent deviance was pronounced in this group. Sixty-
five of the subjects reported stealing from stores. However, the majority of
these subjects who reported stealing (74%) had done so before the onset of
294
Table IV. Criminal Activity in Cocaine Addicts
Of the subjects Of the subjects

who had who had Of the subjects who had Of all the other
engaged in engaged in engaged in this criminal subjects who Percentage of the subjects
this criminal this criminal activity knew or who had engaged in this
activity activity approximated criminal activity
Number who how often they
Average Average engaged in engaged in this Both
age when age last the activity Number who criminal activity before
Percentage first time so many times had engaged Before After and after
Criminal activity: who ever engaged engaged they couldn't in the activity Number cocaine cocaine cocaine
whether arrested engaged in in this in this remember 1,000 times of times addiction addiction addiction
or not this activity activity SD activity SD how often or more ever done SD (%) (%) (%)
Driving when 85% 19 5 32 7 137 23 69 132 15 13 72

drunk high
Stolen from storea 65% 12 5 19 10 30 4 18 55 74 6 21
Sold drugs 64% 23 8 31 7 93 20 78 141 16 36 47b
Gambling 61% 18 6 32 7 94 14 69 135 20 14
66
Carried a weapon 50% 20 7 30 8 78 14 57 104 28 23 49
Carried drugs for 47% 24 8 31 8 33 4 34 76 21 38 39b
other
III • Psychology
Damaged other's 28% 16 6 22 9 5 0 6 10 59 11 27 b
propertya
Beat someone se- 28% 20 7 27 7 3 0 6 10 43 12 32 b
verely
Threatened with a 27% 23 8 28 9 9 1 7 15 34 38 27 b
weapon
Broken into a 23% 17 6 22 9 2 0 8 21 68 14 17b
buildinga
Broken into a car 19% 18 6 21 7 5 0 6 9 74 18 8
Stolen a car 18% 19 6 20 6 1 0 8 34 63 23 13 b
Prostitution/ 13% 24 6 29 7 11 4 98 176 40 33 28
pimpinga
12 • Alcohol–Cocaine Interaction
Used force for 12% 23 8 27 8 7 0 32 93 32 39 26 b

profit
Shot someonea 11% 23 6 26 7 0 0 2 3 49 31 20
Threatened for 8% 22 6 27 6 3 0 39 106 22 44 30 b
profit
Used weapon for 6% 23 8 29 7 2 0 18 26 15 50 20b
profita
Committed forcible 2% 23 7 24 7 0 0 1 0.4 33 67 0
rape
a
Subject(s) (1 to 3 subjects only) refused to answer whether engaged in activity at all.
b
Subject(s) did not know if engaged in activity before or after cocaine addiction, or both.
295
cocaine use. Sixty-one percent of the subjects also reported engaging in gam-
bling activities. This behavior was reported most frequently to have occurred
both prior to and during their cocaine addiction (66% of these subjects).
Episodes of violent behavior and acknowledgment of carrying weapons
were common, being reported by about 50% of the sample. Fifty percent of
subjects reported carrying guns, knives, and blunt objects. Twenty-seven
percent reported threatening others with a weapon and 8% reported threat-
ening others for profit. Serious acts of violence were frequently mentioned:
for example, beating another person severely was reported by 28% of the
sample, shooting someone by 11%, and forcible rape by 2%.
A significant proportion of subjects reported participation in the illicit
drug supply system. Sixty-four percent reported selling drugs and 47% car-
ried drugs for others. Sixteen percent of the subjects who had sold drugs and
21% who had carried drugs for others did so before initiation of cocaine use.
These activities increased after the use of cocaine became habitual; 36% of
subjects sold drugs and 38% carried drugs for others only after the onset of
their cocaine addiction.
4.8. Interactions of Cocaine and Alcohol Use

Subjects reported their alcohol and cocaine use in a total of 3608 (aggre-
gate for the sample) monthly segments of their cocaine use history (FCU to
TE). In 52% of these monthly periods, subjects used both alcohol and cocaine;
in 27% of the segments, they used alcohol only; in 12%, they used cocaine
only; and in 9% of the monthly periods, they used neither alcohol nor co-
caine.
Table V shows the number of monthly episodes for each type use and
percentage of monthly periods the subjects were engaged in drug dealing or
other criminal activity. There was a statistically significant relationship be-
tween whether subjects were using cocaine, alcohol, both, or neither and
whether they were engaged in criminal activity (x2 = 25.196, df = 3, P <
0.001). When subjects used cocaine and alcohol in combination or just co-
caine, they were more likely to engage in criminal activity (10% of those
Table V. Alcohol and Cocaine Use by Crime and Drug Dealing
Percent of monthly Percent of monthly

Monthly segments where segments engaged in segments engaged
patients were using: Frequency Percent criminal activitya in drug dealing
Both alcohol and cocaine 1880 52 10 24

Cocaine only 439 12 10 21
Alcohol only 979 27 6 15
Neither alcohol nor cocaine 310 9 3 9
Total monthly segments 3608 100 9 20
Statistically significant relationship at P < 0.001; x2 test.
a
monthly episodes) than when they were using neither cocaine or alcohol (3%)
or when they were using alcohol only (6%). Also, when subjects were using
both cocaine and alcohol (24%), or when they were using cocaine but not
alcohol (21%), they were more likely to engage in drug dealing than when
they were abstinent from both cocaine and alcohol (9%) or just using alcohol
(15%), (x2 = 63.695, df = 3, P < 0.001). Please note that amount used was not
taken into consideration for this analyses in this section.
4.9. Violence in Cocaine Addicts according to Amount of Alcohol Used

Amount of alcohol used was calculated from the number and types of
alcoholic beverages subjects reported they used per week in each monthly
period. The number and types of alcoholic beverages were converted into
number of ounces of alcohol absolute ethanol (100% pure alcohol) for each
monthly period. Thus, the amount of alcohol variable reports the average
ounces of alcohol consumed per month by subjects from first cocaine use to
their treatment entry. These variables were obtained for 314 subjects. Con-
suming 120 ounces of alcohol per month or more (or an average of 4 ounces or
more per day) was considered excessive alcohol use or excessive drinking.25
Only 20% of the subjects were classified as excessive drinkers using this
method, with the highest reported average amount of alcohol consumed over
their cocaine use career being 207 ounces per week. Twenty-five subjects
reported they did not consume any alcohol from their first cocaine use to their
treatment entry.
Table VI shows criminal activities these subjects engaged in according to
their alcohol use and the percentage of those who engaged in the behavior
prior to, during, or both prior to and during their cocaine addiction. There
was a statistically significant relationship between alcohol use and whether
subjects had ever beaten someone severely (x2 = 5.370, df = 1, P = 0.020).
Forty percent of all subjects, who reported on average having consumed more
than 28 ounces of alcohol per week, reported having beaten someone se-
verely, while only 26% of subjects, who were not classified as engaging in
excessive alcohol use over their cocaine use careers, reported having beaten
someone severely. However, there was no statistically significant relationship
between when, in the subjects’ cocaine-using careers (only before, or after, or
both before and after the onset of their cocaine addiction), they primarily
engaged in this behavior (x2 = 0.015, df = 2, P = 0.992).
There was also a statistically significant relationship between alcohol use
and whether subjects ever threatened someone with a weapon (x2 = 5.898, df
= 1, P = 0.015). Thirty-nine percent of subjects, who reported on average
having consumed more than 28 ounces of alcohol per week, reported having
threatened someone with a weapon, while only 24% of subjects, who were
not classified as engaging in excessive alcohol use over their cocaine use
careers, reported having done so. Again, there was no statistically significant
association between alcohol use and when in the subjects’ cocaine using
Table VI. Violence in Cocaine Addicts by Alcohol Use (n = 313)
Percentage of those who engaged in

excessive alcohol use or not and had
Percentage who engaged in this activity
reported ever
Excessive having engaged Both before
Criminal activity: alcohol use in this activity Before After and after
whether arrested (≥28 ounces by excessive cocaine cocaine cocaine
or not per week)a alcohol use addiction addiction addiction
Carried a weapon Yes 56 23 26 51

No 48 29 23 48
Beat someone se- Yes 40 44 24 32
verelyb No 26 43 24 33
Threatened with a Yes 39 25 42 33
weaponb No 24 38 38 24
Damaged other’s Yes 37 45 18 36
property No 27 65 9 26
Used force for Yes 15 11 78 11
profit No 12 39 29 32
Shot someone Yes 13 38 25 38
No 11 50 35 15
Threatened for Yes 6 25 50 25
profit No 9 23 45 32
Used weapon for Yes 6 25 75 0
profit No 6 15 54 31
Committed forcible Yes 2 0 100 0
rape No 2 40 60 0
Average ounces of alcohol consumed per week over cocaine use history (FCU to TE)
a
Statistically significant relationship between alcohol use and whether engaged in this violent behavior at P <
b
0.05, x2 test.
careers they primarily engaged in this behavior (x2 = 1.427, df = 2, P = 0.490).

Under this classification of alcohol use, 50% (3 subjects) of the subjects who
had committed forcible rape had engaged in excessive alcohol use; yet, all of
these subjects who had committed rape had done so after the onset of their
cocaine addiction.
Small sample sizes for each of the violent behaviors resulted in unreliable
statistical comparisons in terms of alcohol use and when in the subjects’
cocaine-using careers (only before, or after, or both before and after cocaine
onset) they primarily engaged in each violent behavior; neither were any
other statistically significant relationships found between each of the reported
violent behaviors individually and the subjects’ alcohol use.
5. Summary and Discussion

This chapter presents (1) a review of work done on the issue of cocaine-
alcohol use and violence and (2) a longitudinal view of deviant and violent
behaviors that occurred before and during the cocaine use careers of a group
of cocaine-dependent veterans who sought treatment for their dependence.
Alcohol, by itself, has been found to have a calming effect physically but a
disruptive effect psychologically. Inebriation, under high levels of personal
threat, appears to focus attention on negative cues and lead to more intense
aggression.5,6 Cocaine, particularly crack cocaine, has also been associated
with both an increase in the seriousness and rate of criminal activity,18 with
significant increases occurring after the onset of cocaine abuse. Few re-
searchers have looked at the subpopulation of cocaine addicts who also abuse
alcohol19,20 and have found that cocaine–alcohol abusers are more likely to
have higher rates of criminal behavior than both cocaine-only abusers and
alcohol-only abusers.21
The study reported here looked further into the population of cocaine
addicts to examine possible differences in levels and types of criminal, partic-
ularly violent, behaviors between cocaine addicts who abuse and those who
did not abuse alcohol. It also looked at the differences in criminal activity for
cocaine addicts when they were using both cocaine and alcohol, when they
were using cocaine only or alcohol only, or when they were abstinent from
both cocaine and alcohol. In addition, comparisons were made in order to try
to distinguish between behavior these addicts engaged in prior to their co-
caine addiction and during their cocaine use careers.
The sample in this study consisted of adult male cocaine addicts, average
36 years old, which is slightly older than the age of clinical populations of
cocaine addicts examined by other studies. This age difference is most likely
due to the study’s prerequisite that subjects be veterans of the military ser-
vice. African Americans were overrepresented in this study, which is consis-
tent with other urban studies of cocaine users. While this could represent a
sampling bias, there has been no evidence of ethnicity being a major influence
on the variables studied.
These subjects also had an average level of education higher than that of
the general population or the population of veterans, which is consistent with
the findings of other investigators who have reported above-average levels of
education in cocaine-using populations.26–30 The average age at first cocaine
use was 24 years old, which is also consistent with epidemiologic investiga-
tions that have reported that the period of high risk for cocaine experimen-
tation is between the ages of 18 and 24.31 Still, the variance in age of FCU
was considerably high, with the age at first use ranging between 12 and
58 years old.
The average length of these subjects’ cocaine careers from FCU to TE
spanned 10 years. Most subjects reported that during the majority of this time
they were also drinking alcohol, and 20% of these subjects were classified as
excessive alcohol users when their alcohol use was averaged over the period
of their cocaine careers. When subjects reported using another drug with
cocaine, before or after their cocaine use, or as a substitute for cocaine,
the most frequently reported drug was alcohol. These subjects’ alcohol use
started an average of 10 years prior to their FCU.
In addition to sociodemographic and drug use data, we examined pat-

terns of behavior occurring during the cocaine-use career. Retrospective infor-
mation covered a time span from FCU to the point at which the addiction
became sufficiently severe to lead to help-seeking behavior and treatment for
cocaine dependence. Most subjects began using cocaine intranasally, all of
them progressed to smoking crack at some point in their cocaine career, but at
the time of TE, 64% were smoking crack as their preferred route of administra-
tion. Most also had used alcohol and had been exposed to other drugs of
abuse, particularly marijuana.
A subset of the sample exhibited socially deviant behavior, including
violence. Deviant behaviors were present during early youth and adulthood,
both before the onset of cocaine use as well as during the cocaine-use career.
While the proportion of individuals who engaged in particular behaviors was
relatively small, the frequency of the behaviors was large.
The extent of deviant or violent behavior during periods of cocaine use,
periods of combined cocaine–alcohol use, periods of alcohol use, and periods
of sobriety were described. These observations on cocaine-dependent indi-
viduals illustrate the complexity of the relationships between cocaine use,
combined alcohol and cocaine use, and violent behavior. It should be noted
that only a small segment of this cocaine-dependent sample engaged in vio-
lent behavior. Cocaine use did not appear to bring out tendencies that were
not already existent, based on the fact that in most instances subjects who
engaged in criminal behaviors were doing so prior to their cocaine use onset.
Use of alcohol with cocaine seemed to increase the risk of a cocaine user
engaging in violent behavior. The conclusions drawn from our results are
preliminary; thus, further systematic inquiry of these issues is necessary.
Of all the delinquent behaviors these addicts participated in as juveniles,
their variance in the frequency of participation remained high. While the
majority of the subjects who had acted aggressively or violently as teenagers
had done so prior to the onset of their cocaine addiction, the ages when they
threatened an adult or hit an adult fall within the same time period as the
initiation of their alcohol use. Further investigation within this population
would need to be conducted in order to ascertain whether alcohol use during
this time is correlated with these behaviors.
The nature of criminal activities in adulthood also changed as these sub-
jects progressed in their cocaine addiction. The majority of the subjects who
had ever broken into a building or car, stolen from a store, stolen a car, or
damaged others’ property had done so prior to the onset of their cocaine
addiction. Those who had either beaten someone severely or shot someone
had more often done so either prior to their cocaine addiction or both before
and after the onset of their addiction. Driving while under the influence,
gambling, carrying a weapon, and selling or carrying drugs for others were
most often done both prior to and after the onset of their cocaine addiction.
More violent crimes (such as threatening with a weapon or threatening for
profit, using a weapon or force for profit, and committing rape), though
reported by fewer subjects, were more frequently reported to have been com-
mitted after the onset of cocaine addiction.
As was expected, subjects reported having engaged in more criminal
activities during periods of their cocaine career when they were also drinking
alcohol. Subjects reported engaging in fewer criminal activities when they
were drinking alcohol and abstinent from cocaine. When these addicts were
abstinent from both alcohol and cocaine, they were least likely to engage in
criminal activities.
In addition, those who had used alcohol to excess (an average of at least 4
ounces of pure alcohol per day) during their cocaine career were more likely
to have engaged in serious violent behavior (beating someone severely and
threatening with a weapon) than whose who had not used alcohol to excess.
As would be expected, there was no difference in aggressive behaviors en-
gaged in for profit by those who had used alcohol to excess and those who
had not. There were also no significant differences between the number of
cocaine addicts who drank alcohol to excess and those who did not in terms of
shooting someone or committing rape; yet, all the subjects who drank alcohol
excessively during their cocaine careers who had committed rape had done so
only after the onset of their cocaine addiction, in comparison to 40% of those
who did not drink excessively and had committed rape prior to the onset of
their cocaine addiction.
6. Comments and Research Implications
These observations on deviant behavior among cocaine-dependent indi-

viduals, most of whom also drank alcohol during some point of their cocaine-
using careers, illustrate the complexity of the relationships between cocaine
use, alcohol use, and violent behaviors. Our findings indicate that only a
small proportion of our cocaine-dependent sample engaged in violent behav-
ior; cocaine use did not appear to induce behaviors that had not been engaged
in previously. On the other hand, use of alcohol seemed to increase the
likelihood of the cocaine user engaging in violent behavior. These conclusions
drawn from our findings are preliminary; thus, further systematic inquiry of
these issues is necessary and several research topics require much more ex-
plicit attention.
Further analyses need to be conducted with our data. For example, to
distinguish between different levels of alcohol use during different periods in
the cocaine career may provide better information on the dynamic links be-
tween both addictions and crime.
Further analyses are planned distinguishing between specific criminal
activities subjects engaged in at different phases of the cocaine career (during
the year prior to first cocaine use, between first cocaine use and first severe
cocaine use, and between first severe cocaine use and treatment entry) in
relation to the subjects’ alcohol use during these time periods. These and
other analyses will hopefully provide a better understanding of the interac-

tions among drug-related behaviors and violence.
References
1. McGonigal MD, Cole J, Schwab CW, et al: Urban firearm deaths: A five year perspective.
J Trauma 35:532–536, 1993.
2. Linnoila M, Virkkunen M, Scheinin M, et al: Low cerebrospinal 5-hydroxyindoleacetic acid
concentration differentiates impulsive from impulsive violent behavior. Life Sci 33:2609–2614,
1983.
3. Bushman BJ, Cooper HM: Effect of alcohol on human aggression: An integrative research
4. Moss HB, Krisci L: Aggressively in adolescent alcohol abusers: Relationship with conduct
disorder. Alcohol Clin Exp Res 19:642–646, 1995.
5. Zeichner A, Allen JB, Gianola PR, Lating JM: Alcohol and aggression: Effects of personal
threat on human aggression and affective arousal. Alcohol Clin Exp Res 18:657–663, 1994.
6. Zeichner A, Giancola PR, Allen JD: Effects of hostility on alcohol stress-response dampening.
Alcohol Clin Exp Res 19:977–983, 1995.
7. Goldstein PJ: The drug/violence nexus; A tripartite conceptual framework. J Drug Issues
15(4):493–506, 1985.
8. Goldstein J: Cocaine and crime in the United States. Presentation for the United Nations
Interregional Criminological Research Institute International Symposium on Cocaine March
1991, Rome, Italy.
9. Miller NS, Gold MN: Criminal activity and crack addiction. Int J Addict 29:1069–1978, 1994.
10. Giannini AJ, Miller NS, Loiselle RH, Turner CE: Cocaine-associated violence and relationship
to rout of administration. J Subst Abuse Treat 10:67–69, 1993.
11. Stanton B , Galbraith J: Drug trafficking among African American early adolescents: Preva-
lence, consequences, and associate behaviors and beliefs. Pediatrics 93:1039–1043, 1994.
12. Black MM, Richardo IB: Drug use-drug trafficking and weapon carrying among low-income
African-American adolescent boys. Pediatrics 93:1065–1072, 1994.
13. Kang SY, Magura S, Shapiro JL: Correlates of cocaine/crack use among inner city-incarcerated
adolescents. Am J Drug Alcohol Abuse 20:413–429, 1994.
14. Sheley JF: Drug and guns among inner-city high school students. J Drug Educ 24:301–321,
1994.
15. Hutson HR, Anglin D, Kyraicou D, et al: The epidemic of gang related homicides in Los
Angeles county from 1979 through 1994. J Am Med Assoc 274(13):1031–1036, 1995.
16. Klein MH, Maxson CL, Cunningham LC: “Crack” street gangs and violence. Criminology
29:701–727, 1991.
17. Meehan PJ, O’Carrol PW: Gangs drugs and homicide in Los Angeles. Am J Dis Child 146:683–
687, 1992.
18. Belenko S, Chin K, Fagan J: Typologies of criminal career among crack arrestees. Paper
presented at the Annual Meeting of the American Society of Criminology, Reno, NV, 1989.
19. Khalsa ME, Paredes A, Anglin MD: Cocaine dependence behavioral dimensions and patterns
of progression. Am J Addict 2:330–345, 1993.
20. Higgins S, Budney AJ, Bickel WK, et al: Alcohol dependence and simultaneous cocaine and
alcohol use in cocaine-dependent subjects. J Addict Dis 13:177–189, 1994.
21. Windle M, Miller-Tutzauer C: Antecedents and correlates of alcohol, cocaine, and alcohol-
cocaine abuse in early adulthood. J Drug Educ 21:133–148, 1991.
22. Nurco DN, Bonito AJ, Lerner M, Bailer M: Studying addicts over time: Methodology and
preliminary findings. Am J Alcohol Abuse 2:107–121, 1975.
23. Bradburn NM, Rips LJ, Shevell SK: Answering autobiographical questions: Impacts of memo-
ry and inference on surveys. Science 236(4798):157–161, 1987.
24. Khalsa ME, Anglin MD, Paredes A: Pre-treatment natural history of cocaine addiction. NIDA
Res Monogr 105:494–500, 1991.
25. Barbor TF, Kvanzaler HR, Lauerman RJ: Social drinking as a health and psychosocial risk
factor: Anstie’s limit revisited, in Galanter M (ed): Recent Developments on Alcoholism New
York, Plenum Press, 1987, pp 373–397.
26. Gawin FH: Cocaine addiction: Psychology and neurophysiology. Science 251(5001):1580–1586,
1991.
27. Gawin FH, Kleber HD: Cocaine use in a treatment population: Patterns and diagnostic
distinctions. NIDA Res Monogr 61:182–192, 1985.
28. Gawin FH, Kleber HD: Abstinence symptomatology and psychiatric diagnosis in cocaine
abusers. Arch Gen Psychiatry 43:107–113, 1986.
29. Schnoll SH, Carrigan S, Kitchen S, et al: Characteristics of cocaine abusers presenting for
treatment, in Kozel NJ, Adams EH (eds): Cocaine Use in America: Epidemiology and Clinical
Perspectives. NIDA Research Monograph 61. Rockville, MD, US Department of Health and
Human Resources, 1988, pp 171–181.
30. Washton AM, Gold MS, Pottash AC, et al: Treatment outcome in cocaine abusers, in Harris LS
(ed): Problems of Drug Dependence. NIDA Research Monograph 67. Rockville, MD, US Depart-
ment of Health and Human Services, 1986, pp 381–384.
31. Anthony JC, Ritter CJ, Von Kerfett MR, et al: Descriptive epidemiology of adult cocaine use in
four US communities, in Harris LS (ed): Problems of Drug Dependence. NIDA Research Mono-
graph 67. Rockville, MD, US Department of Health and Human Services, 1985, pp 283–289.
IV
Family Issues
Edward Gottheil and Ellen F. Gottheil, Section Editors

Overview
Edward Gottheil and Ellen F. Gottheil
According to a recent bulletin of the Bureau of Justice Statistics, the number of

prisoners in the custody of state correctional authorities for a variety of seri-
ous offenses increased from 1980 to 1993 by a factor of 2.8.1 Of these, the
largest increases, by a wide margin, were drug offenses (increasing by a factor
of 9.79) and sexual assaults other than rape (by a factor of 6.68). In today’s
world, drug abuse and partner abuse have become increasingly significant
public health issues.
Widely cited books by Pizzey,2 Gelles,3 and Steinmetz and Straus,4 pub-
lished only 22 years ago in 1974, markedly increased our attention on and
concern about spouse abuse and helped promote an interest in research into
the causes of marital violence. For the most part, the victims, i.e., the battered
wives and children, were the ones initially studied during the 1970s and
1980s. Increasing realization that stopping or preventing these behaviors re-
quired involvement of the abusers gradually led to research and clinical pro-
grams for batterers in the 1980s.5,6
In 1983, Spieker7 pointed out that while it was “common knowledge”
that violence and alcohol “coexist and are exhibited” in the “all-American
home,” there were few studies available correlating family violence and alco-
hol abuse. Indeed, leafing through the articles of an issue of Violence and
Victims recently, I found that only 2 of 150 cited references were to journals in
the substance abuse field. Unfortunately, “substance abuse” and “domestic
violence” still tend to remain separate conceptual entities in theory, research,
and treatment. Partly this is understandable. Systematic studies of substance
abuse, aggressive behavior, and family relationships are rather recent. Each
Edward Gottheil • Department of Psychiatry and Human Behavior, Thomas Jefferson Univer-
sity, Philadelphia, Pennsylvania 19107. Ellen F. Gottheil • Department of Psychiatry and
Behavioral Sciences, University of Washington Medical School, Seattle, Washington 98195.
307
308 IV • Family Issues
has presented problems in definition, few practicable experimental models,

and difficulty in applying controls in either community or clinical situations.
Studies of interactions between substance abuse and aggression within fami-
lies are even more recent and more difficult to execute.
In addition to bringing together materials from different disciplines i.e. ,
biological, psychological, and social studies of aggressive behavior, drug use,
and family interactions, with their different methodologies and terminolo-
gies, there is the need to consider the many different specific parameters
within particular studies. Some of these, to note only a few examples, would
include: (1) type, level, and duration of alcohol or other drug (AOD) intake,
e.g., cocaine binges or chronic inebriation; (2) type, severity, and duration of
domestic violence (DV) involving spouse and/or child abuse; (3) source of
data from community sample, treatment group, jail, etc,; (4) relationship
between perpetrator AOD intake and DV; (5) relationships among victim’s
AOD use, perpetrator’s AOD use and DV; (6) personality factors, antisocial
behavior, neurotic conflicts; and (7) history of AOD and/or DV in family of
origin of the perpetrator and of the victim.
Clearly the design of relevant research, the interpretation of findings,
and the integration of the data across studies are not simple tasks. A question
we face in our attempt to study and treat very troubled families is: How do we
appreciate the complexity without becoming overwhelmed and ineffective?
One way to do this is to organize problems into unitary areas. In situations
where factors coexist, however, we may be more effective if we can keep in
mind and relate more than one model. As in all cases of dual diagnosis, this
requires that we as researchers and clinicians maintain flexibility without
losing our energy and direction. Undaunted, the authors of the following four
chapters have attempted to bring together and summarize information avail-
able on selected topics relating to family violence from the perspective of the
spouse, the batterer, the child, and treatment.
Theories of spouse abuse and theories of alcohol-induced aggression are
reviewed in Chapter 13 by Kaufman Kantor and Asdigian in an attempt to
understand the linkage between them as well as the possible contribution to
the linkage, if any, of the victim’s AOD use. In general, the theories have
focused on male behavior and intoxication-aggression effects. There has been
great sensitivity in the DV field to the problem of blaming the victim, and
perhaps there has been some reluctance to study possible relationships be-
tween the victim’s AOD use and her victimization out of concern that she will
be blamed for “causing” her abuse. There have been suggestions, however,
that intoxication–victimization effects also occur. The chapter then addresses:
(1) hypothesized dimensions and mechanisms of intoxication–victimization
effects; (2) whether the literature indicates that women who drink or use
drugs are at greater risk for spouse abuse; (3) variation according to popula-
tion sampled or type of victimization; and (4) whether such effects are medi-
ated or confounded by other variables.
If intoxication–victimization as well as intoxication–aggression effects are
IV • Overview 309
to be considered, theoretical explanations of the relationship between AOD

and violence will need to take into account and integrate findings about
differences between men and women in their AOD effects and expectations.
Actually, Kaufman Kantor and Asdigian found that the “empirical evidence
supporting an association between women’s intoxication and physical as-
saults by husbands” was mixed according to whether alcohol or other drugs
were involved, whether the assaults were more severe or less severe, whether
use was at the time of the assault or general usage, and whether the research
was conducted with clinical samples of heavy drinking women and/or bat-
tered women.
In summing up, they suggest that while AOD use by women does, for
the most part, seem to be associated with victimization, this may not be
because the victim’s intoxication is an immediate behavioral precedent to her
victimization, but may be due to a number of possible underlying reasons.
Among these could be the lower rate of self-esteem, decreased resources, and
level of functioning of AOD-addicted women, making it more difficult to deal
effectively with violence at home. Another could be a history of victimization
in the woman’s family of origin. Family violence is known to be an important
factor in the history of male batterers, and assaulted women were reported to
have about double the rate of growing up in a violent home as nonassaulted
women. There is also a greater likelihood of having a heavy-drinking or drug-
abusing partner. Finally, the occurrence of substance abuse may be a conse-
quence of victimization rather than an antecedent.
The male who beats his wife and batters his children has long been
regarded more as a repulsive criminal than as an interesting subject for study
or a person in need of care and support. This view has not been especially
conducive to the development of research or therapeutic programs for the
batterer. Nevertheless, if we are to intervene and hope to prevent recurrences
and ultimately develop preventive programs, we need to study, treat, and
better understand spouse and child abusers. In their review of the literature,
Lee and Weinstein, in Chapter 14, report that in recent years (over 80% of the
references they cite were dated 1983 or later) a number of treatment programs
for batterers have been initiated and a considerable amount of research has
been done. In the first part of their chapter they examine research studies that
attempt to identify the causes of partner abuse and note a number of interest-
ing ideas and findings, but they conclude that the causes are varied and that
men abuse their partners for many different reasons. A major part of their
chapter is devoted to attempts that have been made to define unique charac-
teristics of male batterers. Rather than a definition or typical profile, however,
what emerges is a wide variety of background, behavioral, psychological, and
social characteristics of the batterer, as well as several promising typologies,
which taken together strongly suggest that batterers are heterogeneous. The
importance of this concept of heterogeneity8 is that it then behooves us to
search for causes and treatment approaches rather than the cause of or the
treatment for the batterer.
Lee and Weinstein describe many studies that indicate that AOD abuse is
commonly found in batterers. Although neither alcohol or drug abuse are
necessary or sufficient for the occurrence of domestic violence, they are often
present. As more studies are being done, they are becoming increasingly
more sophisticated, with some now seeking to differentiate the characteristics
of spouse abusers from those of substance abusers by controlling for one or
the other; for example, they describe a Jefferson Medical College study in
which the characteristics of cocaine users who do and who do not batter are
compared.
Regarding treatment, a variety of approaches have been employed, three
of which were selected for discussion: divorce mediation, group treatment for
batterers focusing on shame as a core issue and on confessions, and manda-
tory arrests. These were selected and their potential advantages and disad-
vantages described and evaluated in some detail because of their unique
relationship to battering treatment. Regardless of treatment approach, how-
ever, a number of studies were reviewed indicating that alcohol and drug
abuse were important contributors to recidivism in treatment for battering.
In Chapter 15, Miller, Maguin, and Downs provide us with a comprehen-
sive review of the current state of knowledge regarding (1) the extent to which
AOD use by the perpetrator is related to the perpetration of physical or sexual
abuse against children and (2) the relationship between experiencing physical
or sexual abuse during childhood and the subsequent development of AOD
use/abuse. It may be worth noting that when the authors compare differences
in findings between earlier and more recent reviews, the more recent reviews
are more sophisticated in design and the earlier ones, for the most part, were
done only 10 to 15 years ago. Clearly, a great deal has been accomplished
during the last decade, yielding the many new findings reported in this
review.
In studying relationships between AOD use and family violence, addi-
tional methodological considerations are involved when the child is the vic-
tim. For example, if the validity of the information obtained by asking pa-
tients coming to a substance abuse clinic about their substance abuse and
aggressive behavior has been questioned, how questionable is the validity of
the information obtained from parents when asked, no matter how gently
and how indirectly, about their use/abuse of substances in relation to their
physical and sexual abuse of their children? If, instead, retrospective reports
are obtained regarding parental AOD use and abusive behaviors from their
grown-up children, they may be overreported on the basis of false memories
or fabrications, or underreported due to repression, suppression, or a simple
unwillingness to reveal the information (a study is described in which one
third of officially documented cases of childhood sexual abuse were not re-
ported when the victims were later interviewed as adults). Furthermore, re-
garding sexual abuse, the victims may not know about the AOD usage of the
perpetrators, since the majority of them are not parents but strangers.
Results of studies of the effects of experiencing childhood physical
IV • Overview 311
and/or sexual abuse are not always consistent and vary according to the type,
severity, duration, and context of the abuse, whether the victim is male or
female, and whether the perpetrator is of the same or opposite sex. In Chap-
ter 15, several hypotheses are described by Miller and coauthors regarding
the accessibility of perpetrators to the children, including whether AOD ad-
diction decreases parental ability to protect children, framing this important
topic for further research. Taking into account the complexity of the issues
and methodological limitations, the authors conclude that there is suitable
evidence indicating that parental AOD is related to physical and sexual abuse
of their children, that parental AOD problems increase the likelihood of their
children being victimized by others, and that childhood victimization is re-
lated to the later development of AOD. Nevertheless, there are many gaps
and inconsistencies in the theoretical explanations and our state of knowl-
edge, and Miller and coauthors end their chapter with a discussion of theoret-
ical and methodological considerations, special ethical and legal issues (e.g.,
the duty to warn authorities about findings of abuse as well as the need to
warn subjects about this duty), and clearly specified directions for future
research.
In Chapter 16, the desirability and practicability of linking services for the
treatment of alcoholism and domestic violence are addressed by Collins and
associates. Reviewing the role of alcoholism in domestic violence, they find
consistent evidence in the literature of a strong association between hus-
band’s drinking and spouse abuse. A number of studies also indicate that
drinking on the part of the woman may increase the likelihood of being
victimized, and that in some instances drinking by the abused spouse follows
the abuse and perpetuates the cycle. It seems reasonable to expect, then, that
for men who are batterers and also spouse abusers, for women who have
been battered and are substance abusers, and for their families, treatment
efforts would be enhanced by attending to both the substance abuse and the
spouse abuse.
Nevertheless, they report that although staff of substance abuse and
domestic violence programs when asked endorse the idea of treating both
problems, very few programs actually do so. Few of the programs screen for
the other abuse condition; even when the coexistence of these problems is
recognized, the most common form of linkage is referral to another agency,
which is accomplished only some of the time and often without ongoing
provider interaction and cooperation.
Domestic violence programs are concerned, of course, about the safety
and interests of the battered women and providing them with shelter, child
care services, legal assistance, and protection from the batterer. For the bat-
terer, they are most interested in programming to emphasize behavioral man-
agement, anger control, acceptance of personal responsibility, and meeting
family responsibilities. The concept of alcoholism as a disease may be seen as
an excuse for not accepting responsibility, frequent attendance at Alcoholic
Anonymous (AA) meetings as an avoidance of assuming family respon-
sibilities, and ideas regarding enabling and codependency as a means of

shifting blame to the victims. Indeed, staff of substance abuse programs often
view the addiction as the disease and as the primary problem, deemphasize
personal accountability for past behavior during the “active” disease, encour-
age frequent AA attendance and sponsorship, and often consider the spouse
abuse as secondary to the addiction.
Although Collins and his coauthors describe clear advantages to provid-
ing integrated services to the multiproblem families involved in domestic
violence, there are impediments to developing linkages. These include differ-
ent organizational placements, separate funding streams, and lack of re-
sources, but the most important and difficult to deal with perhaps are the
differences in philosophy, goals, training, and management concerns of the
treatment staffs. AOD staff may assume that battering will decrease as addic-
tion is treated, despite evidence that this will not happen unless the battering
is directly addressed. DV staff may not recognize the extent to which victim’s
substance use or abuse makes it harder for her to deal with or leave her
relationship, or the extent to which substance abuse may exacerbate the hus-
bands battering .
The authors suggest that, “It would be premature to recommend imple-
mentation of widespread attempts at linking substance abuse and domestic
violence services in the absence of well-developed and effective models for
doing so.” They complete the chapter by describing some examples of current
linkages, the advantages of linking services, and different models for linkages
and they outline research and clinical steps to be taken to further develop a
better integration of effective services.
Reading these four lucidly written, comprehensive, well-referenced re-
view chapters on substance abuse and domestic violence, one cannot help
being impressed by the complexity of the conceptual and methodological
issues confronting those wishing to study these conditions and their interac-
tions. Much, of course, remains to be learned about aggressive behavior,
substance abuse, and family relationships before one can hope to really un-
derstand how their interactions may culminate in abuse of the spouse or
child. At the same time, one is also impressed by the clear recognition of
methodological problems and limitations, the increasing sophistication of the
studies, and the progress that has been made with respect to new findings,
research techniques, and theoretical development. The authors of these chap-
ters have made a significant contribution toward bringing together informa-
tion and theory about aggressive behaviors, AOD, and domestic violence,
which should serve to further the process of cross-fertilization.
Nevertheless, a continuing problem is insufficient communication across
disciplines. There is much relevant information within the journals and ex-
pressed in the esoteric verbiage of the separate disciplines. This lack of inte-
gration is to the detriment of the disciplines as well as to our patients. For, if
we treat only the problem with which we are familiar and ignore associated
problems, we may unwittingly sow the seeds for relapse of the very problem
IV • Overview 313
we are trying to treat. While Collins and his coauthors wisely caution against
leaping full-scale into linking AOD and DV services in the absence of well-
developed effective models for doing so, it is clear that the development of
these models is urgently needed to address these problems when they are
intertwined. In addition to having caseworkers refer, manage, and integrate
services, we might add that it is not too early to begin educating clinicians and
researchers in training to be sensitive to other kinds of abuse than the one on
which they are specifically focused, to introduce them to complexity, to ex-
pose them to additional theoretical models, and to familiarize them with
chapters like the four to follow in order to keep our patients from falling
through the large cracks between disciplines.
References
1. Beck AJ, Gilliard DK: Prisoners in 1994. Bureau of Justice Statistics Bulletin, August 1995, p 11.
2. Pizzey E: Scream Quietly or the Neighbors Will Hear. London, Penguin, 1974.
3. Gelles RJ: The Violent Home. Beverly Hills, CA, Sage, 1974.
4. Steinmetz SK, Straus M: Violence in the Family. New York, Harper & Row, 1974.
5. Eddy MJ, Myers T: Helping Men Who Batter: A Profile of Programs in the US. Arlington, TX, Texas
State Department of Human Resources, 1984.
6. Roberts AR: A national survey of services for batterers, in Roy M (ed): The Abusive Partner: An
Analysis of Domestic Battering. New York, Van Nostrand-Reinhold, 1982, pp 17–35.
7. Spieker G: What is the linkage between alcohol abuse and violence, in Gottheil E, Druley KA,
Skoloda TE, Waxman HM (eds): Alcohol, Drug Abuse and Aggression. Springfield, IL, Charles C.
Thomas, 1983, pp 125–136.
8. Institute of Medicine: Broadening the Base of Treatment for Alcohol Problems. Washington, DC,
National Academy Press, 1990.
13
When Women Are

under the Influence
Does Drinking or Drug Use by Women
Provoke Beatings by Men?
Glenda Kaufman Kantor and Nancy Asdigian
Abstract. This chapter examines theoretical and empirical evidence for the existence of “intox-
ication–victimization” effects. Theories of victimization and theories of alcohol-induced ag-
gression are examined for their relevance to the phenomenon of concern. The results of our
examination of theory and research indicates theoretical support for an integrated theory of
intoxication–victimization effects. However, we found that the temporal precedence of women’s
drinking related to their victimizations has not been established by prior investigations. Support-
ing evidence for intoxication–victimization effects is strongest among studies of rape, homicide,
and studies of alcoholic women. Despite the strength of alcohol’s association with wife assaults,
intoxication’s centrality and temporal relationship to specific wife-assault episodes is highly
variable. Our review indicates that women’s intoxication might be spuriously associated with
victimization through its association with husband’s intoxication and via the indirect effects of
victimization histories in the family of origin of both partners.
1. Introduction
Most theory and research on the link between drugs, particularly alcohol, and
crimes of violence have focused on male behavior and have been from the
Glenda Kaufman Kantor and Nancy Asdigian • Family Research Laboratory, University of New
Hampshire, Durham, New Hampshire 03824.
315
perspective of “intoxication–aggressor” effects, i.e., intoxication as a precipi-

tant of aggressive acts.1–7 A complementary perspective suggested by Kauf-
man Kantor and Straus,8 and the focus of this chapter, is that “intoxication–
victim” effects may also exist. This perspective is based on the idea that
women under the influence of alcohol and other drugs may become targets of
male aggression (i.e., “fair game”). Many assume that intoxicated women
violate the norms of appropriate female behavior, and thus lose the protection
afforded by other traditional gender role norms, such as “never hit a wom-
an,”9–11 and a physically violent response to “provocation” is therefore legiti-
mated. Similarly, women under the influence of drugs or alcohol are viewed
as sexually available,12 and are more likely to be victims of sexual aggres-
sion.13 However, studies suggesting that males increase their aggression to-
ward women who violate gender role expectations may provide only a partial
explanation of why intoxicated women are at risk for victimization, and the
“fair game” assumption is difficult to establish. Few studies specifically assess
the vocabularies of offender motive, and more often motive is intuited as a
post-hoc explanation of alcohol–victimization effects.8 Additionally, there is a
need to consider whether, in fact, women who drink heavily or use drugs are
at greater risk for spouse abuse, as well as the adequacy of theory and re-
search conducted on intoxication–victimization effects.
There are apparent limitations in generalizing from experimental studies
of alcohol-induced aggression,1,5,14 to studies of alcohol and wife assault, be-
cause in the former condition most confederates are strangers. Moreover, the
vast majority of alcohol and aggression studies do not take into account the
mediating effects of an intoxicated partner. Experimental studies that include
intoxicated dyads with real partners15 or that examine the effects of an intoxi-
cated female partner on male aggression are also uncommon. Although ex-
perimental analogue studies have yielded important information about alco-
hol-induced aggression, the experimental paradigms (e.g., the competitive
reaction-time paradigm) have not been suitable for studying marital aggres-
sion.16 Additionally, many studies of alcohol-related marital violence have
failed to take into account the intoxicant usage pattern of both partners.
Empirical studies documenting the incidence of victimization among heavy-
drinking or alcoholic women compared to moderate-drinking counterparts
are a relatively recent area of study.17,18 The latter studies are important to
consider because they suggest that female intoxication may have direct or
indirect effects on women’s victimization experiences.
Experts on alcohol-related wife assaults concur that alcohol–aggression
relationships are complex phenomena mediated by a number of factors such
as family history, personality, the drinking context, the amount drunk, and
alcohol- and violence-related cognitions. 19–22 However, as Pernanen23 ob-
served, most theoretical approaches have been individual-based, thus mini-
mizing the social dynamics of alcohol-related interactions. In this chapter we
consider the relevance and adequacy of existing theory to explaining intoxi-
cant–victimization effects, as well as the empirical basis for the hypothesized
13 • Women under the Influence 317
relationships. Specifically, we address the following questions: (1) What are

the hypothesized dimensions and mechanisms of intoxication–victimization
effects? (2) Does the empirical literature show that women who drink heavily
or use drugs are at greater risk for abuse by spouses? (3) Does the importance
of female drinking to victimization vary by the event (type of victimization) or
by the population sample studied? (4) Is this a “spurious” relationship; for
example, a relationship that simply reflects drug/alcohol use by the husband
or reflects other confounding variables such as a history of abuse in the family
of origin?
2. What Are the Theoretical Mechanisms?

Both theories of wife abuse and theories of alcohol-induced aggression
must be considered in formulating an integrated conceptual framework for
interpreting intoxication–victimization effects.
2.1. Theories of Victimization

A number of theoretical perspectives have been utilized to explain wife
abuse. For example, Bersani and Chen24 (1988) examine seven theoretical
perspectives on family violence. Gelles and Straus25 identify 15 theories,
which they organize into three broad categories: intraindividual theory,
social psychological theory, and sociocultural theory. Intraindividual theory
has emphasized alcohol–drug effects and psychological traits such as self-
esteem26–28 and antisocial personality disorder.29 Sociocultural theories such
as systems theory have attempted to integrate social structural and family
processes.30,31 Feminist explanations of women’s victimization also empha-
size sociocultural factors, especially the male-dominated social structure and
socialization practices.32,33 Social–psychological approaches have stressed,
for one, social learning through experience and exposure to violence.34–36
It is important to note that even though we cast the above approaches as
theories of victimization, for the most part the theories pertain to explanations
of the perpetrator’s aggression. This theoretical emphasis on the perpetrator
is supported by previous analyses finding that 14 characteristics were consis-
tently associated with being an abuser, whereas only one factor was consis-
tently associated with being a victim of wife abuse.37 Furthermore, Dobash
and Dobash38 point out that blaming the victim of wife beating provides a
justification for batterers’ violence. When women nag, try to have an equal
say in family decisions, or refuse sex, husbands may feel they are justified in
using force.39 In these cases, the woman is blamed for her own victimization.
This same mechanism may apply when women are intoxicated.
Alcohol figures in only two of the many theories discussed above—
intraindividual theory and social-learning theory. For example, O’Leary’s35
social-learning model of spousal aggression incorporates five major factors
that predict spousal aggression: (1) violence in the family of origin; (2) aggres-
sive personality style; (3) stress; (4) alcohol use and abuse; and (5) marital
dissatisfaction. Prior research2 on alcohol–wife assault relationships has con-
cluded that alcohol use at the time of wife assault is far from necessary or
sufficient for wife abuse to ensue, despite the stereotype that all drunks hit or
all hitting involves drunks. Yet, alcohol use by the perpetrator has emerged as
a major risk factor in wife assault research.2,4,15,40–42
2.2. Theories of Intoxicant-Induced Aggression

General theories of alcohol-induced aggression are usually advanced to
explain alcohol-related wife assaults, but their ability to account for the special
case of wife assault or for intoxication–victimization effects needs to be evalu-
ated. A primary theory utilized to explain the relationship between alcohol
and aggression is that alcohol has psychopharmacological effects, i.e., intoxi-
cants affect personality and affective states.43 Researchers employing experi-
mental paradigms in examining alcohol effects on aggression have generally
demonstrated both direct dose effects of alcohol on aggression and indirect
effects mediated by social, psychological, and environmental factors. The
factors include, but are not limited to, the aggressive predisposition of the
individual, expectancies, and perceptions of threat or provocation.1,3,6,14,44–48
Alcohol is the drug most commonly associated with violence19,43; how-
ever, other psychoactive drugs have been associated with aggression such as
barbiturates, amphetamines, opiate (withdrawal), phencyclidine, cocaine,
and alcohol–cocaine combinations.43,49 Altogether, the aggressive effects of
psychoactive drugs other than alcohol have received less empirical scrutiny,
and the results have been inconsistent. Taylor and Chermack’s50 review finds
that drugs with depressogenic characteristics are most likely to facilitate ag-
gression, and Fagan43 finds solid evidence of a psychopharmacological basis
only for the combined effects of an alcohol–cocaine combination. It must also
be concluded that aggression is not an inevitable consequence of any intoxi-
cant usage.
2.3. Mechanisms of Intoxicant-Induced Victimization

2.3.1. Cognitive Effects. Despite ongoing debate about the precise pharma-
cological effects, there is consensus that alcohol disrupts cognitive function-
ing by diminishing ability to reason, reducing ability to perceive or calculate
consequences of aggressive behaviors, and by enhancing perception of
threat.51,50 All of the latter mechanisms bear upon the process of social inter-
action, and are likely mechanisms that could increase the risk of victimization.
It is clear that aggression may be a consequence of alcohol's impairment of
perception, judgment, and memory. Distorted perceptions increase the likeli-
hood of miscommunications, developing resentment, and the inability to take
into account the consequences of aggressive actions. A greater possibility
exists for escalating verbal aggression between partners that in turn leads to
physical aggression. All of the cognitive misprocessing and miscommunica-
tion may be magnified when both partners are under the influence. However,
the latter assumes a certain invariance in the dynamics of the social interac-
tion. For one, men and women may not behave similarly when under the
influence. 52
2.3.2. Expectancy Effects. Although gender differences in alcohol expectan-

cies have been examined in a number of studies, most researchers have not
systematically explored the specific alcohol-related beliefs held by males and
females.53,54 Previous assessments have been very general in scope, involving
comparisons between males and females across a wide spectrum of emotional
and behavioral expectancies. For example, in their analysis of the Alcohol
Expectancy Questionnaire’s (AEQ) six dimensions, Brown and associates55
reported that females were more likely than males to expect alcohol to result
in positive social experiences, whereas males were more likely to expect
heightened arousal and aggression. Rohsenhow56 examined gender differ-
ences on a number of alcohol expectancy dimensions in addition to those
identified by the AEQ. After differences in drinking habits were controlled,
women expected more cognitive and motor impairment than men did, but
less positive affect, social and physical pleasure, and relaxation. Aggressive
alcohol expectancies did not vary, however, as a function of gender in
Rohsenhow’s study or in a subsequent investigation.57 In Kline’s58 study,
gender differences were not observed on any of the five AEQ factors included
in that research (the arousal/aggression dimension was not used). Critchlow-
Leigh54 administered a different expectancy measure to male and female sub-
jects and found that women expected to suffer more cognitive and physical
impairment (e.g., feeling sleepy, sick, dizzy) from drinking than men did.
Expectancies regarding alcohol-induced nastiness (e. g., being mean, becom-
ing aggressive) were, however, stronger among males even after controlling
for gender differences in drinking patterns.
In a more recent study,59 we focused specifically on gender differences in
expectations concerning drinking-related aggression. The results of this study
of nonabstinent couples suggested that the structure of alcohol beliefs varies
by gender, and that gender differences in aggressive alcohol expectancies are
highly specific. Male respondents were more likely than their female counter-
parts to associate drinking with heightened feelings of power and influence
over others and with irritability and short-temperedness, These findings sug-
gest that alcohol consumption, because it enhances feelings of male domi-
nance and increases the ease with which men are provoked, may create a set
of conditions conducive to male-perpetrated aggression. However, both
heavy-drinking men and women had similar perceptions that drinking may
make aggressive acts more forgivable.
An examination of alcohol-related violence among these couples revealed
that men were twice as likely as women to be drinking at the time of a spousal
assault. This study also found that alcohol-related assaults against wives were
more likely when both partners were heavy drinkers. Typical consumption
levels of wives were the only significant predictors of husbands’ alcohol-
related assaults against wives. On the other hand, drinking-related assaults
perpetrated by wives against husbands were predicted from both women’s
typical alcohol consumption patterns as well as aggressive alcohol expectan-
cies held by the women.
2.3.3. Family History. The results of research examining gender differences

in aggressive alcohol expectancies do not preclude the possibility that intox-
ication–aggression and victimization effects for men and women are at least
partially attributable to histories of victimization. Thus, the cultural origins of
aggressive alcohol expectancies, normative approval of violence toward wom-
en, and aggressive behavior may be rooted in the family. We20 developed a
model of the socialization processes underlying alcohol-related wife assaults.
Based on our evaluation of that model, we concluded that men who were
abused by their parents or who witnessed violence between their parents as
children are, as adults, more likely to approve of aggression against wives as
well as more likely to drink heavily. As a result of their heavier drinking and
beliefs regarding the appropriateness of spousal violence, men have stronger
expectancies concerning the aggressive consequences of alcohol and, in turn,
engage in more assaults on wives. Men who grew up with alcohol-abusing
parents, although not necessarily heavier drinkers, expected alcohol use to
engender violence and were themselves more likely to assault their wives.
Family history of alcohol problems indirectly affected wife assaults in the next
generation through their effects on aggressive expectancies. Overall, our re-
sults indicated that violent socialization in the family of origin exerts stronger
effects on current intrafamily violence than does a family history of heavy
drinking. Our findings suggested that heavy alcohol consumption and cogni-
tions concerning the effects of alcohol consumption on aggression represent a
major pathway through which intrafamilial violence is transmitted across
generations.
Exposure to childhood abuse or marital violence in the family of origin
may similarly increase women’s risks of victimizations as it does for men. For
example, assaulted women have at least double the rate of growing up in a
violent home compared to nonassaulted women.8 However, O’Leary35 sug-
gests that this relationship is less consistent for women compared to men.
Where intergenerational effects exist, the mechanism may be that repeated
attacks by parents can lead to damaged self-esteem and suppression of rage.
These effects may be manifested differentially by gender and compounded by
the current life situation. Women who were harshly punished in childhood or
who witnessed parental violence are more likely to be victimized as adults
because they have low self-esteem and have learned that assaults from a
loved one are legitimate, or because they are more likely to engage in mutual
assaults with their spouses.60 Additionally, because alcoholism and family
violence are so intertwined, both women and men exposed to violence at

home also may have suffered the effects of parental alcoholism.61 Widom’s61
review of the literature on child abuse and alcohol use concludes that while
studies of female alcoholics indicate a higher incidence of child abuse history,
the literature is inconsistent and methodologically flawed. There may also be
effects of a family history of parental alcoholism on women’s future choice of
partners. For example, if daughters of alcoholics are more likely to marry
alcoholics,62 then they may also be at risk for intoxication–victimization ef-
fects via this assortative mating process.
2.3.4. Personality. Powers and Kutash47 have emphasized the contributions

of premorbid personality characteristics and the interaction between sub-
stance abuse and personality as central to substance-abuse-related aggres-
sion. They view the personality characteristics of the abuser as the primary
problem, independent of the behavior of the partner. Research on maritally
violent men also suggests that psychopathology and personality disorders are
present among some subtypes of batterers.29 Literature based on clinical and
criminal justice samples of alcoholics or assaultive men also suggests that
aggressive drinking patterns may be a reflection of the concurrence of antiso-
cial personality disorders with alcoholism.63,64 However, research based on
community samples indicates that husband’s alcohol use remains signifi-
cantly related to assaults on wives even after controlling for measures of
hostility, self-consciousness, and other sociodemographic risk factors.4,40
In sum, the diversity of opinion on personality issues is most likely due
to the variability found in different types of samples. Clinical samples of
battered men in treatment may differ from criminal justice samples, alcoholics
in treatment, or men from community samples.
There is little research examining the contribution of women’s person-
ality disorders to their own victimization. Researchers have concluded that
abusive men’s characteristics are the appropriate area of inquiry.37 However,
because studies do find high rates of assaults among alcoholic women, we
need to consider the possibility that personality factors among such women
may play a role in the dynamics of victimization. For example, clinical data on
alcoholics in treatment show that depression and low self-esteem are com-
mon among alcoholic women.65,66 Such symptomatology may perpetuate or
exacerbate marital disharmony by producing resentment and rejection on the
part of the nondepressed spouse, or by serving as justification for the assaults
of an abusive mate. It is also just as likely that poor self-esteem and depres-
sion are reciprocally related and worsened by the denigration and physical
abuse of a partner.
2.3.5. Social Dynamics. Drinking patterns of husbands and wives tend to be

highly similar.67 It is not surprising then that a married alcoholic has a high
probability of having an alcoholic partner.68 As suggested above, assaults are
more common in relationships where both partners are heavy drinkers.59
Furthermore, alcoholic women may be more verbally or physically aggressive

toward a partner.69 However, we also need to consider the social dynamics of
relationships where only one partner (usually the male) is a heavy drinker.
Conflicts may ensue over any number of issues, including the partner’s drink-
ing or drug problems. Additionally, the partner’s violent acts may be un-
predictable or inconsistent with episodes of intoxication.
Powers and Kutash47 suggest that nondrinking partners may resent
drinking episodes and become less nurturant and sympathetic, which, in
turn, increases the abuser/drinker’s rage. They allege that this dynamic is
especially true in the case of the drinker with high dependency needs, who
experiences frustration in response to his perception of a nonempathic part-
ner. Some support for this description of the social dynamics of alcohol-
related conflicts is found in the experimental literature as well. For example,
Jacob, Ritchey, Cvitkovic, and Blane70 report greater negativity of alcoholic
couples (more disagreements by wives) on “drink night” compared to normal
couples. But there may also be differences in communication patterns, de-
pending on the drinking pattern of the husband. When binge and steady
drinkers were compared, binge drinkers were found to have more negative
communications when drinking, while steady drinkers had more positive
communications when drinking during the experimental manipulation. Such
effects may occur because partners adapt or become habituated to steady
drinking behaviors or because binge drinkers may themselves have other
characteristics such as greater aggressiveness that are confounded with drink-
ing. Additionally, chronic alcohol use can create economic stresses on the
family such as unemployment, which increase family conflict and the likeli-
hood of abuse.41 Finally, alcohol may be used to excuse the violence by the
husband or wife, or by both partners.
3. How Common Is Drinking by Both Parties?

Evidence for drinking by both parties or the association of female intox-
ication with victimization is most apparent in studies of homicide and rape
victims. Wolfgang’s71 classic study of victim-precipitated homicide carefully
defines this category as one in which the victim was the first to use physical
force or display a weapon. He found that alcohol was present in almost two
thirds of the cases. In forty-four percent of those cases, both victim and
offender were drinking. However, all of the case illustrations Wolfgang pro-
vides of victim precipitation in families are instances where the victim is a
male batterer killed by his beaten wife. More recent research on the alcohol-
homicide relationship indicates that between one third and two thirds of
victims had positive blood alcohol levels. Consistent with the greater preva-
lence of heavier drinking by males, the majority of epidemiological data
shows that intoxication is more common among male homicide victims com-
pared to female homicide victims.72–74
Murdoch, Pihl, and Ross75 reviewed 19 studies of homicides and 7 stud-

ies of assaults to determine who was drinking during crimes of violence. Our
reinterpretation of the data presented by these authors indicates that more
than half of all homicide offenders were drinking in 13 of 19 studies. It was
rare for findings on victim intoxication to parallel those for offenders. For
example, more than half of all victims were drinking in only 3 of the 19
homicide studies reviewed. Of the seven assault studies reviewed, more than
half of all offenders were drinking in three of the seven studies, but in no
cases were more than half of the victim group drinking. It should also be
noted that in three of the seven assault studies the incidence of victim drink-
ing ranged from 25 to 40% of all victims. These authors75 conclude that the
instigator of the attack is more likely to be intoxicated, but the victim is “as
likely as the offender to initiate the altercation by attacking or moving in a
way that can be interpreted as an attack” (p. 1070). However, gender issues
are blurred in this analysis, and the fact that intoxication can distort percep-
tions of threat and provocation needs to be considered.
Data provided by rape victims indicate that 40% report drinking or intox-
ication prior to their rape.76 Roizen’s77 analysis of data on arrested popula-
tions finds considerable variability regarding alcohol involvement by per-
petrators or victims in rape cases. Roizen rejects the concept of “victim
precipitation” relevant to crimes of rape, noting that intoxication is responsi-
ble for the misreading of cues by victim and offender. Women’s greater vul-
nerability to rape when intoxicated may also occur because they are less able
to escape undesirable or dangerous situations. Her comments regarding the
potential mechanisms involved in the association between alcohol and rape
are also relevant to alcohol’s role in wife-assault victimizations. She notes the
following:
Thus drinking in rape, as in other crimes, may play any one of a number of
different roles: It may be present but have no effect; it may enhance
chances of victimization when the parties are strangers; it can be present in
the offender alone and exert an effect only on the offender, such as mis-
reading social cues in relation to prevailing norms; or it may begin an
evening gathering of a group of men that ends in drunkenness and rape.
(P. 17)
4. Empirical Studies of Wife Assault and Intoxication

by Female Victims
Empirical evidence supporting an association between women’s intoxica-
tion and physical assaults by husbands is mixed. In one early review of the
risk markers for husband-to-wife assaults, Hotaling and Sugarman37 exam-
ined existing case-control research on women’s drug use and alcohol con-
sumption. Drug use was declared an “inconsistent risk marker,” bearing a
significant relation to husband-to-wife violence in 60% of the studies re-
viewed. Nevertheless, the fact that three of the five studies of drug use by
wives did find an association with wife abuse suggests that drug use may put
women at risk of being a victim of spouse abuse. In contrast, alcohol con-
sumption by wives was related to violence in only 17% (1 of 6) of the studies
reviewed. As such, Hotaling and Sugarman concluded that women’s use of
alcohol does not aid in the prediction of spousal assaults and should be
considered a “consistent nonrisk marker.”
4.1. Evidence from General Population Surveys

Since the publication of the aforementioned review, other investigations
have obtained somewhat stronger support for an intoxication–victimization
effect. Based on data from the 1985 National Family Violence Survey (NFVS),
Kaufman Kantor and Straus8 found that female respondents who reported
minor physical assaults by husbands during the survey year were more than
three times as likely as nonvictims to have been high on drugs during the
preceding year and more than twice as likely to have been drunk at least once.
In addition, any prior year, drunkenness was three times as likely among
victims of severe assaults compared to nonvictims and rates of drug intoxica-
tion were six times higher among the former. It should be emphasized, how-
ever, that the measures used by Kaufman Kantor and Straus8 were limited.
They only assessed the use of any prior year illicit drug use and any episodes
of drunkenness by the husband or the wife.
The National Alcohol and Family Violence Survey (NAFVS)31 also exam-
ined questions concerning substance abuse–aggression and –victimization
relationships. One of the parallel purposes for both survey years (1985 and
1992) was to consider whether substance abuse by the wife or husband con-
tributed to assaults on wives. In 1992, Kaufman Kantor and Jasinski78 exam-
ined this association using much more precise measures of alcohol and other
drug use, yet found similar results. Wives’ substance abuse showed similarly
strong bivariate associations to husband-to-wife violence in Kaufman Kantor
and Jasinski’s78 analysis of the 1992 NAFVS. That analysis examined relations
between husband-to-wife violence and both husbands and wife’s use of alco-
hol and other illicit drugs.
As shown in Fig. 1, rates of wife assault were between two and six times
higher among women who abused alcohol and/or drugs in the previous year
than among women who reported no substance abuse. The highest rates of
victimization were observed among women who either used both alcohol and
marijuana or marijuana only. However, because relatively few women report-
ed abusing alcohol or other drugs, some caution is needed regarding the
reliability of these results. Indeed, the higher frequency of substance abuse
among men may be responsible for the somewhat stronger role of husband’s
substance abuse in wife assaults apparent in Fig. 1. The results of the bivariate
analyses for men indicate that multiple drug use patterns, i.e., the combined
Previous Year Substance Abuse Patterns
*Reflects heavy drinking
Figure 1. Rates of wife assault as a function of husbands and wife’s substance use
patterns.
pattern of alcohol and marijuana use and use of alcohol and hard drugs are
associated with the highest rates of assaults against women.
Although impressive, the bivariate data reviewed above paints only an
incomplete picture of intoxication–victimization effects in the general popula-
tion. In Kaufman Kantor and Straus’8 multivariate analysis of NFVS data,
neither wife’s drinking nor drug use were significantly related to severe as-
saults (a low-base-rate phenomenon in general population samples). How-
ever, past-year drinking and drug use by the wife did predict minor physical
assaults, after controlling for husband drinking and drug use and other socio-
demographic risk factors as covariates in the prediction of husband-to-wife
violence. Kaufman Kantor and Jasinski79 failed to show significant effects of
wife’s alcohol use on husband-to-wife violence when tested along with a
similar set of covariates (see Table I).
Table I summarizes the results of comparable multivariate analyses
across surveys. In 1985, the most significant predictors were husbands illicit
drug use, drunkenness by husband and wife, low income, violence in the
family of origin, and normative approval of violence. In 1992, the results of
multivariate analysis showed that husband’s illicit drug use and heavy drink-
ing again strongly predicted wife assaults, as did family of origin violence,
Table I. Substance Abuse & Wife Assault: Significant Predictors
1985 NFVS 1992 NAFVS
Husband illicit drug use 1 or >X Husband illicit drug use 1 or >X
Wife drunk 1 or >X Problem drinking husband
Husband drunk 1 or >X Wife marijuana use
Low income Youth
Family of origin violence Family of origin violence
Violence approval Violence approval
Unemployment
violence approval, youth, and unemployment. Heavy drinking by the wife

did not emerge as a significant predictor.
Additional analyses of the 1992 survey data examined models of specific
types of substance abuse, while controlling for important sociodemographic
risk factors. The results of these analyses suggested that past-year usage of
illicit hard drugs (e.g., nonprescription amphetamines, barbiturates, tran-
quilizers, cocaine, and heroin) by the husband and marijuana use by the wife
significantly increased the odds of wife assault, net of other risk factors in the
models.
4.1.1. Drinking at the Time of Wife Assault. Moreover, it is important to

emphasize that the measures used by both Kaufman Kantor and Straus8 and
Kaufman Kantor and Jasinski79 assessed prior-year drug and alcohol use rath-
er than usage in relation to specific events, such as assaults by husbands.
Given the omission of data on the temporal ordering of women’s intoxication
and their victimizations, no conclusions can be made about the precise role of
women’s intoxication in the etiology of husband-to-wife violence. When data
on drinking at the time of wife assaults has been assessed, women are much
less likely than men to report intoxication during violent episodes. As shown
in Table II, only 10% of all wife assaults reported in the 1985 NFVS involved
drinking by wives compared to the 22% that involved drinking by husbands.
Likewise, wives were drinking in only 8% of the violent episodes reported in
the 1992 NAFVS in contrast to 28% of those episodes in which husbands were
drinking. Thus, at least part of the association between wife’s intoxication
patterns and histories of victimization may involve postabuse drinking or
drinking unrelated to the specific episodes of assault. Overall, it is rare for
women to be the sole substance abuser in intimate relationships, while this is
not uncommon among husbands. The results show that drinking by the
woman at the time of violence is uncommon among abused women in general
population samples.
4.1.2. Drug Use at the Time of Wife Assault. One strength of the 1992 study
is that data were collected on the use of drugs other than alcohol at the time of
Table II. Drinking at the Time of Violence
1985 NFVS 1992 NAFVS
Neither drinking 76% 71%

Husband only drinking 14% 22%
Wife only drinking 2% 2%
Both drinking 8% 6%
assaults on wives. We found that alcohol is the drug most commonly associ-
ated with intoxicated aggression. Other drug use at the time of violence
occurred in only 5% of all families and more often in conjunction with alcohol.
Although wives infrequently used alcohol at the time of their victimizations,
those who did so were significantly more likely than nondrinking abused
women to be using other drugs at the time of assault.
Leonard and Senchak80 obtained similar results to those of the two na-
tional surveys discussed above when they examined the relationship between
heavy drinking among newly married women and levels of premarital aggres-
sion by husbands. At the bivariate level, frequent excessive drinking by wives
was modestly correlated with husband’s use of minor violence prior to mar-
riage, as was husband’s excessive use of alcohol. When the overlap between
husband and wife alcohol abuse was disentangled in multivariate analyses,
women’s drinking no longer predicted husband-to-wife violence. As with the
analyses by Kaufman Kantor and Jasinski,79 these findings suggest that the
confounding of husband and wife intoxication might underlie the apparent
effects of drug or alcohol use among female victims of domestic violence.
4.2. Evidence from Clinical Populations

Unlike the analyses based on data from the two national probability
surveys, most intoxication–victimization research has been conducted with
clinical samples of heavy-drinking and/or battered women. For example,
Frieze and Schafer’s52 report of a study of battered women suggests that
alcohol use by either partner is related to marital violence in varying ways. In
a primary group of families identified by factor analysis, both husband and
wife were frequently intoxicated and marital violence was consistently associ-
ated with drinking. In other families, drinking was less consistently asso-
ciated with fighting. Based on their sample of battered women, these authors
conclude that high levels of violence in both husband and wife are more
likely in families where the husband has the drinking problem and the wife
does not.
Miller, Downs, and Gondoli81 compared rates of spousal assault among
women in treatment for alcoholism and nonalcoholic women in the general
population. Reports of both minor and severe physical assaults were higher
among alcoholic women, and husband-to-wife violence was a highly signifi-
cant predictor of female alcoholism in multivariate analyses that controlled for

husband drinking and family of origin alcoholism and childhood physical
abuse. Miller and associates, however, did not report whether female alcohol-
ism aided in the prediction of spousal victimization after controlling for its
overlap with family of origin dysfunction and husband’s alcohol problems.
Miller, Downs, Testa, and Panek82 extended the above analysis by com-
paring the female alcoholics to groups of nonalcoholic women sampled from
battered women’s shelters, mental health centers, driving while intoxicated
classes, and the general population. The alcoholics reported higher levels and
more frequent verbal and physical aggression than all other groups except
shelter women, who obtained the highest scores on all violence measures.
Alcoholism, however, did predict high-frequency verbal and physical assaults
by husbands after controlling for group differences in sociodemographic char-
acteristics and mental health symptomatology. The results of these multivari-
ate analyses, although more informative than the bivariate findings, are not
entirely conclusive because they fail to consider the overlap between women’s
alcohol abuse and that of husbands.
Additional findings from Miller and associates82 suggest that the severity
of a woman’s alcohol abuse may explain why intoxication–victimization ef-
fects are obtained in some investigations but not others. They failed to find a
difference in husband-to-wife verbal and physical aggression between alco-
hol-abusing women in the mental health and shelter groups (who had been
excluded from the primary analyses) and their nonabusing counterparts. It is
possible that the alcohol problems of these women were not as advanced as
those among the alcohol-treatment women. If so, it may be that heavy drink-
ing is not as strong of a risk factor for spousal assaults than is the clinical
syndrome of alcoholism. Such an explanation would be consistent with the
tendency of researchers to observe intoxication–victimization effects in clini-
cal samples but not in community samples.80
Using a different sample, Bergman, Larsson, Brismar, and Klang83 com-
pared the victimization experiences of hospitalized female alcoholics to those
of women seeking treatment for either spousal assaults or medical conditions
unrelated to domestic violence. Compared to the medical control group, alco-
holics were more likely to have been battered at least once and to have had
multiple experiences of battering. Moreover, assault victims in the alcoholic
group were more likely than victims in both the medical control and battered
group to have suffered assault-related injuries severe enough to necessitate
medical treatment. In a supplementary analysis, Bergman and associates also
found rates of alcohol dependency to be three times as high among women in
treatment for spousal assaults than among women in treatment for medical
conditions. However, it should also be noted that 70% of the women in the
alcoholic and spousal assault groups had husbands who abused alcohol com-
pared to only 14% of the husbands in the medical control group. Thus, it is
unclear whether the association between alcohol abuse and victimization ob-
served in this investigation is due to women’s drinking or that of husbands.
In addition to comparing rates of wife assault among alcoholic and non-

alcoholic women, researchers have also assessed levels of alcohol and/or drug
use among couples in violent and nonviolent relationships. This approach
was used by Barnett and Fagan,84 who asked four groups of men to report on
their own and their wife’s typical pattern of drinking. Two groups included
physically abusive husbands who either were or were not in treatment for
wife battering. The remaining groups included maritally discordant but non-
abusive husbands and maritally satisfied husbands.
Compared to the self- and partner-reports provided by men in both non-
abusive groups, quantity of alcohol consumption was higher among both
maritally violent men and their wives. Moreover, when asked about drinking
context, violent men revealed that they were significantly more likely than
their abused wives to be drinking during a violent episode. In contrast, post-
abuse drinking by wives was much more common than postabuse drinking
by husbands. These findings are consistent with Kaufman Kantor and Straus'2
report that drinking at the time of husband-to-wife violence is more common
among men than among women.
5. Conclusions
5.1. The Role of Women’s Intoxication in Husband-to-Wife Violence
The evidence reviewed above, although suggestive of an association be-
tween women’s intoxication and assaults by husbands, does not clearly indi-
cate the nature or direction of that association. The findings, however, can be
used to evaluate which of several mechanisms might underlie such an asso-
ciation.
5.1.1. Women’s Intoxication Provokes Assaults by Husbands. Perhaps the

most frequently espoused hypothesis is that drinking or drug use by women
can incite their male partners to behave violently. Such attacks are presumed
to stem from men’s anger at their intoxicated women’s violation of gender role
prescriptions against female drunkenness.8 In a similar vein, it has also been
suggested that assaults on drunken wives might be precipitated by an in-
crease in verbal aggression that often accompanies female intoxication.59
Support for this hypothesis is weak, if not nonexistent, because the tem-
poral precedence of women’s drinking has not been established in any intox-
ication–victimization investigation that we have reviewed. Most research ei-
ther reports correlational data on women’s typical usage patterns over a
specified referent period and their victimization experiences during the same
time period8,84 or examines the relation between current alcohol status and
victimization history.81–83 When drinking at the time of the violence has been
specifically assessed,2,59,79,84 intoxication is much more common among hus-
bands than it is among wives. Thus, as frequently as the above hypothesis is
advanced, there is little evidence from the empirical literature that women’s
drinking provokes or even precedes aggression by husbands.
However, much of the empirical literature that specifically examines is-
sues of temporal precedence is based on general population samples in which
problem-drinking women are a low-base-rate phenomenon. Research exam-
ining clinical samples of alcoholic women does suggest that such women have
high rates of victimization over multiple relationships. Such victimizations
appear to be more a function of their partner’s drinking and propensity to
violent behavior. However, the empirical literature has rarely attended to
specific motivations for assaults on wives. There is evidence that cultural
norms about the appropriateness of hitting wives are associated with wife
assaults. Similarly, the likelihood of victimizations may also be influenced and
perpetuated by societal beliefs that denigrate alcoholic women, as well as the
low self-esteem and depression characteristic of many alcoholic women.
These processes may operate independently of whether the woman is con-
suming alcohol or other drugs at the time of an assault. Among couples where
both are intoxicated at the time of a conflict situation, there is, at minimum,
theoretical evidence that the psychopharmacological effects of intoxication
can increase cognitive distortions, which escalate aggressive behavior in men
and women. However, theory development and empirical testing are needed
in regard to understanding the social dynamics of alcohol-related interactions.
5.1.2. Women’s Intoxication Is a Consequence of Victimization. An alterna-

tive explanation for the intoxication–victimization relation is that women
drink or use drugs as a means of coping with a violent partner. As mentioned
previously, Barnett and Fagan84 reported that postviolence drinking was more
common than drinking prior to and/or during a violent incident among the
wives of assaultive men. This explanation is also consistent with another body
of literature showing that battered women are at particular risk for depressive
symptomatology,85,86 and thus may use alcohol and/or drugs to self-medicate
their dysphoria. Analysis of hospital record data in one study revealed greater
problem drug use for battered women treated for injuries than for nonbat-
tered women.87 This suggests that at minimum, among the subpopulation of
the most severely battered and injured women seeking help, that substance
abuse may be a means of coping with victimization or it may be a correlate of
family life where there is limited commitment to conventional behaviors.
5.1.3. Women’s Intoxication Is Indirectly Related to Victimization via Hus-

band’s Intoxication. A large body of literature has established that drinking
by husbands increases a woman’s risk for physical assault.2,4,8,15,40,41,88 In ad-
dition, it is also well documented that the substance abuse patterns of wives
are highly correlated with those of their husbands. Women’s intoxication
might then be spuriously associated with victimization through its association
with husband’s intoxication. Recall that the strong bivariate associations that
Kaufman Kantor and Straus8 observed between women’s drinking/drug use
and victimization were either eliminated or substantially weakened in multi-

variate analyses that statistically controlled for husband’s drinking and drug
use. A comparison of the bivariate and multivariate results reported by
Leonard and Senchak80 showed the identical pattern. Moreover, the strongest
evidence of an intoxication–victimization effect has come from studies that
failed to control for husband drinking.82,83 Unequivocal support for a direct
role of women’s drinking in spousal assaults can only come from analyses that
disentangle the overlap between husband and wife intoxication.
5.1.4. Women’s Intoxication Is Indirectly Related to Victimization via Fami-

ly of Origin Abuse. Intoxication–victimization effects might also be observed
because both excess drinking among women and spousal victimization have
common roots in childhood experiences of physical and sexual victimization.
A host of studies8,34,36,89 have demonstrated that victimization in one genera-
tion is often repeated in the next generation, such that childhood abuse at the
hands of parents or the witnessing of parental violence significantly increases
a woman’s risk of spousal violence in adulthood. In addition, evidence from a
number of investigations90,91 indicates that female alcoholics experienced
higher levels of verbal and physical aggression (particularly at the hands of
fathers) than did nonalcoholic women. The common effects of violence in the
family of origin could thus produce significant, albeit noncausal, relationships
between wife’s drinking and victimization by husbands.
In sum, the evidence for intoxication–victimization effects are strongest
among studies of rape, homicide, and studies of alcoholic women. We con-
clude that the theoretical basis for intoxication–victimization effects is best
exemplified in situations in which both partners are heavy drinkers and
where the social dynamics of situations are taken into account. We also con-
clude that, in reality, women are less often intoxicated than their partner at
the time of an assaultive episode. Despite the strength of alcohol’s association
with wife assaults, intoxication’s centrality and temporal relationship to spe-
cific wife assault episodes is highly variable, regardless of whether it is a
component of the husband’s or wife’s behavior.
Among alcoholic women, risks for victimization are greater for a number
of reasons, including their violation of traditional female societal norms that
stigmatize alcoholism in women more so than in men.11 They are also at risk
due to their increased use of aggression59 and because they are more likely to
have a heavy-drinking male partner. Finally, women drinking to intoxication
are at risk because they are more likely to have a history of victimization in
their family of origin. They may bring a low sense of self to the relationship
that reinforces beliefs that they deserved to be beaten owing to their drunken-
ness, aggression, or unworthiness. Often, abusive men will instill or reinforce
such beliefs by their own accusations and behaviors toward the woman.
Women growing up in a violent home are also more likely to consider violent
modes of conflict resolution as a “normal” part of intimate behavior. Outreach
efforts to alcoholic women are important because women who believe they
provoked the violence or were responsible for the violence may perpetuate
these patterns by avoidance of help-seeking due to guilt, shame, or even
denial that they have been battered. Unfortunately, similar mechanisms of
denial and minimization exist for both alcoholism and battering. Treatment
providers must also address the complex mechanisms of intoxication–victim-
ization and aggression processes, rather than emphasizing simple justifica-
tion explanations (e.g., drunkenness as a means of excusing the violence) or
explanations of intentionality (e.g., men drink in order to beat their wives) as
a means of reducing wife assaults.
ACKNOWLEDGMENTS. The research reported here was supported by research
grant No. RO1AA09070 from the National Institute on Alcohol Abuse and
Alcoholism. We are grateful to Kelly Foster for her assistance in manuscript
preparation.
References
review. Psychol Bull 107(3):341–354, 1990.
2. Kaufman Kantor G, Straus MA: The “drunken bum” theory of wife beating. Soc Prob
34(3):213–230, 1987.
drinkers. J Abnorm Psychol 84:508–518, 1975.
4. Leonard KE, Bromet EJ, Parkinson DK, et al: Patterns of alcohol use and physically aggressive
behavior in men. J Stud Alcohol 46(4):279–282, 1985.
5. Taylor SP, Gammon CB: Effects of type and dose of alcohol on human physical aggression.
J Personal Soc Psychol 32(1):169–175, 1975.
6. Taylor SP, Gammon CB, Capasso DR: Aggression as a function of the interaction of alcohol
and threat. J Personal Soc Psychol 34:938–941, 1976.
7. Taylor SP: Experimental investigation of alcohol-induced aggression in humans. Alcohol
Health Res World 17(2):108–112, 1993.
8. Kaufman Kantor G, Straus MA: Substance abuse as a precipitant of wife abuse victimizations.
9. Young DM, Beier EG, Beier P, Barton C: Is chivalry dead? J Commun 28:57–64, 1975.
10. Frodi A, Macaulay J, Thome PR: Are women always less aggressive than men? Psychol Bull
84:634–660, 1977.
11. Sandmaier M: The Invisible Alcoholics: Women and Alcohol Abuse in America. New York, McGraw-
Hill, 1980.
12. Goodchilds JD, Zellman GL: Sexual signaling and sexual aggression in adolescent relation-
ships, in Malamuth NM, Donnerstein E (eds): Pornography and Sexual Aggression. Orlando, FL,
Academic Press, 1984, pp 233–243.
13. Koss MP, Dinero TE: Discriminant analysis of risk factors for sexual victimization among a
national sample of college women. J Consult Clin Psychol 57:242–250, 1989.
14. Taylor SP, Leonard KE: Alcohol and human physical aggression, in Geen RG, Donnerstein EI
(eds): Aggression: Theoretical and Empirical Reviews, vol 2, Issues in Research. New York, Aca-
demic Press, 1983, pp 77–101.
15. Leonard KE: Alcohol consumption and escalatory aggression in intoxicated and sober dyads.
J Stud Alcohol 45:75–80, 1984.
16. Leonard KE: Drinking patterns and intoxication in marital violence: Review, critique, and
future directions for research, in National Institute on Alcohol Abuse and Alcoholism Re-
search Monograph-24 (ed): Alcohol and Interpersonal Violence: Fostering Multidisciplinary Perspec-
tives. NIH Publication No. 93-3496. Rockville, MD, NIAAA, pp 253–280.
17. Miller BA, Downs WR: The impact of family violence on the use of alcohol by women. Alcohol
Health Res World 17(2):137–143, 1993.
18. Miller BA, Downs WR, Testa M: Relationship between women’s alcohol problems and experi-
ences of childhood violence. Paper presented at the annual convention of the American
Psychological Association, Boston, MA, 1990.
of violent events, in National Institute on Alcohol Abuse and Alcoholism Research Mono-
graph-24 (ed): Alcohol and Interpersonal Violence: Fostering Multidisciplinary Perspectives. NIH
Publication No. 93-3496. Rockville, MD, NIAAA, 1993, pp 161–191.
20. Kaufman Kantor G, Asdigian N: Socialization to alcohol-related family violence: Disentan-
gling the effects of family history on current violence. Paper presented at the Annual Meet-
ings of the American Society of Criminology, Phoenix, AZ, 1993.
21. Pernanen K: Theoretical aspects of the relationship between alcohol and crime, in Collins JJ
(ed): Drinking and Crime: Perspectives on the Relationship between Alcohol and Consumption and
22. Pernanen K: Alcohol and Human Violence. New York, Guilford Press, 1991.
23. Pernanen K: Research approaches in the study of alcohol-related violence. Alcohol Health Res
World 17(2):101–107, 1993.
24. Bersani CA, Chen H: Sociological perspectives in family violence, in Van Hasselt VB, Mor-
rison RL, Bellack AS, Hersen M (eds): Handbook of Family Violence. New York, Plenum Press,
1988, pp 57–86.
25. Gelles RJ, Straus MA: Determinants of violence in the family: Towards a theoretical integra-
tion, in Burr WR, Hill R, Nye FI, Reiss IL (eds): Contemporary Theories About the Family. New
York, Free Press, 1979, chap 21.
26. Hamburger LK, Hastings JE: Personality correlates of men who abuse their partners: A cross
validation study. J Fam Viol 1(4):323–341, 1986.
27. Hudson WW, McIntosh SR: The assessment of spouse abuse: Two quantifiable dimensions.
J Marriage Fam 43:873–885, 1981.
28. Roy M: A current study of 150 cases, in Roy M (ed): Battered Women: A Psychosocial Study of
Domestic Violence. New York, Van Nostrand Reinhold, 1977, pp 25–44.
29. Holtzworth-Munroe A, Stuart GL: Typologies of male batterers: Three subtypes and the
differences among them. Psychol Bull 116:476–497, 1994.
30. Straus MA: A general systems theory approach to a theory of violence between family
members. Soc Sci Info 12:105–125, 1973.
31. Kaufman Kantor G, Jasinski JL, Aldarondo E: Socioeconomic status and incidence of marital
violence in Hispanic families. Violence Vict 9(3):207–222, 1994.
32. Pagelow MD: Family Violence. New York, Praeger, 1984.
33. Yllo KA, Straus MA: Patriarchy and violence against wives: The impact of structural and
normative factors. J Int Comp Soc Welf 1:16–29, 1984.
34. Kalmus D: The intergenerational transmission of marital aggression. J Marriage Fam 46:11–19,
1984.
35. O’Leary KD: Physical aggression between spouses: A social learning theory perspective, in
Van Hasselt VB, Morrison RL, Bellack AS, Hersen M (eds): The Handbook of Family Violence.
New York, Plenum Press, 1988, pp 31–56.
Newbury Park, CA, Sage, 1980.
37. Hotaling G, Sugarman D: An analysis of risk markers in husband to wife violence: The
current state of knowledge. Violence Vict 1(2):101–124, 1986.
38. Dobash RE, Dobash RP: Wives: The appropriate victims of marital violence. Victimology
1:416–441, 1978.
39. LaRossa R And we haven’t had any problems since: Conjugal violence and the politics of
marriage, in Straus MA, Hotaling GT (eds): The Social Causes of Husband-Wife Violence. Min-
neapolis, University of Minnesota Press, 1980, pp 157–175.
40. Leonard KE, Blane HT: Alcohol and marital aggression in a national sample of young men.
J Interpers Violence 7(1):19–30, 1992.
RL, Bellack AS, Hersen M (eds): Handbook of Family Violence. New York, Plenum Press, 1988,
pp 383–406.
42. Kaufman Kantor G: Parental drinking, violence and child aggression. Paper presented at the
American Psychological Association Annual Meetings: Mini-Conference on Substance Abuse
and Violence, Boston, 1990.
44. Gustafson R Threat as a determinant of alcohol-related aggression. Psychol Rep 58:287–297,
1986.
45. Pihl RO, Zeichner A, Niuara R, et al: Attribution and alcohol-mediated aggression. J Abnorm
Psychol 90(5):468–475, 1981.
46. Schmutte GT, Leonard KE, Taylor SP: Alcohol and expectation of attack. Psychol Rep 45:163–
167, 1979.
47. Powers RJ, Kutash IL: Alcohol, drugs, and partner abuse, in Roy M (ed): The Abusive Partner.
New York, Van Nostrand Reinhold, 1982, pp 39–75.
49. Goldstein PJ, Belluci PA, Spunt BJ, Miller T: Frequency of Cocaine Use and Violence: A Comparison
between Men and Women. Unpublished, New York, Narcotic and Drug Research, 1988.
50. Taylor SP, Chermack ST: Alcohol, drugs and human physical aggression. J Stud Alcohol 11:78–
88, 1993.
51. Pihl RO, Peterson JB, Lau MA: A biosocial model of the alcohol-aggression relationship.
J Stud Alcohol 11:128–139, 1993.
52. Frieze IH, Schafer PC: Alcohol use and marital violence: Female and male differences in
reactions to alcohol, in Wilsnack SC, Beckman LJ (eds): Alcohol Problems in Women. New York,
Guilford Press, 1984, pp 260–279.
53. Critchlow-Leigh BC: The powers of John Barleycorn: Beliefs about the effects of alcohol on
social behavior. Am Psychol 41(7):751–764, 1986.
54. Critchlow-Leigh BC: Beliefs about the effects of alcohol on self and others. J Stud Alcohol
48(5):467–475, 1987.
55. Brown SA, Goldman MS, Inn A, Anderson LR: Expectations of reinforcement from alcohol:
Their domain and relation to drinking patterns. J Consult Clin Psychol 48:419–426, 1980.
56. Rohsenhow DJ: Drinking habits and expectancies about alcohol’s effects for self versus oth-
ers. J Consult Clin Psychol 51(5):752–756, 1983.
57. Rohsenhow DJ, Bachorowski JA: Effects of alcohol and expectancies on verbal aggression in
men and women. J Abnorm Psychol 93:418–432, 1984.
58. Kline RB: The relation of alcohol expectancies to drinking patterns among alcoholics: Gener-
alization across gender and race. J Stud Alcohol 51:175–182, 1990.
59. Kaufman Kantor G, Asdigian N: Gender differences in alcohol-related spousal aggression, in
Wilsnack RW, Wilsnack SC (eds): Gender and Alcohol. New Brunswick, NJ: Rutgers University
Press, 1995, in press.
60. Straus MA, Kaufman Kantor G: Corporal punishment of adolescents by parents: A risk factor
in the epidemiology of depression, suicide, alcohol abuse, child abuse, and wife beating.
Adolescence 29(115):543–562, 1994.
61. Widom CS: Child abuse and alcohol use and abuse, in National Institute on Alcohol Abuse
and Alcoholism Research Monograph-24 (ed): Alcohol and Interpersonal Violence: Fostering Mul-
tidisciplinary Perspectives. NIH Publication No. 93-3496. Rockville, MD, 1993, pp 291–314.
62. Schuckit MA, Tipp JE, Kelner E: Are daughters of alcoholics more likely to marry alcoholics?
63. Abram KM, Teplin LA, McClelland GM: The effects of co-occurring disorders on the relation-
ship between alcoholism and violent crime: A 3-year followup of male jail detainees, in
National Institute on Alcohol Abuse and Alcoholism Research Monograph-24 (ed): Alcohol
and Interpersonal Violence: Fostering Multidisciplinary Perspectives. NIH Publication No. 93-3496.
Rockville, MD, 1993, pp 237–251.
64. Collins JJ: Drinking and violence: An individual offender focus, in National Institute on
Alcohol Abuse and Alcoholism Research Monograph-24 (ed): Alcohol and Interpersonal Vio-
lence: Fostering Multidisciplinary Perspectives. NIH Publication No. 93-3496. Rockville, MD,
1993, pp 221–235.
65. Kaufman E: Psychotherapy of Addicted Persons. New York, Guilford Press, 1994.
66. Wilsnack SC: Patterns and trends in women’s drinking: Recent findings and some implica-
tions for prevention. Paper presented to the Working Group for Prevention Research on
Women and Alcohol, Bethesda, MD, 1993.
67. Wilsnack RW, Wilsnack SC: Husbands and wives as drinking partners. Paper presented at the
16th Annual Alcohol Epidemiology Symposium of the Kettil Bruun Society for Social and
Epidemiological Research on Alcohol, Budapest, Hungary, 1990.
68. Jacob T, Bremer DA: Assortative mating among men and women alcoholics. J Stud Alcohol
47:219–222, 1986.
69. Miller BA, Downs WR, Testa M, Keil A: Thematic analyses of severe spousal violence inci-
dents: Women’s perceptions of their victimization. Paper presented at the Annual Meeting of
the American Society of Criminology, San Francisco, CA, 1991.
70. Jacob T, Ritchey D, Cvitkovic J, Blane H: Communication styles of alcoholic and nonalcoholic
families when drinking and not drinking. J Stud Alcohol 42:466–482, 1981.
71. Wolfgang ME: Patterns in Criminal Homicide. New York, John Wiley, 1958.
72. Goodman RA, Mercy JA, Loya F, et al: Alcohol use and interpersonal violence–alcohol de-
tected in homicide victims. Am J Public Health 76(2):144–149, 1986.
73. Haberman PW, Baden MM: Alcohol, Other Drugs, and Violent Death. New York, Oxford Univer-
sity Press, 1978.
74. Vikkunen M: Alcohol as a factor precipitating aggression and conflict behavior leading too
homicide. Br J Addict 69:149–154, 1974.
25(9):1065–1081, 1990.
76. Koss MP, Dinero TE: Predictors of sexual aggression among a national sample of male college
students. Ann NY Acad Sci 528:1133–1147, 1988.
77. Roizen J: Issues in epidemiology of alcohol and violence, in National Institute on Alcohol
Abuse and Alcoholism Research Monograph-24 (ed): Alcohol and Interpersonal Violence: Fos-
tering Multidisciplinary Perspectives. NIH Publication No. 93-3496. Rockville, MD, 1993,
pp 3–36.
78. Kaufman Kantor G, Jasinski J: Prevention of teen violence: Evaluation of a multidimensional
model. Paper presented at the 4th International Family Violence Conference, Durham, NH,
1995.
79. Kaufman Kantor G, Jasinski J: Family members polydrug use as a risk factor in family vio-
lence. Paper presented at the Research Institute on Additions, Fall Seminar Series, Buffalo,
NY, 1994.
80. Leonard KE, Senchak M: Alcohol and premarital aggression among newlywed couples.
J Stud Alcohol 11:96–108, 1993.
to a random household sample of women. J Stud Alcohol 50(6):533–540, 1989.
82. Miller BA, Downs WR, Testa M, Panek D: The relationship between victim’s perpetrators
alcohol use and spousal violence. Paper presented at the Annual Meeting of the American
Society of Criminology, Baltimore, MD, 1990.
83. Bergman B, Larsson G, Brismar B, Klang M: Battered wives and female alcoholics: A compari-
tive social and psychiatricc study. J Adv Nurs 14:727–734, 1989.
84. Barnett OW, Fagan RW: Alcohol use in male spouse abusers and their female partners. J Fam
Violence 8(1):1–25, 1993.
85. Gelles RJ, Harrop JW: Violence, battering, and psychological distress among women. J Inter-
pers Violence 4:400–420, 1989.
86. Hilberman E, Munson K: Sixty battered women. Victimology 2:460–470, 1977.
87. Stark E, Flitcraft A, Zuckerman D, et al: Wife abuse in the medical setting: An introduction for
health personnel, in: Domestic Violence Monogmph Series No. 7. Rockville, MD, National Clear-
inghouse on Domestic Violence, 1981.
88. Kaufman Kantor G: Refining the brushstrokes in portraits of alcohol and wife assaults, in
National Institute on Alcohol Abuse and Alcoholism Research Monograph-24 (ed): Alcohol
and Interpersonal Violence: Fostering Multidisciplirtary Perspectives. NIHPublicationNo. 93-3496.
Rockville, MD, 1993, pp 281–290.
89. Arias I: A Social Learning Theory Explication of the Intergenerational Transmission of Physical
Aggression in Intimate Heterosexual Relationships. Unpublished doctoral dissertation. Stony
Brook, State University of New York, 1984.
90. Downs WR, Miller BA, Gondoli DM: Childhood experiences of parental physical violence for
alcoholic women as compared with a randomly selected household sample of women. Vio-
lence Vict 2(4):225–240, 1987.
91. Downs WR, Miller BA, Testa M, Panek D: Long-term effects of parent-to-child violence for
women. J Interpers Violence 7(3):365–382, 1992.
14
How Far Have We Come?

A Critical Review of the Research on
Men Who Batter
W. Vernon Lee and Stephen P. Weinstein
Abstract. Although the effects of domestic violence and partner abuse have been known
throughout history, the topic has recently begun to receive attention in the research literature.
Indeed, it was not until 1986 that two interdisciplinary journals were dedicated exclusively to the
study of family violence. Popular lore has frequently cited a connection between substance use
(particularly alcohol) and domestic violence; however, this interaction has now begun to be
formally recognized and addressed in research and treatment paradigms. This chapter presents
some of the research findings regarding the multidimensional relationship between family vio-
lence and alcohol and/or drug abuse. Theories and data about the causes of family violence and
characteristics of the male batterer and of the substance-abusing men who batter are offered. The
final section examines several current models of treatment and their outcome.
1. Introduction
The study of male batterers and their response to treatment interventions is a
relatively new area of exploration. With stories of murder and abuse filling
the pages of our newspapers, the problem of partner abuse has catapulted
into national prominence. Section 2 of this chapter examines research that
attempts to identify the causes of partner abuse and to define what leads
W. Vernon Lee • Penn Recovery Systems, Philadelphia, Pennsylvania 19104-1953. Stephen P.

Weinstein • Department of Psychiatry and Human Behavior, Thomas Jefferson University, Phila-
delphia, Pennsylvania 19107.
337
some men to batter their partners while others do not. Section 3 focuses on
current research regarding characteristics of men who batter and identifying
psychological and behavioral characteristics of batterers. Section 4 examines
some of the characteristics of substance-abusing men who batter, while seek-
ing a connection between battering and chemical abuse. These two disorders
challenge the clinician to find new strategies for working with the male bat-
terer. With the spread of batterers’ programs and mandated treatment for
batterers, the final section examines current methods of treatment and their
outcomes in an effort to determine if anything has worked to turn the tide of
male violence against intimate partners.
2. Causes of Domestic Violence

Society has frequently been accused of being the source of male violence
against their intimate partners.1,2 Some feminists consider aggression toward
women to be a manifestation of a patriarchal social system that seeks to
discriminate against women and perpetuates the myth that men are superior
to women. They maintain that the belief structure used to repress women is
conveyed in attitudes such as sexual conservatism, adversarial sexual beliefs,
and acceptance of interpersonal violence. While sexist attitudes are them-
selves not purported to be a sufficient condition for the occurrence of aggres-
sion against women, their presence increases the likelihood of women being
victimized.3,4
Another point of view is offered by supporters of the psychodynamic
perspective who contend that personality factors specific to the individual
facilitate the commission of aggressive acts. The batterer utilizes violence as a
means to deal with deep-seated resentments and fears that may or may not be
available to the batterer’s conscious thought processes. Gondolf5 summarized
this perspective by stating that men act aggressively against women in order
to express infantile hostility and womb envy; such men simultaneously hate
and fear women.
Although feminist and psychodynamic theorists recognize hostility as a
central aspect of aggression against women by men, in isolation each theoreti-
cal perspective fails to address critical aspects of the issue.6 On the one hand,
feminist theory does not address why two men from similar environments do
not both offend, while on the other hand the psychodynamic approach is
insufficient to explain why two men with roughly similar personality traits do
not both abuse women. Further, neither theory adequately accounts for the
individual man’s choice of offense style, i.e., battering or verbal or sexual
abuse. Dewhurst et al.6 suggested that aggression against women by men
may be facilitated by the misogynist values of a male-dominated culture, as
feminists would argue. But the dynamics that influence an individual’s choice
of abuse style may be more related to the personality characteristics of the
offender.
14 • Research on Men Who Batter 339
In addition to theories that suggest that it is the culture and/or person-

ality characteristics of the batterer that cause abuse, it has also been suggested
that men may become batterers after witnessing abuse in their family of
origin.7 Still another more controversial position is that a man abuses because
his partner pressures him into hitting her. This perspective has not been
validated by experimental study. A most controversial position is taken by
proponents of what has come to be called “the battered-husband syndrome.”
These theorists contend that men are frequently victims in their homes and
view the women as wielding power over them, which leads the men to act out
abusively.8
This entire issue of a reversal of the classic male dominance (i.e., role
reversal) has been studied. Some experimental research has indicated that
when the traditional power hierarchy of husband being in the one-up posi-
tion over his wife is challenged in a marriage, abuse is more likely to occur.
This is particularly true when a woman appears to have more status than her
partner because of a higher educational level or occupation.9 This perspective
is in direct contrast with the “battered-woman syndrome,” which attributes
abuse to a woman’s lack of power and her role as a “victim” in the home.10–12
These and other studies are inconclusive and leave many questions un-
answered regarding the relationship of power in battering relationships.
Given the relative paucity of research on men identified as batterers and
the controversy regarding power in abusive relationships, Claes and Rosen-
thal9 examined the conflict tactics used by men arrested for domestic assault
and their perception of their partner’s power. The results of their study indi-
cated that the severity of abuse toward a female partner is related to the
batterer’s perception of the partner’s power to reward. Men who used the
most severe violent tactics perceived their partners as having high rewarding
power. Claes and Rosenthal’s explanation of this result is that battering males
look to their partners in a dependent fashion to provide their primary person-
al support, and therefore perceive their partner as having high rewarding
power or a level of control that they cannot accept, leading to an abusive
cycle.
In summary, this section explored some of the possible causes of partner
abuse. Feminist and psychodynamic theoretical perspectives as well as the
sociocultural and personality researchers each contribute valuable informa-
tion to this ongoing debate. No one theory has fully explained the causes of
male violence against intimate partners. Men do it for different reasons, and
some whom you might expect to abuse, because of criminal history or family
background, do not abuse at all.
3. Characteristics of Male Batterers

The efforts to define the characteristics unique to male batterers have
produced somewhat inconclusive and often contradictory results.13–15 Much
of the literature directed at describing the batterer has been derived from
three theoretical models or approaches designed to facilitate an understand-
ing of battering. Broadly defined, these are personality–behavioral models,
social learning approaches, and psychodynamic explanations. The person-
ality–behavioral depictions of battering are generally based on empirical as-
sessments, behavioral observations, and reports or combinations of the two.
Social learning approaches are primarily presented as observations fitting a
social learning model of behavior. Psychodynamic explanations of battering
are presented in a framework of modified Freudian theory.
3.1. Personality Behavioral Models

Wodarski11 reports that batterers in an abusive family system tend to
have the following general personality and behavioral characteristics: (1) their
behavior was learned from the family of origin, and violence occurs in the
family of origin among batterers much more frequently than among nonbat-
terers; (2) they project blame for the abuse onto their victim, i.e., batterers
externalize the blame; (3) they are highly possessive of their victim, creating
problems around jealousy and control; (4) the violence directed at an intimate
partner can represent an attempt to displace anger meant for authority fig-
ures; (5) they have inappropriate expectations around their partners' role in
the relationship and a distorted perception of their victims' behavior toward
them; and (6) they often cannot recall details surrounding abuse incidents.11
Among the more specific personality characteristics of batterers most fre-
quently cited in the research are indications that batterers are less assertive,
more hostile, more abusive of alcohol, lower in self-esteem, and more likely
to show identifiable personality disorder than their nonbatterer counter-
parts.7,9,16
Researchers have begun to develop multiple-category typologies de-
signed to tease out and explain characteristics of batterers, based either on the
batterer's personality or behavior. Several of the current existing typologies
based on the batterer's personality/behavioral style have evolved from both
clinical and empirical assessments of batterers in treatment. Utilizing clinical
observations to define batterer's emotional needs, Elbow17 described four
types of batterers: (1) the “controller,” who views his wife as an object to
control; (2) the “defender,” who believes he must overprotect his wife; (3) the
“approval seeker,” who makes excessive demands in order to reinforce
his self-image; and (4) the “incorporator,” a type of batterer who needs his
partner to validate and define himself. Ceasar,18 in an effort to empirically
discriminate the batter from the nonviolent therapy patient, found no differ-
ences on the Minnesota Multiphasic Personality Inventory (MMPI). How-
ever, she was able to develop a four-category typology of batterers based on
extensive clinical interviews. Ceasar18 titled them the “tyrant,” the “exposed
rescuer,” the “nonexposed altruist,” and the “psychotic wife assaulter. ” The
tyrant subgroup was described as self-centered, hostile, and paranoid and
less likely to be arrested than other wife-abuser subgroups. The exposed

rescuers were described as having hysterical personalities and alternating
between sociability and hostility. The nonexposed altruists were unassertive
and constantly trying to please their wives; consequently, they felt unappreci-
ated and victimized. The psychotic wife assaulter’s anger does not stem from
any rationally defineable source and appears erratic.
Hamberger and Hastings19 and Hastings and Hamberger20 attempted to
formulate an empirically based typology of batterers. Male batterers (N = 99)
were recruited from a treatment program, along with 71 help-seeking nonbat-
terers recruited from local marriage and family therapy clinics. All subjects
were tested with the Millon Clinical Multiaxial Inventory-I.21 Factor analyses
indicated three major personality “factors” common to the partner abusers:
factor 1, borderline/schizoidal; factor 2, histrionic/narcissistic; and, factor 3,
antisocialharcissistic. There was very little overlap in the factor structure of
the batterers and nonbatterers, suggesting that qualitatively they represent
largely different populations of men.
Saunders22 used a combination of attitude and behavior measures in a
cluster analysis to study batterers. Scores for depression, anger toward part-
ner, and attitudes toward women were derived, along with measures of
generalized violence, severity of partner abuse, and alcohol use. One cluster
(N = 31) was characterized as being “emotionally volatile,” scoring higher on
measures of anger, depression, and jealousy, while demonstrating moderate
levels of violence and alcohol use. The batterers in the second cluster (N = 48)
were more likely to score low on the anger and depression measures and
were the most likely to be violent outside the home. A third cluster (N = 86)
appeared to be composed largely of batterers who suppressed their feelings,
while confining their violence primarily to home; that is, they scored low on
anger and depression measures but the highest on a scale for social desir-
ability. Gondolf13 points out that it is still unclear whether the “feelings” of
anger and depression are merely associated with different levels of expression
of violence or are in themselves causative of the violence. It is also unclear if
the “types” as empirically assessed actually describe men in different phases
of abuse.
Gondolf13 indicated that the previous research attempting to demon-
strate either behavioral or personality typologies had too many methodologi-
cal (sampling, instrumentation) problems and lacked a theoretical frame of
reference. He believed that the shortcomings of the empirically based person-
ality typologies outweighed those of the more descriptive behavioral typolo
gies. Assessed attitudinal or personality characteristics are insufficient to ex-
plain or predict violent behavior. In his view, behavioral patterns are well
established and less varied than the complex personality attributes associated
with batterers. Therefore, Gondolf argues, because of their destructive impact,
it is the behavior of batterers that is of more immediate concern. Behavior is
more likely to be predictive of future violence, and thus has direct implications
for the victims and for the extent and effectiveness of interventions.
As a result, Gondolf13 attempted to establish a behavioral typology of

batterers by focusing on the descriptions offered by the victims of abuse. His
typology was based on the self-report of battered women admitted to Texas
shelters during an 18-month period, from 1984–1985 (N = 6000). The typol-
ogy he developed does not represent diagnostic categories but rather an
interpretable summary of cluster findings. The batterer “types” include two
highly antisocial and severely abusive types, one of which is typified by an
unusually high number of arrests for general violence and substance abuse,
and a third less antisocial and less abusive type. He labeled the most antiso-
cial as “sociopathic,” given the extreme criminal nature of the behaviors. The
sociopathic batterer type is most likely to have been previously arrested for
property, violent, and drug- or alcohol-related crime. The second antisocial
type was labeled the “antisocial” batterer. This type of batterer is also ex-
tremely abusive physically and verbally. While he is likely to have been gen-
erally violent, he is less likely to have been arrested than the sociopathic
batterer. The third less severe type was referred to as the “typical” batterer,
who may exhibit either chronic or sporadic patterns of abuse. The typical
batterer conforms more to the prevailing clinical profiles of batterers. He is
particularly less likely to have used weapons than the other types of batterers.
These findings substantiate earlier studies based on victims’ reports, which
point to the presence of a ”violence-prone” personality and an association
between the severity of partner abuse and general violence. This also con-
firms recent research on personality types that include a sociopathic/nar-
cissistic type of batterer, whose abuse is not anger driven.13,20
Notwithstanding Gondolf ’s criticism of personality typologies, research
in the 1990s continues to search for typologies that characterize personality
traits and psychological and cognitive characteristics of men who batter wom-
en in order to distinguish them from nonbattering men. Else et al.16 compared
a group of 21 batterers with a comparable group of nonbatterers recruited as a
control group from the local community. Utilizing the MMPI and its person-
ality disorder scales (MMPI-PDS) as well as the Hostility and Direction of
Hostility Questionnaire, Else’s results showed that batterers score higher on
the borderline and antisocial MMPI-PDS and on the acting-out hostility and
self-criticism scales of the hostility questionnaire. No significant differences
were found between the groups’ performances on cognitive measures, de-
pression scale scores, or overall scores on the MMPI. While finding very few
differences on the MMPI, he did find that child abuse was a common experi-
ence among batterers.
Murphy et al.23 examined associations between family of origin violence,
self-reports of psychopathology on the Millon Clinical Multiaxial Inventory-
II,21 and current spouse abuse among partner assaultive men. Compared to
nonviolent men in discordant (N = 24) and well-adjusted (N = 24) relation-
ships, partner-assaultive men (N = 24) were significantly more likely to report
childhood histories of physical abuse and physical abuse of the mother in the
family of origin. The partner-assaultive men also reported significantly higher
scores on a variety of Millon Clinical Multiaxial Inventory-II personality disor-

der and Axis I disorder scales.
Also in search of an explanation for violence in relationships, Wesner et
al.24 examined “explosive rage” in 51 cases of troubled marriages. Physical
violence not involving direct partner abuse (i.e., violence against inanimate
objects not causing injury to persons) had received little attention in the
research literature. However, Wesner et al.24 were impressed with the fre-
quent appearance of several common features to this reaction. Explosive rage
seemed to occur most often in the presence of a female friend or spouse, with
alcohol being consumed, and in the absence of another person other than a
child (family member or male friend). The study strongly suggested a role for
alcohol in sensitizing trigger conditions (it was present in 82% of the sample),
and that the presence of another person not a spouse or girlfriend may exert
an inhibiting effect. In this study, 24% of the sample who engaged in explo-
sive rage involved actual battering at some point in the history of the relation-
ship.
Taking a different route toward attempting to develop an understanding
of the characteristics of the male batterer, Bergman and Brismar25 hypothe-
sized that the same background and precipitating factors would be found in
both male victims and male assailants, and that these factors could predeter-
mine the social career for selected males: that of assailant, victim, or both. The
researchers chose two highly selected and not necessarily representative
groups of assailants and victims. Batterers (MB) included men sent to prison
for wife abuse, a group that constitutes only a limited proportion of all men
who beat their wives or partners. The men identified as “victim” (BM) had
such severe injuries from battering that they required inpatient care; how-
ever, they represent only a small portion of all men who are injured in fights.
Bergman and Brismar’s25 study groups consisted of 18 male batterers and 19
battered males and two control groups selected from the general population
of Stockholm. The control groups, one for the MB group and one for the BM
group, also comprised 18 and 19 men, respectively. The authors believed that
the strength of studying these highly selected groups was that any specific
characteristics, either social or psychological, would probably be more easily
identifiable in these extreme groups.
Although the sample sizes were relatively small, results showed that a
large majority of the MB group (15 of 18) and nearly half (9 of 19) of the BM
group had been abused as a child. Half of the men (9) in the MB group and 7
of the 19 men in the BM group reported that they were alcohol-dependent.
Both the victims and the batterers had a higher proportion of divorces and a
smaller proportion of being married than the controls.
Dutton and Hart26 assessed the prevalence of risk markers for partner
abuse in federal offenders. Evidence suggested that this population may dem-
onstrate high rates of family violence, since demographic and psychological
factors commonly found in incarcerated populations are also mentioned in
most of the literature describing assaultive males. Their study found evidence
of high risk for wife assault in this incarcerated population. Specific risk
markers included abuse in the family of origin, personality disorders, and
prior record of violence.
Bersani et al.27 also studied offenders convicted of partner abuse to deter-
mine if specific psychological variables could be used to create a personal
profile. The revised Taylor-Johnson Temperament Analysis (T-JTA) was used.
This 180-item inventory is divided equally among nine bipolar traits. Based on
these traits, a constellation of behavior patterns were derived. Two factors
were identified for the offenders studied: internal or emotional balance and
social interaction. Under factor 1—internal or emotional balance—batterers
registered high on the Nervous, Depressive, Indifferent, Subjective, Hostile,
and Impulsive subscales. This profile represents men, who as a group, are
conflicted internally. However, it was the negative implications of four of the
six subscale measures that Bersani et al.27 believed to be most significant for
this population: Indifferent—referring to the incapacity to see another’s point
of view or identify with another’s feelings; Subjective—indicating internal
turmoil over unresolved difficulties; Hostile—indicating outwardly expressed
anger; and Impulsive—indicating the tendency to react without reason.
For factor 2, social interaction, the batterers registered high on the Ac-
tive/Social, Expressive/Responsive, and Dominant subscales. This profile
represents a picture of social attributes that are acceptable (even rewarded) in
our society. The T-JTA indicated that these men, as a group, prefer social
interaction (Active/Social) and are highly outgoing (Expressive/Responsive).
The added measure of Dominance, or assertiveness, completes a combined
profile of social interaction that is both admired and required in many occupa-
tional contexts. While this extroverted profile seems inconsistent with the
notion that abusing men are isolated or introverted, it may be characteristic of
the educated, professional, socially successful batterer not frequently studied
in research situations.
Pistole and Tarrant28 also studied men convicted of partner abuse. As
part of his probation, each male participated in a psychoeducational group
designed to improve anger management skills. They examined the relation-
ship between self-reported attachment styles and hostility in intimate rela-
tionships. Attachment styles defined as dismissing, fearful, preoccupied, and
secure were represented in proportions that were similar to those found in
nonviolent samples. Multivariate analysis of variance revealed significant dif-
ferences among the attachment styles displaying resentment, suspiciousness,
and guilt. Attachment and aggression have also been linked to shame. Future
research might explore the empirical relationship of shame and attachment in
male batterers. In addition, measurement at the time of the violence might
reveal differences in attachment behavior and aggression that were not elic-
ited in this study.
Dewhurst et al.6 compared the personality traits, attitudes, and beliefs of
men who batter, men who perpetrate sexual assault, and nonabusive men in
an effort to facilitate the development of a clearer picture of the abusive males
choice process. They found that the attitudes and personality characteristics
of men identified as batterers and men who perpetrate sexual assaults against
adult women were similar and that they differed from a community sample of
men who did not self-disclose violence against women. The attitudes and
personality characteristics of men from the community who voluntarily self-
disclosed physical or sexually assaultive behavior toward women and men
who have been formally identified as perpetrators were similar; however,
they, too, differed from nondisclosing community men.
Davidovich29 explored psychological variables that have been identified
as characteristic of males who physically abuse their partners in an attempt to
determine which psychological variables could explain the actions of male
batterers in the context of broader theories of behavior. Her most cogent
explanation of battering behavior was drawn from social learning theory.
3.2. Social Learning Approaches

Reviewing earlier research on the interaction of multiple stressors and
social learning theory offered by Hofeller,30 Hotaling and Sugarman,14 Mac-
Ewen and Barling,31 and Hastings and Hamberger,20 Davidovich29 noted that
most of the reports were inconclusive. However, two important findings did
emerge from the Hotaling and Sugarman14 study. Witnessing parental vio-
lence as a child or adolescent was found to be a consistent risk marker for
adult violence, while personally experiencing physical abuse as a child
showed inconsistency as a risk marker.
Hastings and Hamberger,20 confirming Hotaling and Sugarmans’14 re-
sults, also found that males who witnessed abuse as children were more
likely to engage in partner abuse. They found, however, that personally
experiencing abuse as a child also led to a greater likelihood of becoming an
adult abuser. In addition, these researchers noted that children who either
witnessed or experienced abuse and who later became adult batterers were
also more likely to have a concurrent alcohol problem. These abusers tend to
come from dysfunctional families, often from alcoholic and/or drug-abusing
parents.
In Davidovich’s29 discussion of violence against women, she describes
batterers as young, generally between the ages of 18 and 34 years old; 54% are
white, 44% are African American, 47% are unemployed, 40% have some prior
criminal record; 60% being intoxicated during the attack on their wives, and
70% are under the influence of drugs and–or alcohol during their violent
episode. Her application of social learning theory to the batterer and to
spouse abuse begins with an individual who learns to respond violently to
stressful situations through modeling of the behavior of family members,
peers, and significant others. When, as an adult, he is confronted with
stressor stimuli, he becomes frustrated and resorts to violence as an appropri-
ate and justified action. Although this theory seems to present a compelling
argument for generational patterns of family violence, Davidovich notes that
the process of translating life events and stressors into violence is not yet
known. Davidovich29 argues that knowledge of the “psychology of the indi-
vidual and personality variables might serve to fill the gap between situation-
al factors and differential responses” (p. 36).
Davidovich29 concluded that the weight of the evidence seems to suggest
that either witnessing and/or experiencing violence in one’s family of origin
strongly increased the likelihood that an individual would engage in spouse
abuse. Furthermore, dysfunctional families of origin, particularly those in
which alcoholism and/or drug addiction were prevalent, play an important
role in the intergenerational transmission of violence.
3.3. Psychodynamic Explanations

Although Davidovich29 indicated that another approach to understand-
ing abusive behavior could be derived from psychoanalytic theory, her
psychodynamic explanations were more descriptive of the psychosocial di-
mensions defining the males who batter. Guerney et al.32 offered a psycho-
dynamic theoretical construct to explain the development and maintenance of
wife battering and provided a rationale for the type of therapy deemed most
likely to be effective with wife batterers. Their constructs are tied to research
evidence along with clinical observations. Although psychodynamic in na-
ture, the explanation was conceived in terms of Dollard and Miller’s33 integra-
tion of Freudian concepts with learning theory.
In their theoretical approach, being abused as a child gives rise to feelings
of rage, but to express this rage to the abuser is dangerous, and the rage is
suppressed or repressed. As is the case with any extremely strong feeling
denied full expression at the time of its occurrence, the individual seeks ways
to express the rage in ways that are relatively safe (e.g., dreams). The inability
to escape from the abuse in childhood leads to a deep sense of powerlessness
that the individual seeks to overcome, often through the artificial sense of
power derived from alcohol or by attempting to exert power/control over
others. However, because of the fear of authority figures that generalizes
from fear of an abusing parent, the attempt to gain power and control over
others is directed mainly toward those who are physically weaker than the
self (e.g., a child and/or wife).
At the same time the individual is attempting to overcome the powerless-
ness and express rage, there is a conflict between fear of one’s rage and the
fear of loss of affection from a loved one. Therefore, the effort to control
others and/or to express rage against others is likely to be inconsistent and
sporadic, often alternating with genuine expressions of concern and affection.
Within this context, while seeking an outlet for the feelings of rage, the
batterer can readily interpret or engineer events in his marital relationship in
such a way as to provide an excuse for expressing his anger. And finally,
since the feelings of rage frequently need to be repressed and repression
requires significant psychological effort, any general disinhibitor (such as
alcohol) strongly increases the probability that control will be lost and rage
overtly expressed.
It appears that personality, behavioral, social learning, and psycho-
dynamic perspectives provide valuable information on characteristics of the
male batterer, and perhaps they should be seen as complementary rather
than mutually exclusive and exhaustive. In order to understand battering, it is
necessary to look at the batterers themselves to comprehend the process of
translating stressors and life events into acts of violence within the socio-
cultural context in which the batterer lives.
In conclusion, research in the 1990s continues to demonstrate that al-
though partner abuse is not strictly limited to men with personality problems
or other identifiable psychological problems, such men do seem to constitute
a larger proportion of the identified treatment population, especially those
with concurrent substance abuse problems. It is clear from the above-cited
studies that men who seek or are forced into treatment as batterers are likely
to have psychological problems that can affect their treatment. The role that
drugs and alcohol play in the mix with personality, sociocultural, and behav-
ioral factors needs further exploration. The influence of substance of abuse
also requires further clarification because presumably nonalcoholic or non-
substance-abusing battering samples may be confounded by the presence of
undetected alcohol or drug problems.
The evidence presented thus far strongly suggests heterogeneity of char-
acteristics of men who batter. Because of this heterogeneity, many research-
ers have identified different specific clusters of psychological and/or behav-
ioral characteristics that they believe identify male batterers. Such analyses
may be helpful if they lead to theoretical formulations, provided they also
recognize the importance of the sociocultural context of violence. Physical
abuse directed at women has only recently begun to be been addressed in
national policy and legislation. There is still relatively little concrete knowl-
edge about the batterer and little real information about the prevalence, sever-
ity, and outcome of violence perpetrated by men against female partners. The
above-cited research describing a variety of background, behavioral, psycho-
logical, and social characteristics of the batterer and suggesting that no one
pathology can be linked to battering is a beginning in the effort to address this
scarcity of information.
4. Substance Abuse and Battering

Although many batterers are alcoholic and/or substance abusers, the role
of alcohol and other substances in domestic violence remains controversial.
At present, neither alcohol nor other drugs are considered necessary or suffi-
cient for the occurrence of domestic violence. However, alcohol and/or drugs
have been found to be present in many domestic violence cases. Estimates of
the percentage of batterers who assault their partners while intoxicated range
from 60 to 70% for alcohol abuse and 13 to 20% for drug abuse.34,35 Alcohol
and drug use are often invoked by batterers to excuse or explain their abusive
behavior. In fact, these substances may well serve a disinhibiting or releasing
function, which may be associated with violence.36
There is a general consensus that alcohol is the drug that is most often
associated with violence.37 According to Miller et al.,38 this is probably due to
the high base rate of alcohol use in the population rather than to anger-
inducing effects of ethanol. Alcohol use can trigger aggression during either
intoxication or withdrawal. The linkage is best viewed as a mixture of person-
ality, social, and physiological causes.39–42 Alcohol consumption under some
circumstances may be associated with an increase in aggressive behavior,42–47
acts of criminal violence,48,49 and assaultive behaviors.50,51 Other studies have
reported a more specific association between drinking and marital violence.
Between 30 and 70% of battered wives report problem drinking or alcoholic
husbands.52–54 Eberle55 reported that victims of alcohol-abusing batterers
were themselves more likely to abuse alcohol than were victims of batterers
who did not use alcohol. Byles56 noted a significant association between
alcohol abuse and spouse battering in 139 people appearing in Family Court.
In studying causes of divorce, Cleek and Pearson57 found a strong association
between alcohol abuse and spouse battering.
Other substances of abuse have also been studied in relation to their
propensity to lead to violence. These have included amphetamines, PCP,
barbiturates, heroin, and most recently crack cocaine.58–63
Roberts64 reported the results of an exploratory study into the relation-
ship between battering and substance abuse. This research was important
since little had been done on the incidence of combined drug and alcohol
abuse among men who batter. Studying the intake records of 234 male bat-
terers who had charges filed against them, Roberts found that men who
committed more serious battery offenses (i.e., an attack resulting in bodily
injury) were significantly more likely than those who committed lesser bat-
tery offenses (i.e., slapping, pushing, etc.) to have either a drug problem or a
dual problem with alcohol and drugs. Almost one half (49.9%) of the serious
offenders abused drugs either alone or in combination with alcohol, while
under one fourth (22.4%) of the lesser offenders had this problem. One sur-
prising finding was that the results did not support previous research that
suggested that more severe violence is associated with the batterer’s abuse of
alcohol.12,34,35 In the Roberts,64 study, the number of alcohol abusers in the
two groups was almost equal: 28 were charged with serious battering and 29
were charged with the lesser offense. However, severe physical abuse was
found to be associated with drug abuse and the dual problem of alcohol and
drug abuse, not solely with alcohol use. This finding suggests the need to
establish specific criteria and assessment methods for determining the nature
and extent of a substance abuse problem that could be “predictive” of ones’
battering behavior.
Miller34 also reported that a combination of drug and alcohol abuse is
more predictive of domestic violence than is alcohol use alone. The subjects
(N = 82) were selected from all male parolees in western New York who were
convicted of either nonviolent or violent offenses. They were given the Diag-
nostic Interview Schedule65 and a modified version of the Conflict Tactics
Scale.66 Alcohol problems were present in 76% of the parolees and 73% of the
parolees reported using some type of illegal drug on a regular basis. Rates of
spousal violence were also high for this sample. During 3 months preceding
the interview, 78% of the parolees committed acts of moderate violence, while
33% of the parolees committed acts of severe violence (some committed both
moderate and severe acts of violence). Examining the contribution of the
parolees‘ alcohol and drug problems to their violence toward a spouse re-
vealed that neither drug abuse nor alcohol problems independently contrib-
uted significantly to the degree of violence, but the interaction effect (alcohol
× drugs) contributed significantly to the level of parolee-to-spouse violence.
However, some of their results appear to contradict popular conceptions
of the impact of substance use. For example, alcohol abuse increased the risk
of parolee-to-spouse violence in the absence of drug abuse, but there was a
tendency for alcohol abuse to decrease violence when there was drug abuse.
One possible explanation for this contradictory finding is that the type of drug
use may lessen the violence that was associated with alcohol problems. The
psychopharmacological effects of marijuana and heroin, for instance, have
been attributed to “mellowing out” or causing individuals to “nod out,” re-
spectively, conditions that are likely to lessen violent tendencies. In addition,
considering the popular notion that cocaine use increases the violence rate,
the researchers expected that cocaine users would have an elevated level of
violence or a level of violence at least similar to that for alcohol abusers only.
However, there were more violent activities (parolee-to-spouse) reported for
barbiturate and marijuana users compared to cocaine users.
Our own inner city, publicly funded, intensive cocaine treatment pro-
gram at Thomas Jefferson University has produced some interesting data
about cocaine users who do and do not batter. A battery of standard psycho-
metric instruments and a brief battering questionnaire were administered to
cocaine-dependent men entering the treatment program that identified those
who were partner batters and those who were not. The partner batterers did
not differ significantly from the nonbatterers on demographic measures. Re-
sults showed a longer history of regular weekly drinking reported by the
batterers and a greater number of years of drinking to intoxication by this
same group. Differences were found on the Addiction Severity Index Family
and Psychological composite scores where the batterers scored significantly
higher than the nonbatterers. Eighty-nine percent of the batterers reported a
history of serious conflict with their sexual partners as compared to 62% of the
nonbatterers, and 79% of the batterers reported that they had experienced
trouble controlling violent behavior as compared to 49% of the nonbatterers.
Regarding psychiatric symptomatology, a comparison of SCL-90-R scores in-
dicated that on four of the nine clinical scales, the partner batterers reported
significantly greater disturbance than the nonbatterers. Specifically, the bat-

terers received higher scores on depression, paranoid ideation, psychoticism,
and interpersonal sensitivity. In our clinical–research setting, efforts are now
being directed at developing and evaluating a treatment approach that may
be effective in dealing with the combined problems of cocaine use and partner
abuse.67
Although the connection between substances of abuse and family vio-
lence has been receiving increased public as well as professional atten-
tion,43,60,62 little has been done or written about the treatment of individuals
who are both chemically dependent and partner abusers and guidelines for
their clinical management are yet to be developed.
5. Treatment Interventions
Social workers, psychiatrists, psychologists, and other mental health
workers are increasingly being asked and funded to offer treatment services
to men who batter their partners, even though the causes of this violence, the
motivations for participation in treatment, and the most effective treatments
are not significantly understood. In addition to avoiding prison, a real incen-
tive for men to enter treatment was reported by Gondolf13 who found that
women were more than twice as likely to return to their relationship after
leaving a shelter if their partner was in counseling. This can be a positive
incentive and should be an excellent initiation to treatment, but do efforts
aimed at treating batterers really work? And if they do work, are the effects
sustained over time?
Tolman and Bennett7 reviewed the literature on interventions directed at
helping men who batter become nonabusive. The studies reviewed employed
follow-up periods ranging from a few weeks to several years. Several meth-
odological concerns were raised, including (1) definitions of successful out-
come, (2) data sources, (3) length of the follow-up period, and (4) percentage
of participants actually contacted for follow-up. The 16 studies reviewed con-
sistently indicated that the majority of men stopped their physical abuse
subsequent to intervention. Percentages of successful outcome ranged from
53 to 85%. Lower percentages of success tended to occur (1) in programs with
lengthier follow-up, indicating that effects may dissipate over time, and (2)
when successful outcomes were based on spouses’ reports rather than on
arrests or self-reports.
Unfortunately, few investigators have examined the impact of treatment
beyond a 12-month period. For example, of the 16 evaluations reviewed by
Tolman and Bennett,7 only three reported follow-up data collected more than
12 months after treatment. Edleson and Syers68 went beyond the 12-month
period with their study of the long-term effects of group treatment on men
who batter. They compared three types of brief treatment groups: (1) a self-
help model, (2) an educational model, and (3) a combined model that inte-
grated the other two. Of 283 men randomly assigned to one of the treatment
conditions, 153 completed 80% or more of the assigned sessions. Eighteen-
month follow-up interviews were conducted with 70 program completers or
their partners. Almost two thirds of the men who completed education and
combined groups were reported to be nonviolent at the 18-month follow-up.
The results indicated that short-term, relatively structured group treatment
tended to produce the most consistent successful results. However, there was
no control group and follow-ups were conducted with only a minority (ap-
proximately 25%) of the original sample of men in the study. No information
was available on the large number of missing subjects.
Additional support for the positive outcome cited by Edleson and Syers68
came from Palmer et al.,69 who examined long-term outcomes of a short-term
treatment program for abusive husbands. Recidivism rates, based on police
reports, were found to be lower than those for a control group of untreated
abusive husbands. The researchers also believe that the results contribute to a
relatively unstudied area by empirically attempting to evaluate an unstruc-
tured treatment program. However, the researchers qualify their results by
indicating that the reliability of self-reports given by the batterers did not
always match police reports, and police reports are considered to be a conser-
vative measure of recidivism. The Palmer et al.69 study did not include an
interview with the victimized partner.
Shepard70 designed a study to address batterer recidivism in relation to
community intervention (law enforcement, criminal justice system, and man-
dated counseling programs). This study was unique in that it followed up
men 5 years after they had completed a treatment program in an effort to
discover background and intervention variables that discriminated batterers
who were recidivists from those who were not. A total of 100 men were
included in the study sample. The sample was drawn from the Duluth Do-
mestic Abuse Intervention Project, which was one of the first community
intervention programs in the country. Results showed that 40% of the men
fell into the category of recidivists. Twenty-two percent had been convicted
for domestic assault, 15% had been the subject of a protection order, and 33%
had been police suspects for domestic assault.70
The long-term study also found that alcohol and drug abuse were impor-
tant contributors to recidivism, results reported previously by Eberle55 and
DeMaris and Jackson.71 Batterers who were abused as children also appeared
to be somewhat more likely to be recividists, replicating results reported by
Grusznski.72 Shepard's70 work reflected the need for a stronger emphasis on
preventive measures in the areas of child abuse and chemical dependency.
The above studies were conducted to evaluate predominantly cognitive-
behavioral counseling programs that have reported lower rates of physical
abuse for treated groups, although most were not able to control for other
variables that may have influenced outcome.70 Such issues as severity and
type of abuse may need to be addressed in future studies.73
In addition to evaluating outcomes, researchers and treatment profes-
sionals are also designing new ways of working with the batterer. Three
studies warrant examination because of their unique approaches to battering
treatment.
Geffner74 argues for the use of divorce mediation with abusive couples.
Because of the inherent dangerousness of working with abusers and their
partners, he suggested that a thorough assessment be completed prior to the
mediation to ascertain the extent of the abuse. The counselor then must
determine if the couple needs or requires individual counseling, reconcilia-
tion, or mediation–divorce counseling. If batterer and victim have indepen-
dently stated that they desire mediation (or that it is court mandated), if the
abused partner is willing and able to participate, and if it appears that any
attempts at intimidation would be manageable, then mediation may be an
appropriate approach.74 He offers guidelines for the counselor in the use of
mediation with a discussion of power and control issues (of the batterer) that
may undermine the mediation attempt as well as put the partner in danger.
However, no information is presented on the efficacy of working with the
batterer in couples' sessions or on the outcomes associated with this counsel-
ing approach. Within this context the capacity to “mediate” with a woman
who may be suffering symptoms of battered women's syndrome, a type of
posttraumatic stress disorder where the victim exhibits many symptoms of
fear and intimidation, should be further researched. Mediation may do more
harm than good. The rational, goal-directed thinking necessary for mediation
is also suspect in a man who has demonstrated difficulty with his own depen-
dency needs and the sharing of power and control in his relationships.
Geffner indicates that if mediation is to be used with couples when abuse has
occurred in the marriage, then it is important to overcome the intimidation
and to balance the power.
Wallace and Nosko75 offer treatment guidelines for working with another
central issue when treating abusive men. They believe that shame associated
with their abuse is the core issue for many men who assault their partners and
that these feelings can best be addressed and resolved in group treatment
approaches. However, its immediate impact on the abusive male may be the
reverse. The requirement of having to attend a batterers' group through infor-
mal or formal coercion can be a shame-inducing experience. Furthermore, the
group norms, necessary to achieve safety and deal with feelings, may be
contrary to traditional masculine norms. Group members are required to
disclose their worst incident of violence toward their partner. This interven-
tion becomes the “confession.” No empirical data are offered to substantiate
the efficacy of this treatment intervention, except for transcripts from treat-
ment sessions. Research utilizing control groups that examines how shame is
elicited and how the shame is negotiated through the various stages of group
development is needed.
A third treatment intervention that has gained considerable attention in
the 1990s is the mandatory arrest of batterers. Although there are no studies
in the literature examining its efficacy, this approach has received widespread
criticism. Stark2 defines the debate that has ensued over the past 10 years
with a summary of the interpretations of the proponents’ positions. The
critics who are against mandatory arrest make their case through three points:
(1) mandatory arrest does not work; (2) mandatory arrest is inhumane; and (3)
the very people we are trying to protect do not want it. Stark argues against
these three points. He believes mandatory arrest helps to control behavior:
There is less opportunity for disregarding the battering, minimizing its conse-
quences, and blaming the victims of abuse. Mandatory arrest offers immedi-
ate protection from current violence. Furthermore, Stark believes that arrest
provides a meaningful opportunity for battered women to consider their op-
tions and give those women ready to end their relationship time to go else-
where or to obtain a protection order. Stark further argues that arrest might
deter recidivism and send a clear message that battery is unacceptable.
Mandatory arrest, like other approaches to treating the batterer, is in its
infancy. With the majority of the women returning to their abusive partners,
there is little doubt that mediation and mandatory arrest will gain in popu-
larity. Only the future will tell how effective these efforts have been. The
debate has begun and it is a lively and necessary debate.
6. Conclusions
In this chapter we have attempted to present a distillation of theory and

research regarding the causes of domestic abuse, the identification of the
psychological and behavioral characteristics of batterers in general, and more
specifically the characteristics of substance abusers who batter their partners.
We have also reviewed some of the more traditional as well as the newer
innovative approaches to treating batterers. None of the findings are convinc-
ingly conclusive in either defining the full range of dynamics, behaviors, and
characteristics that describe the population of batterers or in accurately pre-
dicting a clear at-risk battering group. However, these research efforts have
been highly productive in helping to develop a better understanding of the
complex, sometimes violent, tapestry that characterizes human relationships.
Additional research and continuing debate and discussion are needed as
increasingly we recognize the degree to which domestic violence permeates
our society.
References
1. Browne A: Violence against women by male partners: Prevalence, outcomes, and policy
implications. Am Psychol 48(10):1077–1087, 1993.
2. Stark E: Mandatory arrest of batterers. Am Behav Sci 36(5):651–680, 1993.
3. Scott R, Tetreault L: Attitudes of rapists and other violent offenders toward women. J Soc
Psychol 127(4):375–380, 1986.
4. Koss M, Gidyez C, Wisniewski N: The scope of rape: Incidence and prevalence of sexual
aggression an victimization in a national sample of higher education students. J Consult Clin

Psychol 55(2):162–170, 1987.
5. Gondolf E: Fighting for control: A clinical assessment of men who batter. SOC Casework
66(1):48–54, 1985.
6. Dewhurst A, Moore R, Alfano D: Aggression against women by men: Sexual and spousal
assault. J Offender Rehab 18(3/4):39–47, 1992.
7. Tolman R, Bennett L: A review of quantitative research on men who batter. J lnterpers Violence
5(1):87–118, 1990.
8. Steinmetz S: The battered husband syndrome. Victimology 2(3–4):499–509, 1978.
9. Claes J, Rosenthal D: Men who batter women: A study in power. J Fam Violence 5(3):309–322,
1990.
10. Margolin G, Burman B: Wife abuse versus marital violence: Different terminologies, explana-
tions, and solutions. Clin Psychol Rev 13:59–73, 1993.
11. Wodarski J: An examination of spouse abuse: Practice issues for the profession. Clin Soc Work
J 15(2):172–187, 1987.
12. Walker L: The Battered Woman. New York, Harper & Row, 1979.
13. Gondolf E: Who are those guys? Toward a behavioral typology of batterers. Violence Vict
3(3):187–203, 1988.
current state of knowledge. Violence Vict 1(2):101–124, 1986.
15. Edleson J, Eisikovits I, Guttman E: Men who batter women: A critical review of the evidence.
J Fam lssues 6(2):229–247, 1985.
16. Else L, Wonderlich S, Beatty W, Christie D, Staton R: Personality characteristics of men who
physically abuse women. Hosp Commun Psychiatry 44(1):54–58, 1993.
17. Elbow M: Theoretical considerations of violent marriage, Soc Casework 58:515–526, 1977.
18. Ceasar P: Men who batter: A heterogeneous group. Presented at the American Psychological
Association, Washington, DC, 1986.
19. Hamberger K, Hastings JE: Personality correlates of men who abuse their partners: A cross
validation study. J Fam Violence 1(4):323–341, 1986.
20. Hastings J, Hamberger K: Personality correlates of spouse abusers: A controlled comparison.
Violence Vict 3(1):31–48, 1988.
21. Millon T: Millon Clinical Multiaxial Inventory Manual. Minneapolis, Interpretive Scoring
Systems, 1983.
22. Saunders D: Are there different types of men who batter? An empirical study with possible
implication for treatment. Presented at the Third National Family Violence Research Confer-
ence, Durham, NH, 1987.
23. Murphy C, Meyer S, O’Leary D: Family of origin violence and millon clinical multiaxial
inventory. II. Psychopathology among partner assaultive men. Violence Vict 8(2):165–176,
1993.
24. Wesner D, Patel C, Allen J: A study of explosive rage in male spouses counseled in an
Appalachian mental health clinic. J Counsel Dev, 70:235–241, 1991.
25. Bergman B, Brismar B: Assailants and victims: A comparative study of male wife-beaters and
battered males. J Addict Dis 12(4):1–10, 1993.
26. Dutton D, Hart S: Risk markers for family violence in a federally incarcerated population. Int J
Law Psychiatry 15:101–112, 1992.
27. Bersani C, Chen H, Pendelton B, Denton R: Personality traits of convicted male batterers.
J Fam Violence 7(2):123–134, 1992.
28. Pistole M, Tarrant N: Attachment style and aggression in male batterers. Fam Ther 20(3):165–
173, 1993.
29, Davidovich J: Men who abuse their spouse: Social and psychological supports. J Offender
Counsel Serv Rehab 15(1):27–44, 1990.
30. Hofeller KH: Battered Women, Shattered Lives. CA, R & E Research Associates, 1983.
31. MacEwen K, Barling J: Multiple stressors, violence in the family of origin and marital aggres-
sion: A longitudinal investigation. J Fam Violence 3(1):73–87, 1988.
32. Guerney B, Waldo M, Firestone L: Wife-battering: A theoretical construct and case report.
Am J Fam Ther 15(1):34–43, 1987.
33. Dollard J, Miller NE: Personality and Psychotherapy. New York, McGraw-Hill, 1950.
34. Miller B: The interrelationships between alcohol and drugs and family violence, in De LaRosa M,
Lambert Y, Gropper B, (eds): Drugs and Violence: Causes, Correlates, and Consequences. NIDA
Research Monograph 103. Rockville, MD, US Department of Health and Human Services,
National Institute on Drug Abuse, 1990, pp 177–207.
35. Gorney B: Domestic violence and chemical dependency: Dual problems, dual interventions.
J Psychoact Drugs 21(2):229–238, 1989.
36. Geffner R, Rosenbaum A: Characteristics and treatment of batterers. Behav Sci Law 8:131–140,
1990.
37. Cohen S: Aggression: The role of drugs, in The Substance Abuse Problems, vol 2. New York,
Haworth Press, 1985, pp 358-363.
38. Miller N, Millman R, Keskinen S: Outcome at six and twelve months post inpatient treatment
for cocaine and alcohol dependence. Adv Alcohol Subst Abuse 9(314):101–120, 1990.
39. Cloninger D: Neurogenic adaptive mechanisms in alcoholism. Science 236:410–416, 1987.
40. Van Hasselt VB, Morrison RL, Bellock AS: Alcohol use in wife abusers and their spouses.
Addict Behav 10:127–135, 1985.
41. Schukit J, Russell J: An evaluation of primary alcoholics with histories of violence. J Clin
Psychiatry 14:3–6, 1984.
42. Taylor S, Leonard K: Alcohol and human physical aggression, in Green RG, Donnerstein EI
(eds): Aggression: Theoretical and Empirical Reviews. New York, Academic Press, 1983, pp 77–
111.
43. Leonard K, Blane E: Alcohol and marital aggression in a national sample of young men.
J Interpers Violence 7(1):19–30, 1992.
44. Pihl R, Zacchia C: Alcohol and aggression: A test of the affect–arousal hypothesis. Aggressive
Behav 12:367–375, 1986.
45. Cherek D, Steinberg J, Manno B: Effects of alcohol on human aggressive behavior. J Stud
Alcohol 46:321–328, 1985.
46. Leonard K: Alcohol consumption and escalatory aggression in dyads. J Stud Alcohol 45:75–80,
1984.
47. Zeichner R, Phil R: Effects of alcohol and behavior contingencies on human aggression.
48. Voss H, Hepburn J: Patterns in criminal homicide in Chicago. J Crim Law Criminol Pol Sci
59:499–508, 1969.
49. Wolfgang M, Strohm R: The relationship between alcohol and criminal homicide. Q J Stud
Alcohol 17:411–425, 1956.
50. Leonard K, Bromet E, Parkinson D, et al: Patterns of alcohol use and physically aggressive
behavior in men. J Stud Alcohol 46:279–282, 1985.
Alcoholism: Social Aspects of Alcoholism, vol 4. New York, Plenum Press, 1976, pp 351–444.
52. Fagan J, Stewart D, Hansen K: Violent men or violent husbands: Background factors and
situational correlates, in Finkelhor D, Gelles RJ, Hotaling GT, Straus MA (eds): The Dark Side
of Families. Beverly Hills, CA, Sage, 1983, pp 49–67.
53. Roy M: Four thousand partners in violence: A trend analysis, in The Abusive Partner: An
Analysis of Domestic Battering. New York, Van Nostrand-Reinhold, 1982, pp 17–35.
54. Labell L: Wife abuse: A sociological study of battered women and their mates. Victimology
4:258–267, 1979.
55. Eberle P: Alcohol abusers and non-abusers: A discriminant analysis of differences between
two subgroups of batterers. J Health Soc Behav 23:260–271, 1982.
56. Byles J: Violence, alcohol problems and other problems in disintegrating families, J Stud
Alcohol 39:551–553, 1978.
57. Cleek M, Pearson T: Perceived causes of divorce: An analysis of interrelationships. J Marriage
Fam 47:179–183, 1985.
58. Sterling RC, Weinstein SP, Gottheil E, Shannon D: Psychiatric symptomatology of crack
cocaine abusers. J Nerv Ment Dis 182:564–569, 1995, in press.
59. Honer W, Gewirtz G, Turey M: Psychosis and violence in cocaine smokers. Lancet 451, 1987.
60. Sanchez J, Johnson B: Women and the drugs-crime connection: Crime rates among drug
abusing women at Rikers Island. J Psychoact Drugs 19:2, 1987.
61. Grinspoon L, Bakalar J: Drug dependence: Non-narcotic agents, in Kaplan HI, Sadock BS
(eds): Comprehensive Textbook of Psychiatry/IV. Baltimore, Williams & Williams, 1985, pp 1003–
1015.
62. Cohen S: The cocaine problems. Drug Abuse Alcohol Newslett 12(10): 1983.
63. Allen L: PCP: A schizophrenomimetic. US Pharm 60–66, 1980.
64. Roberts A: Substance abuse among men who batter their mates: The dangerous mix. J Subst
Abuse Treat 5:83–87, 1988.
65. Robins LN, Helzer JE, Croughan J, Ratcliff KS: National Institute of Mental Health: Diagnos-
tic Interview Schedule. Arch Gen Psychiatry 38:381–389, 1989.
66. Straus M: Measuring intrafamily conflicts and violence: The Conflicts Tactics (CT) Scale.
J Marriage Fam 41:75–88, 1979.
67. Lee V, Gottheil E, Weinstein SP: Personality characteristics of cocaine users who batter.
Paper presented at the American Society of Addictive Medicine, Chicago, 1995.
68. Edleson J, Syers M: The effects of group treatment for men who batter: An 18-month follow-
up study. Res Soc Work Pract 1(3):227–243, 1991.
69. Palmer S, Brown R, Barrera M: Group treatment program for abusive husbands. Am J Ortho-
psychiatry 62(2):276–283, 1992.
70. Shepard M: Predicting batterer recidivism five years after community intervention. J Fam
Violence 7(3):167–178, 1992.
71. DeMaris A, Jackson J: Batterer’s report of recidivism after counseling. Soc Casework 68:458–
465, 1987.
72. Grusznski RJ, Carrillo, TP: Who completes batterer’s treatment groups? J Fam Violence
3(2):141–150, 1988.
73. O’Leary KD: Assessment and treatment of partner abuse. Clin Res Dig Suppl Bull 12. Ameri-
can Psychological Association, 1995.
74. Geffner R: Guidelines for using mediation with abusive couples. Psychother Individ Pract,
1992.
75. Wallace B, Nosko A: Working with shame in the group treatment of male batterers. Int J
Group Psychother 43(1):45–61, 1993.
15
Alcohol, Drugs, and Violence in

Children’s Lives
Brenda A. Miller, Eugene Maguin,
and William R. Downs
Abstract. This chapter reviews the current state of knowledge concerning the interrelationship
between the cycle of alcohol and other drugs (AOD) use and the cycle of violence. This issue is
framed in terms of two questions. The first is the extent to which AOD use by the perpetrator is
related to the perpetration of violence toward children, defined here as including both physical
and sexual abuse. The second question is whether the experience of abuse during childhood is
related to the subsequent development of the abuse of alcohol and other drugs. The review
indicates that parental AOD abuse is related to physical and sexual abuse. However, because
most perpetrators are not parents, the relationship is not yet clear. The data do support the link
between experiencing childhood violence and the development of later AOD abuse. Theoretical
explanations for each link are reviewed and mediating variables are identified. The review con-
cludes with a presentation of methodological issues and the directions for future research.
1. Introduction
This chapter addresses the current state of knowledge about the interrelation-
ship between the cycle of alcohol and other drugs (AOD) and the cycle of
violence1 by focusing on two questions. First, to what extent is AOD use* by
* While much of the work covered addresses these relationships for alcohol use/abuse, there is
some evidence that these relationships exist with other types of drugs, particularly illegal or
nonmedical use of drugs. Some cautions are needed in examining the relationships for drugs
other than alcohol. First, the psychopharmacological effects of other substances vary widely.
Brenda A. Miller and Eugene Maguin • Research Institute on Addictions, Buffalo, New York
14203. William R. Downs • Center for the Study of Adolescence, University of Northern
Iowa, Cedar Falls, Iowa 50614.
Recent Developments in Alcoholism, Volume13: Alcoholism and Violence, edited by Marc Galanter.
357
the perpetrator related to the perpetration of physical or sexual abuse against

children?+ Second, is the experience of physical or sexual abuse during child-
hood related to the subsequent development of AOD use/abuse? However,
this chapter does not address the important question of whether AOD use by
children is related to the perpetration of violence toward other children or
adults (see Miczek et al.2 review for a current review of this topic).
Studies pertaining to each question are covered first and are then fol-
lowed by possible theoretical explanations for the links between these phe-
nomena. The chapter next presents a discussion of the methodological con-
siderations for researchers interested in pursuing these lines of inquiry and
concludes with our recommendations for future research directions.
2. Perpetrator’s Substance Use/Abuse

and Physical and Sexual Abuse of Children
2.1. Parental Substance Problems and Perpetration of Child Physical Abuse

Early reviews concluded that, at most, only a modest association existed
between parental alcohol problems and the perpetration of child physical
abuse by the parents.3,4 More recent reviews have concluded that either no
relationship exists,5 or that the association is limited to certain (unidentified)
subgroups of those with alcohol problems.6 However, all of these reviews
have also acknowledged that the research testing these relationships had
serious methodological shortcomings, including a lack of standardized instru-
ments for the two main concepts of child abuse and parental alcohol prob-
lems, and an overrepresentation of minority and poor subgroups in the selec-
tion of samples.3 Another concern was the inclusion of samples based on
identified cases of child maltreatment, which limited the findings to children
officially investigated. Thus, self-report cases that were never investigated
were not represented in early studies. Where clinical samples of alcoholics
were used without adequate comparison groups, it was difficult to separate
the effects of treatment seeking from the effects of parental alcohol problems.
More recent studies that have used stronger methodological approaches
support the association between parental AOD problems and experiences of
physical abuse by the child.7,8 Smyth et al.8 compared 102 mothers with and
Different substances may play different roles. Second, the illegal nature of other drugs compli-
cates the examination of the roles; relationships may exist more as a result of the illegal nature of
the substance rather than the substance itself.
+ Describing adolescents and teenagers as children in this chapter is meant to simplify the presen-
tation but not meant as an oversight of the different developmental issues for the different age
ranges. For an excellent review of the developmental issues as they relate to childhood victimiz-
ation, see Finkelhor.118
15 • Alcohol–Violence in Children's Lives 359
66 mothers without lifetime histories of alcohol problems on their reported

levels of harsh punishment as measured by the Conflict Tactics Scale (CTS)9,10
and the Child Abuse Potential Scale.11 Smyth et al.8 found that mothers with
histories of alcohol problems were more likely to use harsh punishments on
their children. Furthermore, mothers with histories of alcohol problems were
more likely to report that their partner was verbally aggressive toward their
children, suggesting that mothers with alcohol problems may be less capable
of shielding their children from victimization from others. In the Reider et al.7
study, mothers and fathers in alcoholic families and nonalcoholic families
completed the CTS and measures of lifetime alcohol problems. Reider et al.7
found the same relationship for both parents: a higher level of lifetime alcohol
problems was associated with a higher level of parent-to-child violence for
that parent. Thus, there is evidence that the association holds in samples with
clinically significant levels of alcohol problems.
Other recent studies that have used samples drawn from the general
population have also supported a link between parental AOD problems and
perpetration of violence toward children. For instance, Holmes and Robins12
used the Ecological Catchment Area (ECA) study sample, which was de-
signed to determine the general population prevalence and incidence rates of
psychiatric disorders across five different sites in the nation. They found that
parents with diagnoses of alcohol abuse/dependence (primarily fathers) or
with diagnoses of depression (primarily mothers) reported higher levels of
harsh, unfair discipline and parent-to-child violence than parents with nei-
ther diagnoses. However, given that the levels of harsh, unfair discipline and
violence were similar for parents who had either an alcohol or a depression
diagnosis, it would appear that psychiatric diagnoses in parents may raise the
level of risk for violence toward children, rather than specifically alcohol
problems. Thus, the study provides support for a link between alcohol prob-
lems and child maltreatment in the general population. However, given
the relationship between parental depression and child physical abuse
and the high concordance between depression and alcoholism,13 the link
between parental alcoholism and child physical abuse may be due to the
common association with depression specifically or parent psychiatric disor-
ders generally.
Because parents may be biased against reporting their own use of harsh
punishment or severe violence toward their children, some researchers have
asked adults to retrospectively report on use of physical violence and their
parents' AOD use. In one such study, Radomsky14 separated 120 consecutive
female patients from a general medical practice into one of three groups
according to whether they described their parents as (1) alcohol dependent;
(2) harsh, rigid, or difficult but not alcohol dependent; or (3) neither alcohol
dependent nor harsh, rigid, or difficult. Physical or sexual abuse levels were
similar for the women who described their parents as either alcohol depen-
dent or harsh, rigid, or difficult, and these levels were significantly higher
than for women whose parents had neither of these problems. In another
study,15 women’s reports of maternal and paternal alcohol problems (as-

sessed by questions from the Research Diagnostic Criteria16) and both mater-
nal and paternal child violence (assessed by the CTS) were collected from 472
women.* After controlling for race, number of changes in childhood family
structure, and childhood socioeconomic status (SES) each parents’ alcohol
problems was found to be related to that parents’ violence toward the respon-
dent. Thus, there is evidence from several recent studies of a relationship
between parental alcohol use and child physical abuse.
Most, if not all previous studies, have assumed that both parents in a
family with an alcoholic or drug-abusing parent would report elevated levels
of child physical abuse. To our knowledge only two studies have examined
this question. Two different studies, one for male parolees17 and one for
women in treatment settings and the community,15 found that parent-to-
child violence was related to the parent with alcohol problems but not to the
parent without the alcohol problems. Furthermore, Downs and Miller15
found that these relationships remained after controlling for race, SES, and
number of changes in the family structure. While these studies need to be
replicated, the methodological precision of assessing both alcohol depen-
dence and perpetration of violence for the same parent to correctly estimate
the association between parent AOD and perpetration of violence toward
children should not be overlooked.
In summary, the methodological shortcomings of the early literature may
have led researchers to conclude that no relationship between parental AOD
use and child physical abuse existed. Although more recent studies, using
more representative samples, as well as a variety of types of samples and
comparison groups and better assessments of the constructs, have found
parental AOD use to be related to child physical abuse, further studies are
needed to build confidence in this tentative conclusion.
2.2. Perpetrator AOD Problems and the Sexual Abuse of Children

Unlike child physical abuse, which implies a custodial relationship be-
tween the perpetrator and the child, child sexual abuse (CSA) does not neces-
sarily imply such a relationship with the perpetrator. This definitional differ-
ence has a significant consequence for the types of questions that have been
addressed. The question of the relationship between AOD problems and CSA
requires that CSA perpetrators be studied or that CSA victims be able to
accurately describe their assailant’s level of substance use problems. If, in-
stead, victims describe their parents’ levels of AOD problems, the resulting
association is a combination of the association between perpetrator AOD
* This study included samples of women in treatment for alcoholism (n = 98), mandatory educa-
tion programs for driving while intoxicated (n = 100), in a shelter for battered women or in
groups for battered women from an agency affiliated with the shelter (n = 97), in outpatient
treatment for mental health issues (n = 77), and selected from the community using random
digit dialing (n = 82).82
15 • Alcohol–Violence in Children’s Lives 361
problems and CSA, given that their parent was the perpetrator, and the
association between parent AOD problems and being assaulted by a non-
parental adult, given that their parent was not the perpetrator. Thus, two
distinct questions are blurred together.
Early studies of the relationship between substance use problems and
CSA were primarily single-sample designs with subjects drawn from clinical
or correctional populations of males convicted of sexual offenses. An early
review of this literature concluded that alcohol use was involved in a substan-
tial minority (approximately 40%) of sexual abuse incidents and that approx-
imately half of all offenders had histories of alcohol problems.18 However,
because these early studies included no control or comparison group, it was
difficult to definitively conclude that alcohol problems are related to CSA.
More recent studies have incorporated a number of methodological im-
provements to remedy the weaknesses of previous studies. However, these
more recent studies have also approached this question from the perspective
of the victims of CSA rather than that of the perpetrator. For example, Yama
et al.19 studied a convenience sample of 364 female college students who were
asked to retrospectively describe their histories of CSA according to Fin-
kelhor’s20 protocol and their parents’ alcohol problems. The results showed a
significant relationship between parental alcoholism and CSA. In a second
study, Windle et al.21 assessed the presence of a family history of alcoholism
and the presence of physical and sexual abuse in a group of male and female
alcoholic inpatients. Exactly 72% of men who had experienced both childhood
physical and sexual abuse had a positive family history of alcoholism com-
pared to about 39% of men who had experienced sexual abuse only and about
32% of the men who had neither physical nor sexual abuse during their
childhoods. For women in the study, about 40% of women who had, experi-
enced neither form of abuse had a positive family history of alcoholism com-
pared to 53% who had experienced sexual abuse and about 60% who had
experienced both forms of abuse. Thus, experiences of CSA were associated
with a family history of alcoholism for both males and females. Finally,
Downs and Miller15 examined the association between CSA and parental
alcoholism in a sample of 472 women.* A significant bivariate relationship
was found and the relationship remained after controlling for race of respon-
dent, number of changes in childhood family, and childhood SES. In conclu-
sion, more recent studies incorporating methodologically improved assess-
ments of CSA and alcohol problems have consistently found a relationship
between parental alcoholism and CSA.
Because more recent studies have used standardized assessments of
CSA, some data, although sparse, are available on the relationship of the
victim and the perpetrator. Yama et al.19 found that 18% of their subjects had
been sexually abused by a father or stepfather, 33% by another relative, and
the remainder by strangers. Downs and Miller15 reported that only 4% of
women had been abused by the biological father or adoptive father. Although
neither of these studies provide data on the magnitude of the association
between perpetrator AOD problems and CSA (since victims could not report
on the AOD problems of perpetrators who were not their parent), Downs and
Miller15 reported that a significantly higher percentage of women in the alco-
holism treatment sample (37%) had been assaulted by a male family friend
than in either the driving while intoxicated (DWI) sample (12%) or the house-
hold sample (16%).
Although many of the early studies on the parental AOD problems and
CSA relationship suffer from a number of methodological weaknesses, cur-
rent studies have used much improved and standardized methods of assess-
ing CSA and parental alcoholism and substance use. While two recent studies
have reported a significant association between parental AOD use and CSA,
the fact that these studies also indicate that the majority of CSA is perpetrated
by persons who are not the victim’s parents means that very little is known
about the magnitude of the association between CSA and the perpetrator’s
substance use. As a result, it seems advisable to study the CSA–parental
substance use relationship in the context of the type of perpetrator (e.g.,
parent versus nonparent). These studies also highlight the importance of
identifying how nonparental perpetrators gain access to their victims.
2.3. Explanations for the Link between Perpetrators’ Substance Problems

and Violence toward Children
Various hypotheses have been proposed to account for the link between
perpetrator’s alcohol use and his or her violence. One of the more promising
is the cognitive disorganization hypothesis, which proposes that alcohol use
results in the narrowing of attention to some social cues but not others.22
Intoxication results in a “myopic” view of the world and only the most salient
cues are acknowledged.23 This narrowing of focus limits the abilities of the
individual to process information to appropriately avoid behavioral conse-
quences.24 As applied to family violence, alcohol use results in miscom-
munication between family members, a focus on some cues but not others,
overestimation of present threat, underestimation of consequences for ag-
gression, and ultimately an increase in the likelihood of violence.
Collins25 has pointed out the need to distinguish between acute and
chronic effects of alcohol. For example, over a period of many years, chronic
heavy alcohol use results in a number of cognitive impairments that are
evident whether or not the person is currently drinking. Thus, alcohol use
can have both acute effects on cognition (i.e., while drinking) and chronic
cumulative effects on cognition. These possibilities help explain certain data
regarding patterns of drinking and alcohol use. For example, Kantor and
* This study included samples of women in treatment for alcoholism (n = 98), mandatory educa-
tion programs for driving while intoxicated (n = 100), in a shelter for battered women or in
groups for battered women from an agency affiliated with the shelter (n = 97), in outpatient
treatment for mental health issues (n = 77), and selected from the community using random
digit dialing (n = 82).82
Straus26 found parents with the highest levels of abuse to be mothers with
binge-drinking patterns (reflecting perhaps acute effects of alcohol on cogni-
tion) and fathers with high daily alcohol consumption patterns (perhaps re-
flecting chronic alcohol effects).
A second promising explanation is the deviance disavowal hypothesis.
According to this hypothesis, alcohol use allows the perpetrator to attribute
the violence to the alcohol use, thereby avoiding or minimizing personal
responsibility for the violence.27–30 Thus, rather than creating a pharmaco-
logical effect, alcohol use is connected to cultural expectancies, and it is these
cultural expectancies that create a greater likelihood of deviant, aggressive, or
violent behavior. This hypothesis explains why alcohol use might be related
to aggression in some settings but not others and why the cultural expecta-
tions about alcohol use may be important to explaining the behaviors exhib-
ited when drinking.
A third hypothesis that has been advanced is the disinhibition hypothe-
sis, which posits a direct pharmacological link between alcohol use and ag-
gression such that alcohol anesthetizes brain centers that control inhibi-
tions.22,31–33 However, there has been little direct empirical support for the
hypothesis.34–37 In particular, the disinhibition hypothesis cannot explain
why alcohol increases aggression in some experimental paradigms but not
others.38 Finally, cultural norms approve the use of corporal punishment
(i.e., violence) against children,39 and thus it might be argued that there are
fewer inhibitions against harsh or punitive punishments of children.
The finding that an apparent majority of the perpetrators of child sexual
abuse are not the child’s parents or surrogates requires that the mechanism by
which adults outside the nuclear family gain access to the child be identified.
Parental unavailability is one construct that has been found to be associated
with childhood sexual abuse.40 Substance-abusing parents would be pre-
sumed to be less available to protect children from extrafamilial sexual abuse
than nonsubstance-abusing parents. For example, Downs and Miller15 found
that the lack of paternal protection due to father’s alcohol problems may have
contributed to other males known to the family perpetrating sexual abuse
against daughters, and formulated two hypotheses to account for these find-
ings. First, daughters growing up in homes with an alcoholic father may not
receive the typical emotional support, nurturance, and sustenance from their
fathers. As a result, these girls may be more vulnerable to the manipulations
of adult males outside the nuclear family who provide that support and
nurturance, but at the cost of sexually abusing the girl. Second, daughters in
homes with an alcoholic father may be more likely to be placed temporarily
with relatives as the parents attempt to cope with the problems of alcohol
dependence. These girls may then be more likely to be abused at the homes of
these relatives. However, no studies of either theorized linkage have been
conducted.
A second hypothesized explanatory construct is parental protectiveness.
Most of our understanding about parental protectiveness comes from studies
that have examined what mothers do in response to their children’s victimiza-

tion. Faller41 reports a range of maternal responses toward children vic-
timized by violence. Protective responses for children victimized by CSA
include calling the police or protective services, leaving the house with the
children, making the perpetrator leave, and/or initiating divorce proceedings
when the perpetrator is a partner/spouse, placing the child where the alleged
perpetrator cannot have access to the child, and insisting the perpetrator get
treatment. Nonprotective responses include blaming or disbelieving the
child, or allowing the child to remain in the risky environment.
Mothers may have difficulty making protective responses because of
their substance abuse, as well as their own history of physical violence.42,43 In
addition, women with a history of CSA may be unable to protect their chil-
dren from dangerous persons in the child’s environment.44–46 In particular,
women who have been sexually victimized as children are frequently revic-
timized as adults,47,48 an experience that may make it more difficult for these
women to protect their children.49 For instance, if mothers develop relation-
ships with abusive partners, not only are they in danger, but their children
may be in danger also. In their review of long-term effects of child sexual
abuse, Beitchman and colleagues47 suggest several explanations for this pat-
tern, including: victimized children are forced out of the family and into high-
risk situations; the impact on self-esteem may make mothers conspicuous
targets for sexually exploitive men who may also abuse their children; women
may idealize abusive men, seeking to reestablish the special relationship with
their father (when father was the source of abuse); and impaired ability to
identify correctly untrustworthy persons who may abuse their children.
Mothers’ ability to maintain a quality parent–child relationship, in an
environment conducive to skilled parenting, is limited by coexisting comor-
bidity problems.46 A review by Finkelhor and Browne50 report that sexually
abused women in the general population displayed more mental health prob-
lems compared to nonvictims. Furthermore, women who abuse alcohol and
other drugs show a similar high comorbidity of mental health problems when
compared to the general population.51 Jumper’s52 meta-analysis confirms the
relationship between childhood sexual abuse and impaired adult psychologi-
cal adjustment, typically resulting in depression, low self-esteem, childhood
psychopathology, and adult emotional and behavioral problems. Together,
these factors place women at greater risk to abuse their children.53,54
Although these explanations may prove to account for the direct links be-
tween AOD use and the perpetration of physical or sexual abuse against
children, it may also be that the bivariate relationship is completely accounted
for by variables that are strongly correlated with both substance use and
physical or sexual abuse (i.e., third variables). In their review, West and
Prinz55 suggested a number of possible third variables including SES, mater-
nal psychiatric illness, parental discord, divorce, financial problems, parental
imprisonment, and reporting bias. In addition to these variables, it is critical
that comorbid antisocial personality disorder (ASP) be tested as well, since it
is well known that ASP is related to the early onset of alcohol problems,
referred to as primary ASP and secondary alcoholism,56–58 type II alcohol-
ism,59 or as type B alcoholism.60 Also, DiLalla and Gottesman61 have sug-
gested that children of ASP-positive parents are more likely to be physically
abused than children of ASP-negative parents. Thus, the association between
parental alcohol problems and physical child abuse may be due to their com-
mon link with ASP or with other variables such as those suggested by West
and Prinz.55 Furthermore, these same relationships may also be true for CSA.
A few studies have controlled for third variables including a variable
conceptually related to ASP. Among the variables included as possible third
variables include antisocial behavior, depression, and marital aggression7;
social support, stress, and SES62; and race, number of changes in childhood
family structure, and childhood SES.15 In each study cited, the associa-
tion between parent alcohol problems and child physical abuse remained
significant,
Downs et al.63 have argued that there are serious empirical and defini-
tional problems with the ASP construct such as low interrater reliability, low
diagnostic stability, and conceptual ambiguity, which limits its usefulness as
an explanatory construct. Although more work is needed, the results to date
suggest the relationship is not due to a third variable. More research is needed
to unravel additional factors that may be related to both parental alcohol
problems and childhood maltreatment.
The existing body of literature pertaining to perpetrator AOD use and the
perpetration of abusive behavior toward children needs to be enlarged. How-
ever, more than simple bivariate studies are needed. Future studies need to
measure and control for possible third variables as well as include measures
of theoretical constructs suggested by the several theories proposed as expla-
nations of the substance use and violence toward children relationship.
3. Childhood Victimization as Antecedent to Later Substance Use

3.1 Childhood Physical Abuse Prior to Substance Use
Several studies using retrospective reports for adults suggest that indi-
viduals with AOD problems are more likely to report histories of physical
abuse. Straus and Kantor64 found that retrospectively measured severe par-
ent-to-child violence during adolescence was related to increased drinking as
adults in a national probability sample of adults. Holmes and Robins65 found
that alcoholic subjects (males and females) were significantly more likely to
report that their parents had used severe forms of violence than controls.
Brown and Anderson66 found that among a sample of males and females
receiving care at a military psychiatric hospital, physically abused patients
were more likely to receive a alcohol disorder diagnosis than control patients
who reported neither physical nor sexual abuse. In another study, Blane et
al.67 found that a greater exposure to childhood victimization led to an in-
creased likelihood of an alcohol-dependence diagnosis in adulthood for a

sample of male offenders on parole. Finally, Windle et al.21 found that the
prevalence of child physical abuse histories for both male and female inpa-
tients in alcohol treatment facilities were considerably higher than rates for
community-based studies, although the rates did not differ significantly for
males and females. Thus, in samples of adults using retrospective recall, the
evidence indicates a relationship between experiences of childhood physical
abuse and the subsequent development of adult substance use problems.
While this evidence is compelling, neither of the two longitudinal studies
that examined this question were able to find a link between experiences of
childhood physical abuse and the development of later alcohol problems.
Widom et al.68 matched adults who had an official record of physical or sexual
abuse or neglect from juvenile or adult court records as children to adults
without such records as children on age, sex, and race, residence, and family
economic status. Because Widom et al.68 used only official records, it is likely
that the comparison group contained an unknown percentage of cases with a
history of unreported abuse or neglect. As adults, subjects completed an
extensive assessment including diagnostic interview protocols. They found
no relationships between physical abuse and measures of current or lifetime
alcohol symptoms or diagnoses for either males or females. McCord69 fol-
lowed a group of 232 males originally assigned to the control group of the
Cambridge–Somerville Study and interviewed them as adults. In 1957, the
males were divided into four groups (abused, neglected, rejected, or loved)
based on nearly 6 years of case records of interviews with the boys and their
parents conducted by project social workers when the boys were from 5 to 9
years of age. McCord69 found that the proportion of males who later became
alcoholics in adulthood was not related to whether or not they had been
abused, neglected, rejected, or “loved” as children. Further analyses of longi-
tudinal datasets are needed to provide additional clarity on this question.
The few studies of adolescents using retrospective recall of physical
abuse have yielded inconsistent results. A series of studies by Dembo and
associates70–72 of several different samples of adolescents in detention facili-
ties found significant bivariate relationships for males and females separately
and in combined samples between physical abuse and lifetime drug use.
However, physical abuse failed to predict either adverse alcohol or marijuana
consequences after controlling for background characteristics, self-reported
delinquency, referral history, and psychological functioning in either cross-
sectional analyses or in longitudinal analyses.73 Although Dembo and associ-
ates used several different samples, other studies are needed to address this
question with different types of samples. Even so, it may be that effect of
physical abuse on the development of AOD problems becomes distinguish-
able only in adulthood.
Most studies have simply examined the effect of parental violence on
children. The question these studies cannot address is whether father-to-child
violence, for example, has different effects than mother-to-child violence for
female children. In two separate samples of women, Downs et al.74 and Miller
et al.43 found that alcoholic women in treatment reported significantly higher
levels of verbal aggression, moderate violence, and severe violence from their
father but not their mother than did nonalcoholic women. Although the
Miller et al.43 study found that alcoholic women also reported higher levels of
verbal aggression, moderate violence, and severe violence from their mother,
these differences disappeared once background characteristics including pa-
rental alcoholism were controlled. Furthermore, these relationships also held
when involvement in treatment was controlled.43
Overall, the data indicate that the relationship between child physical
abuse and later AOD problems is stronger for adult samples than for adoles-
cent samples. However, further studies using adolescent samples are needed,
since the current database consists of only two samples from a single source.
Furthermore, there is some evidence that the effects of severe violence in the
family depend, at least for female children, on whether paternal or maternal
violence is examined.
3.2. Child Sexual Abuse Prior to Substance Use

Although early reviews of the consequences of CSA47,75,76 did not consid-
er substance abuse, a more recent review did do so. Polusny and Follette77
identified a number of studies, using both community and clinical samples,
that reported significantly higher rates of AOD diagnoses (either alcohol or
drugs) for women who had been CSA victims. Their summary shows that the
prevalence of lifetime AOD diagnoses in community samples ranged from
13.7 to 31.0% among female CSA victims compared to 3.1 to 12.0% among
non-CSA females. In clinical samples the prevalence ranged from 20.9 to
57.0% among female CSA victims compared to 2.3 to 27.0% among non-CSA
females. Polusny and Follette77 included studies that defined CSA as any
forced or coerced sexual behavior imposed on a child and sexual activity
between a child and a much older person, whether or not coercion was
involved. However, because many studies relied on single-item assessments
or other nonstandardized assessments, there was likely some degree of vari-
ability in the prevalence data attributable to methods.
More recent studies also generally support the relationship between CSA
and AOD problems for adult women; however, much less is known about
adolescent women. For example, Wilsnack et al.78 found that CSA, as mea-
sured by the Russell79 and the Wyatt80 methodology, was related to an in-
creased prevalence of alcohol problems over the past 12 months, including
dependence symptoms, problem drinking, and heavy drinking, as well as
lifetime substance use in a large national probability sample of US women.
Also, Mullen et al.81 found that CSA involving intercourse was related to both
heavy drinking and drug dependence in an enriched probability sample of
Dunedin, New Zealand adult women. In this study, sexual abuse was mea-
sured by self-report followed by an interview to confirm the self-reported
experiences. Only one study, that of Dembo et al.,70 was found that studied
adolescent females. This study, which measured CSA by the Finkelhor20
criteria, found that CSA was related to lifetime drug use for a sample of
adolescent females in detention facilities. Thus, these three studies support
the CSA–substance abuse relationship for both adult and adolescent women.
Widom et al.68 also studied this question with their sample of adults with
officially recorded sexual abuse and matched adults. They found that offi-
cially recorded CSA was associated with a higher level of current alcohol
symptoms for females. However, no difference was noted for the lifetime
alcohol symptoms, current alcoholism diagnosis, and lifetime alcoholism di-
agnosis.
Since the mid-1980s, Miller and associates have studied the links between
early childhood victimization and the later development of alcohol and other
drug problems in women (see Miller and Downs82 for a review). In an early
study,42 which compared 45 women who either had previously received
treatment or were currently in treatment to 40 women without alcohol prob-
lems recruited from the community, women with alcohol problems were
found to be 2.5 times more likely to have been CSA victims. In a later study,
Miller et al.43 found that alcoholic women were significantly more likely to
have experienced all forms of sexual abuse (exposure, touching, and penetra-
tion) than were women in either the DWI or community samples. In a second
analysis, women in treatment settings with and without alcohol problems
and women in the community were compared to assess the effects of treat-
ment involvement. Women in treatment with alcohol problems were signifi-
cantly more likely to have CSA histories compared to women in treatment
without alcohol problems after controlling for parental alcohol problems.
Taken together, the evidence from a number of retrospective studies of
adults, a study of adolescents, and a prospective cohort study indicates that
females who have been sexually abused as children are more likely to have
AOD problems later in life. The evidence is clearer for adult women than for
adolescent women by virtue of the fact that more studies have been con-
ducted for adults. Thus, further work is needed with adolescent samples.
By comparison, few studies have included samples of males, regardless
of the source.21 In fact, Polusny and Follette’s77 review included only one
study of males83 and only a few others were located in a literature search.
Stein et al.,83 who used the Los Angeles ECA sample, found that significantly
more males who had been sexually assaulted as children (44.9%) had received
an abuse or dependence diagnosis for drugs as adults than did males who
had not been assaulted (7.8%). However, no difference in the rates of an alco-
holism abuse or dependence diagnosis was found between the two groups. In
a prospective cohort study of adults, Widom et al.68 did not find any associa-
tion between CSA and either current or lifetime alcohol symptoms, or be-
tween CSA and either lifetime or current diagnosis for males. In a third study,
Harrison et al.84 found that sexually abused male adolescent patients in a
substance abuse treatment program were more likely to report daily drinking
than nonabused male patients. Finally, Dembo et al.70 found that CSA was
related to lifetime drug use for male adolescents in a detention facility. Thus,
the data for males is mixed as some studies have found significant results,
while others using different designs have not.
Overall, the data indicate that CSA is related to later AOD problems in
adulthood for women, irrespective of whether the women are drawn from
clinical samples or community samples when retrospective measures of CSA
are used. Although there have been considerably fewer studies of males, CSA
may also be related to later AOD problems. Furthermore, the relationship
between CSA and alcohol problems, at least for women, has remained when
personal and family background characteristics, including the presence of
parental alcoholism, were controlled. Similar analyses are needed for males.
The literature suggests that one of the consequences for children who are
either physically or sexually abused is a greater likelihood of AOD problems
in adulthood and possibly in adolescence. At present, the data are not suffi-
cient to determine whether one type of abuse is associated with a greater
likelihood of AOD problems or whether males or females, given the experi-
ence of abuse, are more prone later AOD problems. Further work is needed
on these points.
3.3. Explanations of the Linkage from CSA and Parent-to-Child Violence

to AOD Problems
Although the explanations of the linkage from CSA and parent-to-child
violence to AOD problems are not well understood, Miller and Downs82
proposed three possibilities for women. In this chapter, these explanations
will be expanded to include men (see also Downs and Miller,15 Downs et al.63).
3.3.1. Substance Use as a Coping Mechanism. Reviews of both CSA and

child physical abuse consequences have concluded that depression, suicidal
ideation, and low self-esteem are among the sequela of CSA and child physi-
cal abuse in adult samples.77,85 And, subsequent studies reaffirm these con-
clusions.64,78 Although these data point to depression as a mediating variable
between victimization and subsequent substance use, the comorbidity data
for alcoholism and depression do not completely support this view. Results
from the ECA studies indicate that for males the observed comorbidity of
depression and alcoholism abuse or dependence is usually—in about 78% of
the cases—due to the secondary depression; that is, depression which fol-
lows the onset of alcoholism.13 However, Hesselbrock et al.57 found the con-
verse for females: that a diagnosis of major depression preceded a diagnosis
of alcohol dependence. Thus, the role of depression as a mediating variable
may depend on gender.
Another perspective is offered by studies of the effects of childhood
abuse for adolescent samples. Reviews of the effects of CSA and physical
abuse on children85,86 have also found elevated levels of depression, with-
drawal, and anxiety. More recently, Boney-McCoy and Finkelhor87 found
that 10- to 16-year-old males from a community sample who have experienced
CSA, but not similar females, reported increased sadness. They also noted no
relationship between parental violence and sadness for either males or fe-
males. However, McCloskey et al.88 found that, after controlling for all forms
of family violence, a canonical correlation analysis revealed associations be-
tween both father-to-child violence and severe father-to-child violence by
both maternal and child report and depression by maternal report in their
community sample of 6- to 12-year-old children. Furthermore, CSA was not
related to depression by either maternal or child report. Studies of clinical
samples of adolescents who had experienced either physical or sexual abuse
have also found increased depression and self-destructive behavior for both
males and females.84,89,90 Thus, it does appear that on the whole depression is
also a consequence of abuse for adolescents, just as it is for adults. However,
further studies are clearly needed to reinforce this conclusion and to examine
whether gender differences also exist for adolescents.
Kaplan’s91 theory of low self-esteem (i.e., self-derogation), although de-
veloped to explain adolescent drug abuse, provides a cogent theoretical link-
age between depression and low self-esteem reported by children and adoles-
cents who have experienced CSA and physical abuse and the development of
substance abuse problems. Briefly (see Miller and Downs82), CSA and physi-
cal abuse are posited to lead to self-devaluation and a loss of self-esteem when
the emotional impact of the abusive events overwhelm the coping abilities of
the child victim. Ultimately, depression and possible self-destructive behavior
may result. Substance use is viewed as a method for keeping both the short-
term and long-term emotional consequences at bay.
The empirical model implied by these data and Kaplan’s91 theory has not
been completely tested with either child or adolescent samples. However,
Dembo et al.70 was able to test a portion of the model in a cross-sectional
analysis of data from a sample of male and female juveniles in detention
facilities. They found that both CSA and child physical abuse predicted in-
creased self-derogation, and that self-derogation subsequently predicted life-
time drug use. However, their data also showed that self-derogation did not
fully mediate the bivariate relationship between either CSA or child physical
abuse and lifetime drug use. Thus, their results indicate that additional vari-
ables may be needed to account for the bivariate relationships.
3.3.2. Externalizing Behavior. Although reviews of studies using adult sam-

ples have not reported any evidence of a possible relationship between CSA
and externalizing behaviors (i.e., aggression, violence, hyperactivity, and de-
linquency) either during adolescence or adulthood, reviews focusing on child
and adolescent samples have done so. Kendall-Tackett et al.86 concluded that
increased aggression, delinquency, and cruel or antisocial behavior was a
consequence of CSA. Malinosky-Rummell and Hansen’s85 review of the con-
sequences of physical abuse also reached similar conclusions. In addition,
Malinosky-Rummell and Hansen85 concluded that physical abuse is associ-
ated with an increased frequency of noncompliance, nonaggressive conduct
disorders in children, and possibly nonviolent delinquency (i.e., property
crimes). Recent studies have also generally supported these conclusions.

Boney-McCoy and Finkelhor87 found that CSA was associated with increased
“teacher trouble” for both males and females. Contact CSA, a more severe
form of CSA, was also associated with teacher trouble for female adoles-
cents (the parallel analysis was not done for males due to insufficient sample
size). They also found that parental violence was associated with teacher
trouble for female adolescents but not for males. Although McCloskey et al.88
did not find any evidence that CSA was associated with increased opposition-
al behavior in their sample of preadolescent children, they did find that mod-
erate mother–child and father–child violence but not severe father–child vio-
lence was associated with oppositional behavior, after controlling for the
global level of family violence for their sample of preadolescent children.
Aggression and other forms of externalizing behavior, particularly when
it appears at a young age and persists, has been repeatedly found to be a
strong predictor of substance use problems (see Hawkins et al.92 for a review).
Involvement in deviant peer groups whose activities are organized around
delinquency or substance use has been posited by several investigators as the
key mediating variable between aggression and substance use.91,93 Another
hypothesis, advanced by Miller and Brown,94 proposes that antisocial or ag-
gressive behavior in childhood underlies delayed skills in self-regulation that
lead to the development of alcohol and drug problems by young adulthood.
In addition to the role of aggressive behavior, Kaplan91 also proposes that
children with low self-esteem and poor coping and social interaction skills
may lose the motivation to conform to more prosocial peer groups, and thus
join with peers involved in deviant peer groups. However, no studies were
located that tested models such as these.
3.3.3. Posttraumatic Stress Disorder Related Symptoms. Along with depres-

sion, posttraumatic stress disorder (PTSD)* is one of the more commonly
occurring consequences of CSA or parent-to-child violence.47,76,77,85,86 More
recently, Boney-McCoy and Finkelhor87 found a significant association be-
tween the presence of CSA and parent-to-child violence and increased PTSD
symptoms for both male and female samples.
Currently, only one study has examined the relationship between victim-
* The Diagnostic and Statistical Manual of Disorders (DSM), 4th ed, states that this disorder occurs
after exposure to extreme traumatic stressors such as: (1) personal experiences that threaten
death, serious injury, or personal integrity in some other manner, or (2) witnessing or learning
about another family member’s or close associate’s unexpected death, serious injury, or threat
to personal integrity. The individual’s response to the event must involve intense fear, helpless-
ness, or horror. Such experiences can trigger symptoms characteristic of PTSD: (1) a persistent
reexperiencing of the trauma (e.g., recurrent distressing dreams of the event); (2) avoidance of
the associated stimuli (e.g., efforts to avoid activities, places, or people that arouse recollections
of the trauma), as well as a numbing of the general responsiveness (e.g., feelings of detachment
or estrangement from others); and (3) persistent symptoms of increased arousal (e.g., hyper-
vigilance, startle response). These symptoms must be present for more than 1 month and create
significant distress or impairment in social, occupational, or other important areas of function-
ing for a diagnosis of PTSD.
ization, PTSD symptoms, and alcohol problems. The prevalence of two or

more alcohol problems among women participating in the National Women's
Study95 was 1.5% of those who had not been victimized, 3.9% among women
who had been victimized but experienced no PTSD symptoms, and 14.5%
among women who had been victimized and experienced PTSD symptoms.
Theoretical underpinnings of these data are not clear. However, it may be
that alcohol, and possibly drugs as well, may facilitate the disassociation
needed to gain relief from the memories of the experience. Support for this
view comes from women's own stated motives for drinking.96–98
Empirically, the association between CSA and parent-to-child violence
and AOD problems may be due to a variable that is associated with both
physical and sexual abuse and the later development of AOD problems.
Parental alcoholism is one variable that has been advanced for this role.
Parental alcoholism has been linked to the perpetration of physical abuse
and CSA. It has also been found to be consistently and strongly linked to the
development of alcoholism in the biological offspring.60,99,100 Although these
data suggest that parental alcoholism may account for the association between
child physical and sexual abuse and later alcoholism, few studies have exam-
ined this issue. However, in two separate studies, Miller et al.42,43 found that
while parental alcohol problems did predict later alcohol problems for women,
CSA also predicted later alcohol problems. In addition, Miller et al.43 also found
that father-to-child severe violence but not mother-to-child severe violence also
predicted later alcohol problems. Thus, there are some data to indicate that
parental alcoholism does not account for the association between abuse during
childhood and later substance use problems. Whether these same relation-
ships would also hold for males is a question in need of investigation.
4. Methodological Considerations
Recent reviews on the connections either between perpetrators' alcohol
problems and their victimization of children or between childhood victimiza-
tion and the development of alcohol problems identify the need for improve-
ments in research design (e.g., better control groups, multivariate rather than
bivariate analyses).5,101 This section covers three additional issues that need
to be carefully considered in future studies: definitions and measurement of
violence toward children, controversies regarding childhood memories of vio-
lence, and ethical and legal concerns.
4.1. Definitions and Measurement of Violence toward Children

There is great variation in how violence is defined. On one end of the
continuum are studies that rely on official records of founded cases to define
violence5 Strict definitions of physical and sexual abuse are used to identify
cases, evidence verifies existence of abuse, and authorities investigate each
case to ensure that legal criteria for physical or sexual abuse are met. Samples
based on these official records contain few false-positive cases. For official
data on incidence and prevalence of child abuse cases brought to the attention
of Child Protective Service agencies, the National Center on Child Abuse and
Neglect has established a national data collection and analysis program (Na-
tional Child Abuse and Neglect Data System).102 This program is authorized
under Public Law 102-295, the Child Abuse, Domestic Violence, Adoption
and Family Services Act of 1992.102
However, most incidents of physical and/or sexual abuse are not offi-
cially investigated, and thus officially identified cases are not representative
of all victimized children. In 1991, states received reports on 2.7 million chil-
dren,102 but estimates of the amount of parent-to-child violence are much
higher. Straus and Gelles103 report from their National Family Violence re-
survey of 1985 that an estimated 6.9 million children are severely assaulted
each year and an unknown number of children are sexually abused each year.
In particular, there may be a class bias as to which children will be reported
and found to be a physically or sexually abused child. Given that states have
different laws defining child abuse and childhood sexual abuse, there will be
differences in definitions across states. Furthermore, official records are often
incomplete reports and missing data can pose problems.
At the other end of the continuum are the studies that have chosen a
more inclusive range of experiences to represent violence. Gelles and
Straus104 propose that violence be viewed as “an act carried out with the
intention or perceived intention of physically hurting another person.” Be-
cause corporal punishment is carried out with the intent to cause physical
pain, Straus105 identifies corporal punishment as a form of parent-to-child
violence. To provide information for these more inclusive definitions of
physical and sexual abuse of children, data from self-reports are gathered. For
self-reports, measurement tools require specificity regarding types of behav-
iors that are considered violent. This specificity is preferable to global ques-
tions or questions that label behavior as “physical abuse” or “sexual ab-
use.”20,43,106 Also, global questions concerning physical or sexual abuse may
result in underestimates of these events, especially if respondents are asked
to label their own experiences as abuse. For this reason, researchers have
constructed instruments that identify a range of specific behaviors for identi-
fying physical and sexual abuse.
For measures of physical (nonsexual) violence, the CTS107,108 continues
to be widely accepted for measuring parent-to-child violence.* Three general
sections are included in the CTS: rational discussion, termed Reasoning; ver-
bal or nonverbal acts that symbolically hurt the other, termed Verbal Aggres-
sion; and use of physical aggression, termed Violence.109 The Violence sub-
scale is further divided into Minor and Severe Violence. The CTS captures
frequency of the behavior within a year’s time frame (although this time
frame is often altered to meet the needs of the research project). Perhaps
* For further information, contact Family Research Laboratory, University of New Hampshire,
Durham, NH 03824.
because of its wide adoption in many different contexts, there are criticisms of
the CTS, including: its framing of violence in response to conflicts, limitations
in the set of violent acts, assessing threats as forms of violence, inaccuracies of
self-reports over a 1-year period, equating acts that differ greatly in serious-
ness, failure to consider context and who initiates violence, distinctions made
between minor and severe violence, how gender differences for specific be-
haviors impact outcomes, and inability of the CTS to determine the process
and sequence of violence.109 In addition, the CTS is not designed to deter-
mine short- or long-term effects of child abuse. Despite these limitations, the
CTS has been one of the most widely used measures of family violence, and
data from two national studies (1975 and 1985) provide an excellent oppor-
tunity for comparing rates in specific samples with the general population.110
Recently, the authors have created a revised version of the parent–child CTS
(PCCTS), which provides more appropriate reasoning and psychological
abuse scales for parent–child relationships. The PCCTS also has three supple-
mental scales: Neglect, Sexual Abuse, and Injury.
Researchers may choose to avoid knowing about actual parent-to-child
violence in order to avoid having to report this violence to legal authorities,
thereby violating the promise of confidentiality to respondents (see Section
4.3). Other measures can be considered when measures of potential for vio-
lence, rather than actual violence, are appropriate. For example, the Child
Abuse Potential scale11 provides a measure of parental propensity for vio-
lence toward children, based on parent characteristics. The Parental Punitive-
ness Scale111 identifies a number of punitive responses toward hypothetical
situations involving children's misbehavior and has been revised to include
more severe forms of punishment.67
Measures of childhood sexual abuse have largely been developed for the
retrospective assessment of victimization.20,43,112 As with physical abuse,
there is great variation in the definition of sexual abuse. Defining the relation-
ship of perpetrators and victims, requiring age differences between the per-
petrators and the victims (e.g., 5-year age differences), and whether to in-
clude perceived consensual events (e.g., boyfriends who are more than 5
years older) are crucial to establishing a definition of childhood sexual abuse.
As with childhood physical abuse, measuring childhood sexual abuse re-
quires careful specification of behaviors, including suggestions of sexual ac-
tivity, touching, as well as penetration, without defining such behaviors as
sexual abuse. In addition, the operational definitions of sexual abuse will
differ between boys and girls. Identifying specific behaviors that have oc-
curred produce higher prevalence rates than do single screening ques-
tions. 20,113
During the research process, sensitivity must be given to the impact of
respondent's having revealed traumatic events. In interview settings, individ-
uals may reveal experiences that have not been previously shared and may
display emotions (e. g., crying, anger) that require sensitive handling. Recall-
ing experiences may exacerbate posttraumatic stress symptoms, and a list of
15 • Alcohol–Violence in Children's Lives 375
community resources that provide counseling may be appropriate for all par-
ticipants, regardless of whether they report histories of violence or not. The
research staff must be carefully trained to address these situations to mini-
mize harm to individuals and to maintain integrity of the research process.
Comparisons of self-administered questionnaires and face-to-face inter-
views in a retrospective survey of childhood sexual abuse suggest that inter-
views are not necessarily superior to self-administered formats in assisting in
disclosure of childhood sexual abuse.113 A small group of women (10%) re-
ported childhood sexual abuse during a self-administered questionnaire but
not during the face-to-face interviews. Similarly, 12.4% of the women report-
ed childhood sexual abuse during the face-to-face interviews but not during
the self-administered questionnaire. However, interviews assist in clarifying
different behaviors and provide better detail regarding the incidents of child-
hood sexual abuse.113 Childhood sexual abuse is perpetrated under a veil of
secrecy. Even after the child becomes an adult, talking about these secrets
may be difficult. Thus, individuals who have abuse histories may assess more
carefully who they tell and whether there is a compelling reason to tell the
person asking about the experience. Researchers need to create an atmo-
sphere of respect and privacy to provide a safe setting where these secrets can
be shared discreetly. Furthermore, respondents may question why the re-
searcher needs to know this intimate and private information and the re-
search team should be prepared to respond to these questions.
In this chapter, our definition of violence toward children focuses on acts
of both physical and sexual abuse; we have not reviewed neglect, psychologi-
cal abuse, or the impact of witnessing violence. These additional acts of com-
mission or omission are also important potential impacts on children's lives
that are interwoven with physical and sexual abuse. Furthermore, studies on
violence toward children should consider the impact of multiple forms of
victimization from different perpetrators, over extended periods of time.
4.2. Controversies Regarding Childhood Memories of Violence

Much of what is known about the connections between substance prob-
lems and childhood victimization is based on retrospective reports from
adults. Three major concerns have emerged for such retrospective data. First,
there is a concern that adults do not necessarily remember or accurately
remember their experiences in childhood.5 There is some indication that the
failure to remember is specifically connected to traumatic events; individuals
repress painful memories as a coping mechanism.114 Empirical evidence ex-
ists that demonstrates that adults do fail to report their childhood victimiza-
tion experiences. In a sample of women substance abusers (n = 105) in outpa-
tient treatment who reported histories of CSA, 19% stated that they had
forgotten the abuse at some point in their lives.112 In a recent study of offi-
cially reported cases of CSA, reinterviewed 17 years later as adults, Wil-
liams115 found that a large proportion of women (38%) did not recall or chose
not to report the index case of sexual abuse.* Even among the women who
did remember the index incident, 16% reported that there was some period of
time in the past when they did not remember the incident. Being younger in
age at the time of abuse and having a greater level of force used in the assault
were positively related to failure to remember. Failure to recall victimization
experiences will result in false-negative cases within the control group, there-
by attenuating differences between groups.
Another concern is that memories that are retrieved may be false or
constructed. Loftus116 proposes that both popular writings and therapists’
suggestions can create false memories. The basis for her argument is largely
anecdotal evidence. Although “false recall” has been raised largely within
therapeutic settings, careful consideration should be given to prevent such
concerns from emerging within research settings. Questions concerning
childhood sexual abuse, asked after rapport has been established but early in
the questionnaire/interview, may avoid opportunities for “created memories”
during the research process, particularly in studies that involve multiple fol-
low-ups. Providing interviewers with training to ensure that they impart
empathy without suggesting or leading respondents is another important
concern, particularly when open-ended questions are used to probe for more
detailed information. Careful selection of the research team to avoid potential
assistants who may be overzealous in finding reported cases of child victimiz-
ation is essential.
The final concern is that subjects may refuse to tell researchers about
their victimization. Koss106 suggests that matching sex and ethnicity may
improve willingness to divulge information about prior sexual victimization.
Providing confidential settings that are comfortable and safe may also pro-
mote willingness to report victimization. Concerns about adequacy of recall
are important for any research project relying on self-report data. Such con-
cerns should not devalue the importance of individual reports about their life
experiences. Especially with regard to long-term consequences, memories
that are recalled may have more relevance to later impact on other life events
than does a “factual” accounting of the event.1
4.3. Ethical and Legal Issues

A number of ethical and legal issues are involved in conducting research
with or about children who may be experiencing or have experienced violent
victimization. For instance, asking parents or other adult figures to report on
their use of violence toward children presents a dilemma when these respon-
dents report violence. According to Socolar and Amaya-Jackson,117 health
care professionals, teachers, and other child care professionals are required
* Williams did not reveal the true purpose of the follow-up interview and some of the failure to
report may not be due to lack of memory but rather a conscious choice not to report.
by law to report suspected maltreatment disclosed by subjects in all states,

and in nearly half of the states anyone suspecting child maltreatment is re-
quired to report. Socolar and Amaya-Jackson117 report different techniques
used to avoid such dilemmas, including pursuing anonymous data. Use of
anonymous data precludes the possibility of follow-up research designs or
matching individuals with official data sources. Regardless of the legal re-
quirements, researchers have ethical responsibilities for reporting suspected
child abuse, and some would argue that researchers have an ethical respon-
sibility to find out about abuse.
Institutional Review Boards require that subjects be informed of risks that
may result from participation in the research project. Parents or other adult
figures, once informed that their answers to questions concerning violence
could potentially result in official action against them, may refuse to partici-
pate. Another research concern is the bias that occurs if individuals who
perpetrate violence against children disproportionately refuse to participate.
Even among participants who choose to participate, a chilling effect on the
research process may be expected as participants weigh their answers for the
possible actions that might be taken. If a researcher informs parents that
children who participate in a research project may be at risk for intervention
by the authorities in cases where there is a suspicion of child abuse, children
who are victims of parental violence may not be allowed to participate.
Researchers have used various techniques to avoid such dilemmas, such
as limiting questions asked. As mentioned earlier, measurements can be used
that provide information about punishments for hypothetical situations
involving children’s misbehavior. Hypothetical situations can include re-
sponses that are both appropriate and abusive. Answers that parents give to
hypothetical situations represent parental attitudes regarding what is socially
acceptable, but may not measure propensity for violence adequately. Re-
searchers may ask about punishments, excluding punishments that would be
defined legally as child abuse, but that includes corporal punishment. Finally,
subjects may be warned not to divert from answering the specific questions
asked. Interviewers can be trained to warn or “head off ” any divergence from
the established questions that might lead to potential conflicts between confi-
dentiality and protection of children.
Another difficulty for researchers involved in prospective designs is that
identification of children who are suspected cases of child abuse will result in
some type of intervention. Any intervention may impact the long-term out-
comes under investigation. The specific issue of whether substance problems
are one of the consequences of untreated childhood victimization is difficult
to address in prospective designs that identify children who experience
abuse. Identifying samples of official cases of child abuse avoids some of the
problems of other approaches in so far as the researcher does not have to
report the case. However, children may reveal additional abuse not known to
authorities.
5. Future Directions
5.1. Considerations for Research Designs and Data Analyses
5.1.1. Developmental Victimology. Child development has received rela-
tively little attention in the study of childhood Victimization. Recently, a
framework for examining “developmental victimology” has been proposed
with two separate domains.118 First, there are developmental aspects of risk,
which are typified by the changing risks children encounter as a result of their
age and level of development. Second, there are developmental aspects of
impact, which identify variation in responses to victimization dependent on
the developmental stage. Children at different stages of development have
differing capacities for resiliency and differing vulnerabilities.119,120 The role
of alcohol or drugs in perpetrating childhood violence and the role of child-
hood violence in developing alcohol or drug abuse problems may operate
differentially, depending on the developmental stage of the child at the time
of the event. Thus, future studies should consider how risks for violence
might be assessed in view of the child’s age. Furthermore, to assess develop-
mental aspects of consequences, age of the child at the time of victimization
and duration of victimization across developmental stages need to be consid-
ered. Males and females may respond differently to victimization experiences
at the same developmental stage. Finally, children at different developmental
stages will report on victimization experiences differently. Assessments of
violence that are appropriate for different stages of development are needed.
5.1.2. Examination of Mother and Father Violence Separately. Results from
studies7,43 indicate differences in impact between mother and father violence
toward children. Collecting data to allow for separate analyses would benefit
and further our understanding of this issue. There are several possible expla-
nations for why this difference may exist. Mother and father violence are
qualitatively different both in terms of threat and harm caused. Mother vio-
lence may occur in the context of numerous mother–child interactions, while
father violence may occur within the context of fewer total interactions. Moth-
ers may not be able to protect children from father violence, but fathers can
protect children from mother violence, subsequently affecting the long-term
impact of violence on children’s lives. Finally, parent violence from the same-
sex parent may have different meanings than parent violence from the oppo-
site-sex parent.
5.1.3. Inclusion of Multiple Forms of Victimization. Investigations of the
relationships between childhood violence and alcohol or drugs have often
focused on only one form of victimization. However, research indicates that
multiple forms of violent victimization often occur in children’s lives.43,21
More research is needed that includes both physical abuse and sexual abuse
histories. Consideration needs to be given to the number of different types of
perpetrators that may victimize children. Children may have multiple parent
figures during their lives and multiple primary caregivers. Exposure to moth-
er’s boyfriends and extended families including siblings of different ages can
also broaden the range of risk to children. As discussed earlier, parental
alcohol and drug problems can impact the protection of children from violent
victimization. Further investigation of this parental role omission should con-
sider the vulnerability to different types of perpetrators for these children. In
addition, impact of violent victimization from multiple perpetrators may have
differential impact on the long-term consequences, including drug or alcohol
abuse, for children.
5.1.4. Contexts and the Meanings Associated with the Victimization. One of
the least developed areas of this research has been the contexts of experiences
and meanings associated with victimization. For example, assumptions are
sometimes made that physically invasive sexual acts (e.g., sexual intercourse)
are more harmful than less invasive sexual acts (e.g., showing sexual parts of
the body). Yet, when asked to choose the most traumatic or upsetting sexual
experience (when multiple experiences are present), women have reported
that the context of the experience can make the less physically intrusive
experience more traumatic. Research that provides in-depth exploration of
the meanings of these events is needed. This type of research is often avoided
because of the lengthy process of analyzing and coding open-ended data.
However, understanding the meanings of these events to people’s lives can
provide a better understanding of connections between alcohol, drugs, and
childhood violence.
5.2. Summary
Our understanding of the relationships between alcohol, drugs, and
childhood victimization has been aided by the growth of studies on family
violence and child maltreatment and the development of alcohol and drug
research in general. Combining these two major arenas into a research agen-
da requires a comprehensive understanding of both fields of research and
how they overlap as well as differ. The prospect of pursuing research on these
questions can seem overwhelmingly difficult because of the complexity of the
issues and the diversity of subjects who are affected by these issues. Yet, a
careful review of people’s lives and the problems that they present indicates
the relevance of this research. Histories of childhood violence and substance
abuse problems overlap within generations and reappear in subsequent gen-
erations. Investigations that examine either cycles of violence or cycles of
addiction do not provide adequate understanding of synergistic effects. There
is a need to study these problems in combination and across generations.
Although our knowledge about the relationships between AOD prob-
lems and violence toward children is far from adequate, there is evidence that
perpetration of violence toward children may be related to the perpetrator’s
AOD problems, and that parental AOD problems may increase the vul-
nerability of children for victimization from others. More recently, studies

have focused on the relationships between victimization experiences and the
development of AOD problems in subsequent years, and results indicate
support for this relationship, as well. Given these results, intergenerational
studies of AOD problems may also want to include investigation of inter-
generational patterns of violence. Furthermore, treatment programs may
need to consider the impact of victimization experiences on initiation and
maintenance of dependence on alcohol and drugs. Finally, preventing vio-
lence may help prevent future AOD problems.
References
1. Miller BA: Investigating links between childhood victimization and alcohol problems, in
Martin SE (ed): Alcohol and Interpersonal Violence: Fostering Multidisciplinary Perspectives,
Monograph 24. 1993, pp 315–323. Rockville, MD: US Government Printing Office.
2. Miczek KA, DeBold JF, Haney M, et al: Alcohol, drugs, and violence, in Reiss AJ, Roth JA
(eds): Understanding and Preventing Violence: vol 3: Social and Psychological Perspectives on
Violence. Washington, DC, National Academy Press, 1993, pp 377–570.
3. Orme TC, Rimmer J: Alcoholism and child abuse: A review. J Stud Alcohol 42:273–287, 1981.
4. Black R, Mayer J: Parents with special problems: Alcoholism and opiate addiction. Child
Abuse Negl 4:45–54, 1980.
5. Widom CS: Child Abuse and Alcohol Use and Abuse: Alcohol and Interpersonal Violence: Fostering
Multidisciplinary Perspectives (Research Monograph Series 24). Rockville, MD, NIAAA US
Department of Health and Human Services, 1992, pp 291–314.
RL, Bellack AS, Hersen J (eds): Handbook of Family Violence. New York, Plenum Press, 1988,
pp 383–406.
7. Reider E, Zucker RA, Maguin ET, et al: Alcohol involvement and violence towards children
among high risk families. Paper presented at the 97th Annual Meeting of the American
Psychological Association, New Orleans, LA, August 1989.
8. Smyth NJ, Miller BA, Janicki PI, et al: Mothers’ protectiveness and child abuse: The impact
of her history of childhood sexual abuse and an alcohol diagnosis. Paper presented at the
American Society of Criminology Conference, Boston, MA, November 1995.
Garden City, NJ, Anchor Books, 1980.
10. Miller BA, Downs WR, Testa M: The relationship between women’s alcohol problems and
experiences of childhood violence. Paper presented at the Annual Meeting of the American
Psychological Association, Boston, MA, 1990.
11. Milner JS, Wimberly RC: An inventory for the identification of child abusers. J Clin Psychol
35:95–100, 1979.
12. Holmes SJ, Robins LN: The role of parental disciplinary practices in the development of
depression and alcoholism. Psychiatry 51:24–36, 1988.
13. Helzer JE, Pryzbeck TR: The co-occurrence of alcoholism with other psychiatric disorders in
the general population and its impact on treatment. J Stud Alcohol 49:219–224, 1988.
14. Radomsky NA: The association of parental alcoholism and rigidity with chronic illness and
abuse among women. J Fam Prac 35:54–60, 1992.
15. Downs WR, Miller BA: Intergenerational links between childhood abuse and alcohol-related
problems, in Harrison L (ed): Alcohol Problems in the Community. London, Routledge Press,
1995, pp 14–51.
16. Andreasen NC, Endicott J, Spitzer RL, Winokur G: The family history method using diag-
nostic criteria. Arch Gen Psychiatry 34:1229–1235, 1977.
17. Miller BA: The interrelationships between alcohol and drugs and family violence, in De La
Rosa M, Lambert EY, Gropper B (eds): Drugs and Violence: Causes, Correlates, and Conse-
quences, Monograph 103 (DHHS Publication No. ADM-90-1721), Rockville, MD, National
Institute on Drug Abuse, 1990, pp 177-207.
18. Aarens M, Cameron T, Roizen J, et al: Alcohol, Casualties and Crime. Berkeley, CA, Social
Research Group, 1978.
19. Yama MF, Fogas BS, Teegarden LA, Hastings B: Childhood sexual abuse and parental
alcoholism: Interactive effects in adult women. Am J Orthopsychiatry 63:300–305, 1993.
20. Finkelhor D: Sexually Victimized Children. New York, Free Press, 1979.
21. Windle M, Windle RC, Scheidt DM, et al: Physical and sexual abuse and associated mental
disorders among alcoholic inpatients. Am J Psychiatry 152:1322–1328, 1995.
22. Pernanen, K: Alcohol and crimes of violence, in Kissin B, Begleiter H (eds): Social Aspects of
Alcoholism. New York, Plenum Press, 1976, pp 351–444.
23. Steele CM, Josephs RA: Alcohol myopia: Its prized and dangerous effects. J Stud Alcohol
45:921–933, 1990.
24. Zeichner A, Pihl RR: Effects of alcohol and instigator intent on human aggression. J Stud
Alcohol 41:265–276, 1980.
25. Collins JJ: Drinking and violations of the criminal law, in Pittman DJ, White HR (eds):
Society, Culture, and Drinking Patterns Reexamined. New Brunswick, NJ, Rutgers Center of
Alcohol Studies, 1991, pp 650–660.
26. Kantor GK, Straus MA: Parental drinking and violence and child aggression. Paper pre-
sented at the Annual Meeting of the American Psychological Association, Boston, MA,
August 1990.
27. Coleman DH, Strauss MA: Alcohol abuse and family violence, in Gottheil E, Druley KA,
Skoloda TE, Waxman HM (eds): Alcohol Abuse and Family Violence. Springfield, IL, Thomas,
1983, pp 104–124.
28. Gelles RJ: The Violent Home. Beverly Hills, CA, Sage, 1972.
29. Lang AR, Goeckner DJ, Adesso VJ, Marlatt GA: Effects of alcohol on aggression in male
social drinkers. J Abnorm Psychol, 84:508–518, 1975.
30. MacAndrew C, Edgerton RB: Drunken Comportment: A Social Explanation. Chicago, IL, Al-
dine, 1969.
31. Wikler A: Mechanisms of action of drugs that modify personality function. Am J Psychiatry
108:590–599, 1952.
32. Shuntich RJ, Taylor SP: The effects of alcohol on human physical aggression. J Exp Res
Personal 6:34–38, 1972.
33. Fitzpatrick JP: Drugs, alcohol and violent crime. Addict Dis 1:353–357, 1974.
34. Fagan J: Interactions among drugs, alcohol, and violence. Violence Public Health 12:65–79,
1993.
35. Gelles RJ: Alcohol and other drugs are associated with violence-They are not its cause, in
Gelles RJ, Loseke DR (eds): Current Controversies on Family Violence. Newbury Park, CA,
Sage, 1993, pp 182–196.
36. White HR: The drug use-delinquency connection in adolescence, in Weisheit R (ed): Drugs,
Crime, and the Criminal justice System. Cincinnati, OH, Anderson, 1990, pp 215–256.
37. Pernanen K: Theoretical aspects of the relationship between alcohol use and crime, in
Collins JJ (ed): Drinking and Crime. New York, Guilford Press, 1981, pp 1-69.
38. Graham K: Theories of intoxicated aggression. Can J Behav Sci 12:141–158, 1980.
39. Straus MA: Discipline and deviance: Physical punishment of children and violence and
other crime in adulthood. Soc Problems 38:133–152, 1991.
40. Finkelhor D, Baron L: High-risk children, in Finkelhor D (ed): Sourcebook on Child Sexual
Abuse. Newbury Park, CA, Sage, 1986, pp 60-88.
41. Faller CK: Child Sexual Abuse: An Interdisciplinary Manual for Diagnosis, Case Management, and
Treatment. New York, Columbia University Press, 1988.
42. Miller BA, Downs WR, Gondoli DM, Keil A: The role of childhood sexual abuse in the
development of alcoholism in women. Violence Vict 2:157–172, 1987.
43. Miller BA, Downs WR, Testa M: Interrelationships between victimization experiences and
women’s alcohol use. J Stud Alcohol Suppl 11:109–117, 1993.
44. Cole PM, Woolger C, Power TG, Smith KD: Parenting difficulties among adult survivors of
father–daughter incest. Child Abuse Negl 16:239–249, 1992.
45. Mian M, Maron P, LeBaron D, Birtwistle D: Familial risk factors associated with intrafamilial
and extrafamilial sexual abuse of three- to five-year-old girls. Can J Psychiatry 39:348–353,
1994.
46. Davis SK: Chemical dependency in women: A description of its effects and outcomes on
adequate parenting. J Subst Abuse Treat 7:225–232, 1990.
47. Beitchman JH, Zucker KJ, Hood JE, et al: A review of the long-term effects of child sexual
abuse. Child Abuse Negl 16:101–118, 1992.
48. Wyatt GE, Guthrie D, Notgrass CM: Differential effects of women’s child sexual abuse and
subsequential sexual revictimization. J Consult Clin Psychol 60:167–173, 1992.
to a random household sample of women. J Stud Alcohol 50:533–540, 1989.
50. Finkelhor D, Browne A: Assessing the Long-Term lmpact of Child Sexual Abuse: A Review and
Conceptualization. Newbury Park, CA, Sage, 1988.
51. Regier DA, Farmer ME, Rae DS, et al: Comorbidity of mental disorders with alcohol and
other drug abuse. J Am Med Assoc 264:2511–2517, 1990.
52. Jumper SA: A meta-analysis of the relationship of child sexual abuse to adult psychological
adjustment. Child Abuse Negl 19:715–728, 1995.
53. Silber S, Bermann E, Henderson M, Lehman A: Patterns of influence and response in
abusing and nonabusing families. J Fam Violence 8:27–38, 1993.
54. Newberger CM, Gremy IM, Waternaux CM, Newberger EH: Mothers of sexually abused
children: Trauma and repair in longitudinal perspective. Am J Orthopsychiatry 63:92–102,
1993.
55. West MO, Prinz RJ: Parental alcoholism and childhood psychopathology. Psychol Bull
102:204–218, 1987.
56. Liskow B, Powell BJ, Nickel E, Penick E: Antisocial alcoholics: Are they clinically significant
diagnostic subtypes? J Stud Alcohol 52:62–69, 1991.
57. Hesselbrock VM, Meyer RE, Keener RE: Psychopathology in hospitalized alcoholics. Arch
Gen Psychiatry 42:1050–1055, 1985.
58. Schuckit MA: Alcoholism and sociopathy: Diagnostic confusion. J Stud Alcohol 34:157–164,
1973.
60. Babor TF, Hofmann M, DelBoca FK, et al: Types of alcoholics, I. Evidence for an empirically
derived typology based on indicators of vulnerability and severity. Arch Gen Psychiatry
49:599–608, 1992.
61. DiLalla LF, Gottesman II: Biological and genetic contributions to violence: Widom’s untold
tale. Psychol Bull 109:125–129, 1991.
62. Muller RT, Fitzgerald HE, Sullivan LA, Zucker RA: Social support and stress factors in child
maltreatment among alcoholic families. Can J Behav Sci 26:438–446, 1994.
63. Downs WR, Smyth NJ, Miller BA: The relationship between childhood violence and alcohol
problems among men who batter: An empirical review and synthesis. Review of Aggression
and Violent Behavior, in press.
64. Straus MA, Kantor GK: Corporal punishment of adolescents by parents: A risk factor in the
epidemiology of depression, suicide, alcohol abuse, child abuse, and wife beating. Adoles-
cence 29:543–561, 1994.
65. Holmes SJ, Robins LN: The role of parental disciplinary practices in the development of
depression and alcoholism. Psychiatry 51:24–36, 1988.
66. Brown GR, Anderson B: Psychiatric morbidity in adult inpatients with childhood histories
of sexual and physical abuse. Am J Psychiatry 148:55–61, 1991.
67. Blane HT, Miller BA, Leonard KE et al: Intra- and lntergenerational Aspects of Serious Domestic
Violence and Alcohol and Drugs. Buffalo, NY, Research Institute on Alcoholism, 1988.
68. Widom CS, Ireland T, Glynn PJ: Alcohol abuse in abused and neglected children followed-
up: Are they at increased risk? J Stud Alcohol 56:207–217, 1995.
69. McCord J: A forty-year perspective on effects of child abuse and neglect. Child Abuse Negl
7:265–270, 1983.
70. Dembo R, Dertke M, La Voie L, et al: Further examination of the association between heavy
marijuana use and crime among youths entering a juvenile detention center. J Psychoactive
Drugs 19:361–373, 1987.
71. Dembo R, Dertke M, Borders S, et al: The relationship between physical and sexual abuse
and tobacco, alcohol, and illicit drug use among youths in a juvenile detention center. Int J
Addict 23:351–378, 1988.
72. Dembo R, Williams L, Wish ED, et al: The relationship between physical and sexual abuse
and illicit drug use: A replication among a new sample of youths entering a juvenile deten-
tion center. Int J Addict 23:1101–1123, 1988.
73. Dembo R, Williams L, Schmeidler J, Wothke W: A longitudinal study of the predictors of the
adverse effects of alcohol and marijuana/hashish use among a cohort of high risk youths. Int
J Addict 28:1045–1083, 1993.
74. Downs WR, Miller BA, Gondoli DM: Childhood experiences of parental physical violence
for alcoholic women as compared with a randomly selected household sample of women.
Violence Vict 2:225–240, 1987.
75. Browne A, Finkelhor D: Impact of child sexual abuse: A review of the literature. Psychol Bull
99:66–77, 1986.
76. Beitchman JH, Zucker KJ, Hood JE, et al: A review of the short-term effects of child sexual
abuse. Child Abuse Negl 15:537–556, 1991.
77. Polusny MA, Follette VM: Long-term correlates of child sexual abuse: Theory and review of
the empirical literature. Appl Prevent Psychol 4:143–166, 1995.
78. Wilsnack SC, Klassen AD, Vogeltanz ND: Childhood sexual abuse and women’s substance
abuse: National survey findings. Paper presented at the American Psychological Associa-
tion Conference, Psychosocial and Behavioral Factors in Women’s Health: Creating an
Agenda for the 21st Century. Washington, DC, May 1994.
79. Russell DEH: The prevalence and seriousness of incestuous abuse: Stepfathers vs. biological
fathers. Child Abuse Negl 8:15–22, 1984.
80. Wyatt GE: The sexual abuse of Afro-American and white American women in childhood.
Child Abuse Negl 9:507–519, 1985.
81. Mullen PE, Martin JL, Anderson JC, et al: Childhood sexual abuse and mental health in
adult life. Br J Psychiatry 163:721–732, 1993.
82. Miller BA, Downs WR: Violent victimization among women with alcohol problems, in
Galanter M (ed): Recent Developments in Alcoholism, vol 12 : Women and Alcoholism. New York,
Plenum Press, 1995, pp 81–101.
83. Stein JA, Golding JM, et al: Long-term psychological sequelae of child sexual abuse: The Los
Angeles epidemiologic catchment area study, in Wyatt GE, Powell GJ (eds): Lasting Effects of
Child Sexual Abuse. Newbury Park, CA, Sage, 1988, pp 135–154.
84. Harrison PA, Hoffman NG, Edwall GE: Differential drug use patterns among sexually
abused adolescent girls in treatment for chemical dependency. Int J Addict 24:499–514,
1989.
85. Malinosky-Rummell R, Hansen DJ: Long-term consequences of childhood physical abuse.
Psychol Bull 114:68–79, 1993.
86. Kendall-Tackett KA, Meyer Williams L, Finkelhor D: Impact of sexual abuse on children: A
review and synthesis of recent empirical studies. Psychol Bull 113:164–180, 1993.
87. Boney-McCoy S, Finkelhor D: Psychosocial sequelae of violent victimization in a national
youth sample. J Consult Clin Psychol 63:726–736, 1995.
88. McCloskey LA, Figueredo AJ, Koss MP: The effects of systemic family violence on children’s
mental health. Child Dev 66:1239–1261, 1995.
89. Lanktree C, Briere J, Zaidi L: Incidence and impact of sexual abuse in a child outpatient
sample: The role of direct inquiry. Child Abuse Negl 15:447–453, 1991.
90. Becker JV, Kaplan MS, Tenke CE, Tartaglini A: The incidence of depressive symptomatol-
ogy in juvenile sex offenders with a history of abuse. Child Abuse Negl 15:531–536, 1991.
91. Kaplan HB: Self-esteem and self-derogation theory of drug abuse, in Lettieri DJ, Sayers M,
Wallenstein Pearson H (eds): Theories on Drug Abuse (Research Monograph Series 30). Rock-
ville, MD, NIDA US Department of Health and Human Services, 1980, pp 128–131.
92. Hawkins JD, Catalano RF, Miller JY: Risk and protective factors for alcohol and other drug
problems in adolescence and early adulthood: Implications for substance abuse prevention.
Psychol Bull 112:64–105, 1992.
93. Dishion TJ, Capaldi D, Spracklen KM, et al: Peer ecology of male adolescent drug use. Dev
Psychopathol 7(4):803–824, 1995.
94. Miller WR, Brown JM: Self-regulation as a conceptual basis for the prevention and treatment
of addictive behaviors, in Heather N, Miller WR, Greeley J (eds): Self-control and the Addictive
Behaviors. Sydney, Australia, Macmillan, 1991, pp 3–79.
95. Crime Victims Research and Treatment Center: A Report to the Nation. Arlington, VA, Na-
tional Victim Center, 1992.
96. Edwards G, Hensman C, Peto J: A comparison of female and male motivation for drinking.
Int J Addict 8:577–587, 1973.
97. Fillmore KM: Drinking and problem drinking in early adulthood and middle age. Q J Stud
Alcohol 35:819–840, 1974.
98. Fillmore KM: Relationships between specific drinking problems in early adulthood and
middle age. J Stud Alcohol 36:882–907, 1975.
99. Chassin L, Rogosch F, Barrera M: Substance use and symptomatology among adolescent
children of alcoholics. J Abnorm Psychol 100:449–463, 1991.
100. Lewis CE, Bucholz KK. Alcoholism, antisocial behavior and family history. Br J Addict
86:177–194, 1991.
101. Leonard KE, Jacob T: Alcohol, alcoholism, and family violence, in VanHasselt V, Morrison
RL, Bellack AS, Hersen M (eds): Handbook of Family Violence. New York, Plenum Press, 1988,
pp 383–406.
102. National Center on Child Abuse and Neglect: National Child Abuse and Neglect Data System:
Working Paper 2—1991 Summary Data Component. Washington, DC: US Government Printing
Office, 1992.
103. Straus MA, Gelles RJ: Physical Violence in American Families: Risk Factors and Adaptations to
Violence in 8,145 Families. New Brunswick, NJ, Transaction Publishers, 1990.
104. Gelles RJ, Straus MA: Determinants of violence in the family: Toward a theoretical integra-
tion, in Burr WR, Hill R, Nye FI, Reiss IL (eds): Contemporary Theories about the Family, vol 1.
New York, Free Press, 1979, pp 549–581.
105. Straus MA: Beating the Devil Out of Them: Corporal Punishment in American Families. New York,
Lexington Books, 1994.
106. Koss MI': Detecting the scope of rape: A review of prevalence research methods. J Interper
Violence 8:198–222, 1993.
107. Straus MA: Measuring intrafamily conflict and violence: The conflict tactics (CT) scales.
J Marriage Fam 41:75–88, 1979.
108. Straus MA: The national family violence surveys, in Straus MA, Gelles RJ (eds): Physical
Violence in American Families: Risk Factors and Adaptations to Violence in 8,245 Families. New
Brunswick, Transaction Publishers, 1990, pp 3–16.
109. Straus MA: The Conflict Tactics scales and its critics: An evaluation and new data on validity
and reliability, in Straus MA, Gelles RJ (eds): Physical Violence in American Families: Risk
Factors and Adaptions to Violence in 8,245 Families. New Brunswick, Transaction Publishers,
1990, pp 49–74.
110. Straus MA, Gelles RJ (eds): Physical Violence in American Families: Risk Factors and Adaptions to
Violence in 8,145 Families. New Brunswick, Transaction Publishers, 1990.
111. Epstein R, Komorita SS: The development of a scale of parental punitiveness toward aggres-
sion. Child Dev 36:129–142, 1965.
112. Loftus EF, Polonsky S, Fullilove MT: Memories of childhood sexual abuse: Remembering
and repressing. Psychol Women Q 18:67–84, 1994.
113. Martin J, Anderson J, Romans S, et al: Asking about child sexual abuse: Methodological
implications of a two stage survey. Child Abuse Negl 17:383–392, 1993.
114. Herman JL: Trauma and Recovery. New York, Basic Books, 1992.
115. Williams LM: What does it mean to forget child sexual abuse? A reply to Loftus, Garry, and
Feldman (1994). J. Consult Clin Psychol 62:1182–1186, 1994.
116. Loftus EF: The reality of repressed memories. Am Psychol 48:518–537, 1993.
117. Socolar RRS, Amaya-Jackson L: Methodological and ethical issues related to studying child
maltreatment. J Fam Issues 16:565–586, 1995.
118. Finkelhor D: The victimization of children: A developmental perspective. Am J Orthopsychia-
try 65:177–193, 1995.
119. Cicchetti D, Toth SL: A developmental psychopathology perspective on child abuse and
neglect. J Am Acad Child Adolesc Psychiatry 34:541–565, 1995.
120. Newberger CM, DeVos E: Abuse and victimization: A lifespan perspective. Am J Orthopsy-
chiatry 58:505–511, 1988.
16
Issues in the Linkage of Alcohol

and Domestic Violence Services
James J. Collins, Larry A. Kroutil, E. Joyce Roland,
and Marlee Moore-Gurrera
Abstract. It is well established that alcohol is a risk factor for male against female domestic
violence. Some evidence also suggests that some women victims of domestic violence develop
substance abuse problems in response to their victimization. Although interpretations vary re-
garding the exact nature of the relationship of substance abuse and domestic violence offending
and victimization, there is evidence that linking substance abuse and domestic violence services
could have a positive impact on batterer cessation and victim support services. Currently, how-
ever, service linkage for the two problems is rare. There are major barriers to linkage of substance
abuse and domestic violence services, including philosophical differences of treatment perspec-
tive between program types and structural impediments that make linkage difficult. The chapter
discusses the barriers to linkage, examines potentially useful linkage models, and suggests the
next steps to examine the feasibility of linking services for the two problems.
1. Introduction
Violence within the family is an ancient problem, but one that was largely
viewed as a private problem until the past 20 or 25 years. Particularly impor-
tant to the increased visibility of domestic violence was the first nationally
representative survey of violence in the American family conducted in the
mid-1970s. The survey found that 16% of those surveyed reported some kind
James J. Collins, Larry A. Kroutil, E. Joyce Roland, and Marlee Moore-Gurrera • Health and
Social Policy Division, Research Triangle Institute, Research Triangle Park, North Carolina
27709-2194.
387
of violence between spouses in the year of the survey, while 28% reported
marital violence at some time during the marriage.1 This survey and numer-
ous subsequent national and less comprehensive studies have focused atten-
tion on the family violence problem and have provided data to study its
characteristics.
The reduction of violence against women by their male partners is a
priority health objective for the United States2 and was a major focus of the
1994 federal crime legislation. Violence against women is believed to affect
approximately 4 million women annually in the United States.3 Violence
against women is considered to be the leading cause of nonfatal injury for
women, and an estimated 1 million women each year seek emergency treat-
ment for injuries sustained during battering.4 This figure probably underesti-
mates the number of women with battering-related treatment needs because
battering goes largely undiagnosed as a potential cause of women’s injuries or
health problems.5–9 National victimization data indicate that only 56% of
incidents where women are attacked by an intimate are reported to police.10
Research also indicates that sustained exposure to violence leads to the devel-
opment of chronic and acute health problems, including chronic pain,8,9,11–13
miscarriage,14,15 irritable bowel syndrome,16 depression and anxiety,17–20 and
suicide.5,8,21 Too often, battering culminates in homicide.22,23 In addition, and
of significant importance, battered women are at higher risk for inappropriate
use of prescription drugs, illicit drugs, and alcohol.8,12,21 In fact, battering is
recognized as a major risk factor for substance abuse and mental illness
among women in the United States.2
In this chapter, we discuss the role of substance abuse in domestic vio-
lence and the implications that the association between substance abuse and
domestic violence have for programs that deal with these problems. There are
good reasons to think that a substance abuse–domestic violence program
service linkage would be beneficial to clients. Such linkage is infrequent,
however, and there are significant impediments to linkage. The chapter ad-
dresses these issues based on a review of the literature in the substance abuse
and domestic violence fields and on interviews conducted with service pro-
viders in these fields.
2. Role of Alcohol in Domestic Violence

One of the aspects of domestic violence receiving attention has been the
relationship of drinking to its occurrence. Three studies using data from two
national domestic violence surveys found that the frequency of the husband’s
drinking was associated with wife abuse.24–26 Evidence of the magnitude of
the relationship between alcohol use and domestic violence varies widely,
with statistics ranging from as low as 25% to as high as 80%.24–31
Studies of battered women indicate that their partner’s alcohol use plays
a role in domestic violence. In a study by Rounsaville,32 29% of the battered
women reported that their partners were drinking when the violent episode
16 • Alcohol–Domestic Violence Services Linkage 389
occurred. Similarly, Gayford33 reported that the male partner was intoxicated
at the time of the abuse 44% of the time; and Carlson34 found that alcohol was
involved in 67% of abusive incidents. Roy35 also found that more than 80% of
the men who were occasional drinkers were inclined to abuse their partners,
but only when they had been drinking. Several studies of women who were
in shelters or used crisis hotlines indicate that there is an association between
alcohol and domestic violence.36 For example, data from New York‘s Abused
Women’s Aid in Crisis indicate that alcohol abuse by the husband is an
underlying factor in over 80% of domestic violence cases. Similarly, a study in
a Michigan emergency shelter indicates that over 66% of the assaults were
associated with alcohol use. Furthermore, data from a Philadelphia commu-
nity hotline for abused women were examined, and 55% of the callers said
that their partners had become abusive when the latter had been drinking.
There is some evidence that drinking on the part of women increases the
likelihood of domestic violence victimization. In a multivariate analysis of
national survey data, Kantor and Straus26 found that women who drink
heavily have a higher risk of being the victim of minor (but not severe)
violence by their partners. However, the bulk of empirical research indicates
that only a small percentage of domestic violence events involve situations in
which the woman was the sole drinker. The most common patterns are both
partners drinking, no drinking, and only male partner drinking (see review
by Hamilton and Collins,28 pp. 261–275).
Substance use by domestic violence victims has been suggested to be a
response to the violence they suffer. The substance use may be used to cope
with the physical and emotional pain after abuse.37,38 As stated previously,
there is evidence that women who are the victims of domestic violence are at
increased risk of abusing alcohol, illicit drugs, and prescription medication.2
However, limited research is available documenting the prevalence of alcohol
and drug use among battered women. In one review of the literature, 7 to
21% of female battering victims reported alcohol abuse or alcoholism.28 Fur-
thermore, recent literature compared with older literature suggests that the
proportion of battered women engaging in the use of alcohol and other drugs
may be increasing.8,9,39–43 In support of this hypothesis, a recent survey of
North Carolina’s domestic violence programs found that 85% of domestic
violence shelter providers believed that the number of battered women with
substance abuse problems is growing, and 97.5% believed that substance-
abusing battered women are an underserved population in their counties.44
One explanation for the increased risk of substance abuse among domes-
tic violence victims is that battered women may suffer from posttraumatic
stress disorder (PTSD). Women who suffer from PTSD and battered women
in general may turn to alcohol and drugs to cope with the physical pain,
emotional pain, and fears associated with being in a battering relation-
ship.37,38,45 Consequently, both substance abuse and traumatic stress disor-
ders must be addressed in order to adequately serve this population.
Regardless of etiology, substance abuse by battered women is associated
with and exacerbates a range of problems.4 Substance abuse is likely to make
it even more difficult for battered women to manage the complicated and
dangerous process of leaving violent relationships and maintaining violence-
free lives.35 These factors, in turn, are likely to increase the morbidity and
injury associated with battering,41 contribute to increased use of health care
services,46 enhance the risk of partner homicide,22 and increase the chances of
attempted and completed suicide.47 At the same time, the chronic nature of
battering—coupled with battered women’s fear, shame, loss of self-esteem
and personal power, isolation, lack of social support, and development of
other health problems (e.g., chronic pain, depression)—may make it harder
for them to break free of drug or alcohol dependency.9,40,48–50 For some
women, chronic violence and substance abuse become enmeshed. Substance
abuse among battered women also is thought to increase the likelihood that
these women will be held inappropriately responsible for their victimization
by those in a position to help them, such as the police, judges, social service
workers, and health care workers.44,51–53
Among men who are batterers and are also substance abusers, there are
good reasons to believe that their substance abuse is an impediment to getting
them to stop using violence against their partners. In the next section, we
discuss the ways that drinking might be a factor in the etiology of men’s
violence (e.g., cognitive impairment, drinking associated with male power
needs, alcohol as an excuse for violence). Although there is not a clear con-
sensus about the nature and magnitude of alcohol’s relationship to battering,
there is little disagreement that it is contributory in some way. Moreover,
there appears to be a growing disagreement that it is contributory in some
way, Moreover, there appears to be a growing consensus that it is necessary
to address alcohol and drug problems to achieve the best results with respect
to the battering problem.54,55 However, batterers’ programs may not wish to
deal with the substance abuse problems of their clients within their own
programs. Nevertheless, many such programs recognize the need for sub-
stance abuse treatment to maximize long-term violence-free outcomes for
violent men. It also seems likely that family conflict may be a risk factor for
relapse to substance abuse among those batterers in treatment.27 Substance
abuse, battering, and family conflict are probably related to each other in
complex ways.
3. Explanation of the Alcohol-Domestic Violence Relationship

Despite the high correlation of drinking and drinking problems to do-
mestic violence, there is considerable controversy about alcohol’s role in con-
nection with the violence. Psychosocial interpretations of the alcohol–domes-
tic violence relationship include the following:
1. Alcohol-induced cognitive impairment may result in misinterpretation
of spouses’ verbal–behavioral cues and intentions and increase the
likelihood of violent interactions56,57; a similar interpretation suggests
that alcohol causes a “myopia” that can have dangerous behavioral

effects.58
2. Alcohol interacts with the male’s need to exert personal power and
control in marital relationships55,59–61; the relevance of power and
control is also thought to be a major factor in family violence indepen-
dent of alcohol.62–64
3. Socially learned alcohol expectancy effects lead individuals to think
that alcohol induces violence and that it may be an acceptable excuse
for family violence24,65–67; an expectancy effect for alcohol-induced
aggression has been demonstrated in controlled laboratory experi-
ments.68–70 These expectancy effects have sociocultural roots.71
4. Using a complex explanatory framework that combines multiple caus-
al views, Leonard,72 in a National Institute on Alcohol Abuse and
Alcoholism (NIAAA) monograph, suggested that multiple distal fac-
tors (background influences, drinking patterns, marital discord) com-
bine with proximal influences, such as situational cues and acute alco-
hol effects, in the context of aversive interpersonal interactions to
produce physical aggression between spouses (p. 256).
A common feature of these psychosocial interpretations appears to be

that alcohol use in and of itself is not sufficient for a man to commit acts of
domestic violence. In particular, interpretations that focus on issues of male
power and control, or on socially learned expectancies related to alcohol,
suggest that underlying social or attitudinal problems must be dealt with if a
male batterer is to stop his behavior.73 Under these theoretical interpreta-
tions, should a batterer stop drinking but nothing be done about these other
issues, the battering is likely to continue, or the batterer may resume drinking
to provide an excuse to resume the pattern of abuse (i.e., the socially learned
expectancy explanation). Not surprisingly, then, programs for batterers con-
sider the violence to be a behavioral choice on the part of the batterer.55,74
Further, both the Minnesota Coalition for Battered Women38 and Levy and
Brekke74 recommended that programs for batterers place a strong emphasis
on the batterers assuming responsibility for their behaviors and that the pro-
grams challenge excuses for battering (e.g., “I was drunk”).
In contrast to programs for batterers that focus on psychosocial issues
and behavioral responsibility, the “disease concept” of alcoholism and the
Minnesota (or Hazelden) model of treatment are popular with many alcohol
treatment programs.75–77 Specifically, Roman and Blum77 stated that alcohol-
ism is characterized as a progressive disease (p. 758), marked by the inability
to control one’s drinking, and in later stages by physical tolerance and with-
drawal symptoms.78,79 As summarized by Chiauzzi and Liljegren75 (p. 305),
the Minnesota model of treatment is characterized by (1) psychological coun-
seling and educational approaches to build awareness of the consequences of
addiction; (2) involvement of recovering personnel in therapeutic roles; (3)
acceptance of the disease concept of alcoholism; (4) emphasis on attending
Alcoholics Anonymous (AA) meetings and doing what the Institute of Medi-
cine76 referred to as “stepwork” (p. 56); and (6) use of group counseling to
confront denial. An important underlying feature of this model is that alco-
holism is viewed as a primary problem requiring treatment in its own right.76
By definition, alcohol treatment programs will focus primarily on a cli-
ent’s use of alcohol, the consequences of this use, and development of a way
of living that involves abstinence from alcohol. However, the primary focus
on recovery from alcohol abuse or dependence can mean that perpetrators of
domestic violence who are in alcohol treatment may go undetected, allowing
the domestic violence to continue. If issues of domestic violence are detected
in the course of alcohol treatment, the emphasis on the disease concept of
alcoholism can lead domestic violence to be viewed as a problem secondary to
the abuse of alcohol, as opposed to a problem requiring special intervention
in its own right.80 Consequently, programs, counselors, and clients adhering
to an alcoholism treatment approach that emphasizes the disease concept of
alcoholism may assume that behavior change (i.e., cessation of battering) will
occur once a client stops drinking and begins to achieve some measure of
sobriety.80,81
A different set of issues is raised if the female victim of battering has a
substance abuse problem. Good evidence on the relationship between the
female domestic violence victim and substance abuse is sparse, but it seems
most likely that the woman’s substance abuse problem is usually either inde-
pendent of the battering problem or is a response to battering. Regardless of
the etiology of battering and substance abuse for the battering victim, it is
advisable to treat both problems. If a woman has an active alcohol or drug
problem, she will be less likely to have the cognitive, emotional, and other
resources needed to take constructive actions to protect herself and her chil-
dren, leave the violent relationship, or do both.
4. Examples of Current Linkage

Bennett and Lawson54 surveyed substance abuse programs and domestic
violence programs in Illinois to learn more about the types of service linkages
between both types of programs, as well as barriers to cooperation. Their
sample consisted of all 45 domestic violence programs in Illinois and a ran-
dom sample of 150 licensed substance abuse treatment programs in the state.
A total of 388 staff from 74 programs participated in the study (249 staff from
53 substance abuse treatment programs and 139 staff from 21 domestic vio-
lence programs), for agency response rates of 47% for the domestic violence
programs and 35% for the substance abuse treatment programs.
Although response rates were low and the focus of the study was on
service linkages within a single state, the study represents an attempt to use
survey research methods to document the extent and types of service linkages
between substance abuse treatment and domestic violence programs, as well
as to indicate the relative importance of possible barriers to linkage. Key
findings from their study regarding the extent and types of service linkages
include the following:
1. Formal screening for cross-problems (e.g., screening for domestic vio-
lence among substance abuse treatment clients) was relatively rare,
done possibly by about 10% of programs. However, “screening” tend-
ed to consist of one or two questions and tended not to be very
systematic. Among substance abuse treatment clients, for example,
histories of domestic violence tended to be identified only if a client
reported incidents in the course of counseling.
2. One exception to the general pattern of haphazard screening was that
screening for substance abuse tended to be more systematic and struc-
tured in programs for male batterers. Many of these programs used
the Michigan Alcoholism Screening Test (MAST).82
3. Although sizable percentages of substance abuse treatment and do-
mestic violence programs indicated that they had an in-house “spe-
cialist” on staff (35% of the treatment programs and 24% of the domes-
tic violence programs), such expertise was not often based on formal
training. For example, substance abuse specialists in domestic vio-
lence programs often consisted of staff who were recovering from
their own chemical dependency.
4. Approximately 70% of substance abuse treatment and domestic pro-
grams indicated some form of formal linkage agreement with the com-
plementary program. However, only about 20% of directors of sub-
stance abuse treatment programs indicated that they met sometimes
or frequently with staff from domestic violence programs, compared
with nearly 70% of domestic violence program directors who indicated
that they met with staff from substance abuse treatment programs.
5. Nearly one in four (23%) of substance abuse treatment staff indicated
that they never referred clients to domestic violence programs, com-
pared with only 5% of domestic violence staff who did not refer clients
to substance abuse treatment.
We have also had conversations about service linkage with approx-
imately 25 individuals who do research or deliver services in the domestic
violence and substance abuse fields. These conversations confirm that linkage
to address both domestic violence and substance abuse is infrequent. And
although our conversations cannot be said to have been with a representative
sample of researchers and service providers, two observations can be made:
Programmatic linkage attempts are more common for batterer programs than
for victim programs, and programs that do try to deal with the coexistent
domestic violence and substance abuse problems typically do so by referral.
In other words, the most common pattern we have seen is for batterers’
programs to refer their clients who have alcohol or drug problems to sub-
stance abuse treatment. This is probably not surprising given the impedi-
ments to linkage we have already discussed (e.g., differing treatment philoso-
phies, organizational boundaries).
We have identified a few batterer treatment programs operating at sev-

eral locations that incorporate both substance abuse and domestic violence
treatment as part of a single programmatic entity.55,80,83 The Amend Pro-
gram, for example, operating at several locations in Colorado, has counsel-
ors trained as both substance abuse and domestic violence counselors.80
The Amend Program attempts to deal with both substance abuse and vio-
lent behavior problems among batterers' problems in spite of the diffi-
culties of integrating the two treatments already discussed. It is not clear
whether it is feasible to adopt this approach widely, or whether it requires
special circumstances, such as a strong commitment, to deal explicitly with
both problems.
5. Linking Services
Given the strong evidence of the relationship between alcohol abuse
and domestic violence, a number of researchers and service providers have
suggested the need for greater integration of treatment services for alcohol
and domestic violence problems, such that both problems can be ad-
dressed.38,54,74,80,84 The situation with substance abuse and domestic violence
may be analogous to other dually occurring problems, such as psychiatric
comorbidities. Although the problems are interrelated, the systems that have
been designed to intervene in the problems operate essentially independent
of one another. Stated more generally, the dilemma is that although research
evidence continues to mount, showing that many psychosocial problems are
highly interrelated (i.e., people with one problem tend to have many other
problems as well), our systems of care tend to be narrowly focused on a
specific problem, and the systems operate independently.
The set of problems existing around the alcohol–domestic violence rela-
tionship may be a particularly complex situation for integrating services. The
configuration of problems includes the following:
1. The needs of the victim and her children related to injury, housing,
subsistence, safety, and so on, and whether to continue the existing
family unit.
2. Possible substance abuse treatment needs of the victim.
3. Intervention to deal with the violent behavior of the offender.
4. Possible substance abuse treatment needs of the offender.
It is unlikely that the full set of needs for families with the dual problems of
violence and substance abuse can be dealt with by a single program.
Attempting to link services provided by multiple programs raises several
issues that may make linkage problematic:
1. The philosophies and goals of treatment for programs may not be
compatible.
2. Mechanisms and logistics for linking services may not exist.
3. Funding and other resources to support linkage may be limited.
Domestic violence programs usually have a strong advocacy component

that focuses on protecting and promoting the safety and interests of women
victims. A corollary of this perspective emphasizes personal accountability for
men who batter. As discussed previously, substance abuse programs often
operate under the view that addiction is a disease, implicitly deemphasizing
personal accountability for actions that took place during the active disease
state. Moreover, most domestic violence programs view the violence that is
perpetrated against women to have its roots in men’s desire to exercise power
and control over their partners. Substance abuse programs often view addic-
tion as a pervasive condition that has broad negative physiological, psycho-
logical, and behavioral effects that can include violence. The logical implica-
tions of the views of domestic violence and substance abuse treatment
programs for dealing with the problems they try to ameliorate are at odds
with each other. This is an impediment to linking substance abuse and do-
mestic violence services. Indeed, Bennett and Lawson54 found that having
conflicting beliefs about the issue of personal responsibility for behavior was
the leading factor cited as a possible reason for noncooperation between
substance abuse treatment and domestic violence programs; this factor was
endorsed by 55% of all respondents as a possible reason for noncooperation,
as well as by 65% of respondents in domestic violence programs and 45% of
staff in substance abuse treatment programs.
The problem of conflicting perspectives that underlie programmatic in-
terventions is further complicated for female domestic violence victims with
alcohol or drug problems who may find themselves in substance abuse treat-
ment programs, operating under the disease concept. In such programs,
there may be an implicit tendency to see the domestic violence they suffer as
associated with their own alcohol or drug problem (i.e., that their substance
abuse contributed to their victimization). Under this view, the substance
abuse problem receives primary attention. Sobriety first is likely to be the goal
of substance abuse programs; safety first will usually be the first priority for
domestic violence programs.38,73 Fundamental differences of perspective
make it difficult for collegial cooperation between domestic violence and sub-
stance abuse treatment programs.
The problem of different treatment paradigms also impedes dealing ef-
fectively with the multiple problems of substance abuse and domestic vio-
lence, even within single program types. Because domestic violence pro-
grams view some of the traditional substance abuse treatment approaches
with skepticism or hostility, and because they may not have an alternative
treatment paradigm that is effective, successful responses to alcohol and drug
problems for their clients may be absent. Substance abuse programs may
ignore the violence that their clients engage in, assuming that if the substance
abuse treatment is successful, the violence problem will be solved. There is
evidence that battering does not cease unless the problem is explicitly ad-
dressed.74,80 And substance abuse treatment programs do not typically focus
on the violence problem, particularly for the male batterer.
Resource scarcity or the absence of appropriate resources is a common
problem in domestic violence and substance abuse treatment programs. Do-

mestic violence programs may not have the expertise to recognize the pres-
ence of an alcohol or drug problem. Even if a problem is recognized, the
resources or programmatic expertise to deal with it may not be available. In
fact, some domestic violence programs will not accept women with active
substance abuse problems because it may interfere with their care of women
and children in their programs who are not substance abusers. Similarly,
substance abuse treatment programs do not usually have programmatic com-
ponents that address either battering problems or problems associated with
being a victim of domestic violence.
The linkage of alcohol and domestic violence services is also impeded by
organizational boundaries and financial limitations. Domestic violence and
alcohol treatment programs are usually independent of one another. They
are organized separately, have intervention philosophies and program ap-
proaches shaped to their unique perspectives and goals, and may have re-
source limitations that prevent integration of services. Financial support for
both alcohol and domestic violence programs is usually modest and limits the
scope of programming. Financial impediments to integration of services may
also be formal. For example, during our interviews at both types of programs,
we were told of an alcohol treatment program that was prevented from dealing
with domestic violence because hours could not be billed for these services.
The organizational placement and method of funding domestic violence
and substance abuse treatment services make service linkage difficult. Domestic
violence and substance abuse treatment programs tend to be located in different
parts of the social service and health bureaucracies. The domestic violence and
substance abuse treatment functions are often independent entities in state,
county, or municipal governments. Or, the two program types are located in
different parts of the governmental structure (e .g., health, mental health). These
formal boundaries make it more difficult to embark on joint ventures, such as
attempting to deal with the related aspects of domestic violence and substance
abuse problems. The collaboration problem is difficult enough when the two
kinds of services are organized separately but in the same departmental unit,
such as when domestic violence and substance abuse programs come under a
Department of Mental Health. When different departments are involved, a
consensus to do something collaborative is even more complicated.
In recent years, the courts have begun to mandate that domestic violence
offenders participate in batterers’ programs as a condition of their sentences.
This approach, however appropriate for protecting women victims, may in-
troduce another layer of complexity for linkage, given the involvement of the
criminal justice system.
6. Models for Service Linkage

Recognition that many individual and family problems are multifaceted
has led to a movement toward services integration, which has been defined
by Agranoff85 as “the quest for the development of systems that are respon-
sive to the multiple needs of persons at risk” (p. 533). Born largely out of the
War on Poverty that was launched by the Johnson Administration in the
mid-1960s, efforts at improving services integration have flourished, but they
also have encountered substantial barriers.86
Although many services integration efforts have been broad-based, it is
clear that there are potential benefits to integration on a more limited scale.
For example, experiences of programs attempting to establish integrated
substance abuse treatment and primary care have demonstrated that it
is feasible to improve the integration of these services and beneficial to do
so.87 On a somewhat broader scale, evaluation of the National Institute of
Mental Health’s Child and Adolescent Service System Program demonstra-
tion showed that integration of services for severely emotionally disturbed
children and adolescents could be improved at the state and local levels, and
that doing so could result in more effective service delivery.88
Examples of attempts to integrate alcohol abuse and domestic violence
interventions include programs in the Amend Program at several Colorado
locations,80 the Intercede Program of Longford Health Sources in Massillon,
Ohio,83 and Pittsburgh Veterans Affairs Medical Center.55 Beyond these ex-
amples, however, little information has been documented about the extent of
integration of alcohol-related and domestic violence services and mechanisms
for integrating both services when integration is attempted. An exception is a
study of domestic violence and chemical dependency program linkages in
Illinois. Bennett and Lawson54 surveyed statewide samples of these programs
and found that programs thought linkage would benefit clients; however, the
authors noted that “in-house expertise in the cross-problem was minimal for
both chemical dependency and domestic violence agencies’’ (p. 281).
A recent study of linking drug abuse treatment to primary medical by
Schlenger et al.87,89 provides some guidance for the linkage of domestic vio-
lence and substance abuse programming. The study developed an a priori
taxonomy of basic linkage approaches that were being implemented. This
taxonomy was based on the perspective of the service user, and it identified
four different models: (1) centralized, where drug treatment and primary care
services are offered at a single location (“one-stop shopping”); (2) decentral-
ized, where different services are offered at different locations, with clients
being referred to different locations, depending on their service needs; (3)
mixed, where a limited number of primary care and drug treatment services
are offered at a single location, but most services are delivered at separate
locations; and (4) transitional, in which the location of services changes over
the user’s treatment history.
It was observed, however, that the a priori conceptualization of models
of linkage was probably more complicated than what existed in reality. Based
on observations of linkage demonstration projects, it was hypothesized that
specific linkage efforts may be best understood as lying along a continuum
that ranges from decentralized through centralized. At one extreme, virtually
all drug treatment and primary care services are delivered at separate loca-
tions, and linkage occurs by referral; in these highly decentralized systems,

case management appears to be a key ingredient for ensuring that referrals
are kept and for providing some continuity of care. At the other extreme of
the linkage continuum, a full range of both primary care and drug treatment
services is located in a single location (i.e., the one-stop shopping approach),
with linkage occurring through the physical proximity of services.90 It would
not be surprising to find a similar decentralized to centralized continuum of
domestic violence and substance abuse treatment integration among the link-
ages observed in this area although too few examples exist to be sure.
In addition, findings from the national linkage evaluation suggested
some important benefits of having services in a single location. In particular,
having services in a single location appeared to make services more accessible
to clients and facilitated client enrollment and retention in treatment. In addi-
tion, having services in a single location almost always increased both formal
and informal interaction among providers, contributing to a more holistic
approach to treatment.
Furthermore, the linkage demonstration study revealed that attempts to
bring together two service systems that have historically been separate can
often expose philosophical differences in approaches to treatment that need
to be addressed and overcome if linkage is to occur. Specifically, it was
learned through contacts with some grantees that primary care providers
sometimes viewed the structure and rules of some drug treatment programs
as being overly rigid, while drug treatment providers sometimes viewed pri-
mary care providers as “rescuing” or “enabling” a substance abuser. Sim-
ilarly, then, we might expect philosophical differences between alcohol treat-
ment and domestic violence providers to be an important reason why these
services remain separate, as well as a potential barrier to integration when
programs attempt to establish linkages.
It may also be difficult to sustain the linkage of services across different
agencies over time. A research demonstration project to link substance abuse,
health, employment, and housing services for homeless drug-using adults
illustrates the problem. In a recent discussion on service linkages, Erickson et
al.90 noted that the Amity Settlement Services for Education and Transition
Program found that the networking needed to sustain service linkage “wilted
away” over time (p. 343).
Some domestic violence service providers and advocates downplay the
role of alcohol in incidents of domestic violence. Many domestic violence
service providers consider drinking as an excuse for acting violently or think it
is offered as an excuse after the fact to avoid accountability. Moreover, a
commonly held view is that the principal causes of domestic violence are
sociocultural factors that support the use of violence against women and
reinforce the social and economic advantages of men. From this perspective, a
focus on alcohol as an important contributory factor in domestic violence may
detract from a focus on what are viewed as the more important cultural,
social, and economic causes. For example, Cayouette91 cited the concern that
men in a batterers’ program who subsequently enter a recovery program for

alcoholism can have their focus diverted from attention to their battering
problem. Cayouette also expressed the concern that some men can use their
involvement in a recovery program (e.g., frequent attendance at AA meet-
ings) as a way of continuing to avoid family responsibilities. Furthermore,
application of terms from the self-help movement, such as “codependent” or
“enabler,” to female victims of domestic violence has been cited as a new form
of victim-blaming, in which the woman is made to feel partly responsible for
the alcoholic batterer’s problem.80,91 It is the view of some domestic violence
advocates and service providers that empowerment for women victims and
behavioral accountability and attitudinal change for batterers are the major
goals to be sought.
Based on anecdotal evidence regarding the substance abuse–domestic
violence service linkage and on general findings about linking multiple kinds
of services, some tentative inferences can be drawn. Our conversations with
individuals in both the domestic violence and substance abuse treatment
fields suggest that linkage is rare or at least infrequent. (The Bennett and
Lawson54 study in Illinois is consistent with this conclusion.) This is true for
each program type and for both domestic violence victims and offenders.
Domestic violence programs for victims often will not accept women with
substance abuse problems. Programs for male batterers usually do not deal
with alcohol and drug problems; if they recognize that a substance abuse
problem exists, they will usually refer the individual to a substance abuse
treatment program. Substance abuse treatment programs do not usually rec-
ognize or deal with the violent experiences or violent behavior of their clients.
One implication of the scarcity of existing linkages is that well-developed
models for accomplishing linkage do not exist. This means that significant
programmatic development and evaluation are needed. It would be prema-
ture to recommend implementation of widespread attempts at linking sub-
stance abuse and domestic violence services in the absence of well-developed
and effective models for doing so.
Based on anecdotal evidence from domestic violence treatment pro-
viders, linkage is sometimes not possible until the victim’s or offender’s sub-
stance abuse problem is dealt with, particularly for male batterers. The sub-
stance abuse problem can be a serious impediment to meaningful cognitive
and psychological engagement in programming to address problems related
to domestic violence. The anecdotal evidence we have gathered indicates that
domestic violence treatment providers usually refer victims and batterers who
have serious alcohol or drug problems to a substance abuse program to deal
with this problem. In our experience, it is rare that victim or offender domes-
tic violence programs deal simultaneously with the domestic violence pro-
gramming.
The domestic violence program services provided to victims and of-
fenders differ. Programming for victims emphasizes safety, shelter, child care
services, legal advocacy, and so on. Programming for male batterers empha-
sizes anger and conflict management, acceptance of responsibility for violent

behavior, meeting family responsibilities, and so on. Due to the different
program foci, the optimal approaches to linking substance abuse and domes-
tic violence services may differ for victims and offenders. For example, the
simultaneous treatment of both substance abuse and problems related to
domestic violence victimization may work well, whereas a serial approach in
which the substance abuse problem is dealt with first for batterers may be
better in the case of offenders.
At the current state of development of domestic violence and substance
abuse treatment linkage, the linkage mechanism that seems most appropriate
is a brokering or case management approach. A case manager approach to
delivery of therapeutic services involves assessing individual's needs and
arranging for services to address those needs. So, in the case of linking do-
mestic violence and substance abuse services, for example, a case manager
would evaluate whether a victim or offender has a substance abuse problem,
assess its severity, identify available treatment resources, make a referral to
treatment, and monitor treatment progress. Because so little is currently
known about dealing with the domestic violence and substance abuse prob-
lems within a single programmatic context, the case management model
seems appropriate. When more is learned about how to integrate program-
ming for the two kinds of problems, dealing with both in a single program
might be effective. What does seem clear though is that domestic violence and
substance abuse are highly correlated. Regardless of the exact nature of their
causal relationship (or the absence thereof), dealing with both problems may
be beneficial. It remains to be demonstrated whether synergistic benefits
result from linking domestic violence and substance abuse services, but there
is every reason to expect that dealing with both problems when they coexist
will pay high individual and social dividends.
7. Next Steps
Based on the strong empirical association between alcohol and domestic
violence and on evidence from practitioners, it is reasonable to infer that
services for the two kinds of problems should be linked. But linkage is both
problematic and uncertain. There are philosophical, structural, and practical
impediments to linkage, as we have discussed. Moreover, no existing scien-
tific evaluation evidence indicates that linkage is successful. Two needs are
certain: (1) study of the feasibility of linkage and identification of promising
approaches to linkage, and (2) evaluation of the effects of linkage.
One focus of feasibility work should be the study of attitudes and beliefs
toward the alcohol–domestic violence relationship among service providers
and the implications of these beliefs for service linkage.54 If the attitudes and
beliefs among a range of service providers in both the domestic violence and
chemical dependency fields were documented, it likely would be possible to
16 • Alcohol-Domestic Violence Services Linkage 401
develop plans for linkage that have a high probability of succeeding. Another
line of linkage inquiry should focus on structural factors likely to affect efforts
to link services. The bureaucratic placement and organization of the two
kinds of services, as well as the method and level of financing of the services,
are likely to influence both the feasibility and implementation of linkage.
Ultimately, evaluation of the domestic violence-chemical dependency
service linkage will be the arbiter of its long-term use. At the present time,
this presents a real challenge, given the infrequency of linkage in the real
world. The first set of steps should probably involve conducting process and
implementation evaluation work with the few existing examples of program-
matic service linkage. Subsequently, outcome evaluation work using experi-
mental or quasi-experimental designs could address questions of effective-
ness.
The evaluation of linkage will be a complex undertaking, given the need
to consider three major outcomes (success of victim services, batterer violence
cessation, and chemical dependency recovery) and the involvement of multi-
ple program types. Even when the technical evaluation methodology chal-
lenges are dealt with successfully, the findings are likely to be ambiguous. For
example, some aspects of linkage such as batterer violence cessation may
succeed, but others such as chemical dependency recovery may fail.
Given the high coincidence and seriousness of the domestic violence and
chemical dependency problems, it is reasonable to begin linkage activities
before evaluation results are in. Bennett and Lawson54 argued that chemical
dependency and domestic violence programs ought at least to assess for the
cross-problems, making an argument that not to do so is at best programmat-
ically unsatisfactory and at worst irresponsible. Given the very high personal
and societal costs involved, we agree with their conclusion.
It is not premature to document the existence of chemical dependency
problems among domestic violence victims and offenders and to attempt to
determine whether violent victimization or violent behavior co-occur with a
chemical dependency problem. Uncovering the problems can be the first step
toward amelioration, and establishing the magnitude of the coexistent prob-
lems among clients may itself create an impetus toward service linkage.
ACKNOWLEDGMENT. The authors appreciate the editorial support of Richard

S. Straw and the word-processing support of Catherine A. Boykin during the
preparation of this chapter.
References
1. Straus MA, Gelles RJ, Steinmetz SK: Behind Closed Doors: Violence in the American Family. New
York, Anchor Doubleday, 1980.
2. US Department of Health and Human Services: Healthy People 2000: National Health Promotion
and Diseuse Prevention Objectives (DHHS publication PHS 91-50212). Washington, DC, US
Government Printing Office, 1991.
3. Louis Harris and Associates Inc: The Commonwealth Fund Survey of Women’s Health. New York,
The Commonwealth Fund, 1993.
4. National Committee for Injury Prevention and Control: Inju ry Prevention: Meeting the Chal-
lenge. New York, Oxford Press, 1989.
5. Kurz D, Stark E: Not-so-benign neglect: The medical response to battering, in Yllo K,
Bograd M (eds): Feminist Perspectives in Wife Abuse. Newbury Park, CA, Sage, 1988, pp 249–
266.
6. McLeer S, Anwar R: Education is not enough: A systems failure in protecting battered
women. Ann Emerg Med 18:651–653, 1989.
7. McLeer S, Anwar R: A study of battered women presenting in an emergency room. Am J
Public Health 79:65–66, 1989.
8. Stark E, Flitcraft A, Frazier: Medicine and patriarchal violence: The social construction of a
private event. Int J Health Serv 9:461–493, 1979.
9. Stark E, Flitcraft A: Spouse abuse, in Rosenberg ML, Fenley MA (eds): Violence in America: A
Public Health Approach. New York, Oxford University Press, 1991, pp 123–157.
10. Harlow CW: Female Victims of Violent Crime (NCJ-126826). Washington, DC, US Department
of Justice, Office of Justice Programs, Bureau of Justice Statistics, 1991.
11. Domino JV, Haber JD: Prior physical and sexual abuse in women with chronic headache:
Clinical correlates. Headache 26:310–314, 1987.
12. Haber JE, Roos C: Effects of spouse abuse and/or sexual abuse in the development and
maintenance of chronic pain in women. Adv Pain Res Ther 9:889–895, 1985.
13. Larsson G, Anderson M: Violence in the family: Morbidity and mortality consumption. Scand
J Soc Med: 161–166, 1989.
14. Helton A: Protocol of Care for the Battered Women: Prevention of Battering during Pregnancy
(prepared June 1986 at Texas Women’s University through grant from Metropolitan Houston
Chapter of March of Dimes Birth Defects Foundation). White Plains, NY, March of Dimes
Birth Defect Foundation, 1987.
15. McFarlane J: Battering during pregnancy: Tip of an iceberg revealed. Women Health 15:69–84,
1989.
16. Drossman DA, Leserman J, Nachman G: Sexual and physical abuse in women with function-
al or organic gastrointestinal disorders. Ann Intern Med 113:828–833, 1990.
17. Koss MP: The women’s mental health research agenda: Violence against women. Am Psychol
45:374–380, 1990.
18. Post R, Willett AB, Franks RD, et al: A preliminary report on the prevalence of domestic
violence among psychiatric inpatients. Am J Psychiatry 137:974–975, 1987.
19. Rosewater LB: Battered or schizophrenic? Psychological tests can’t tell, in Yllo K, Bograd M
(eds): Feminist Perspectives on Wife Abuse. Newbury Park, CA, Sage, 1988, pp 200–216.
20. Walker LE: The Battered Women Syndrome (Springer Series: Focus on Women, vol 6). New
York, Springer, 1984.
21. Amaro H, Fried LE, Cabral H, Zuckerman B: Violence during pregnancy and substance use.
Am J Public Health 80:575–579, 1990.
22. Browne A: When Battered Women Kill. New York, Free Press, 1987.
23. Mercy IA, Saltzman IE: Violence among spouses in the United States, 1976-85. Am J Public
Health 79:595–599, 1989.
24. Coleman DH, Straus MA: Alcohol abuse and family violence, in Gottheil E, Druley KA,
Skoloda TE, Waxman HM (eds): Alcohol, Drug Abuse, and Aggression. Springfield, IL, Charles
C. Thomas, 1983, pp 104–124.
25. Kantor GK, Straus MA: The “drunken bum” theory of wife beating. Soc Probl 34:213–231,
1987.
26. Kantor GK, Straus MA: Substance abuse as a precipitant of wife abuse victimizations. Am J
27. Gondolf EW, Foster RA: Wife assault among VA alcohol rehabilitation patients. Hosp Commu-
nity Psychiatry 42:74–79, 1991.
28. Hamilton CJ, Collins JJ: The role of alcohol in wife beating and child abuse, in Collins JJ Jr
nal Behavior. New York, Guilford Press, 1981, pp 253–287.
current state of knowledge. Violence Vict 1:101–124, 1986.
30. Leonard KE, Bromet EJ, Parkinson DK, et al: Patterns of alcohol use and physically aggressive
behavior in men. J Stud Alcohol 46:279–282, 1985.
31. National Woman Abuse Prevention Project: Alcohol Abuse and Domestic Violence (Domestic
Violence Fact Sheet). Springfield, IL, Illinois Department of Public Aid, n.d.
32. Rounsaville BJ: Theories in marital violence: Evidence from a study of battered women.
Victimology 3:11–31, 1978.
33. Gayford JJ: Wife battering: A preliminary survey of 100 cases. Br Med J 1:194–197, 1975.
34. Carlson BE: Battered women and their assailants. Soc Work 22:455–460, 1977.
35. Roy M: A current survey of 150 cases, in Roy M (ed): Battered Women: A Psychosociological
Study of Domestic Violence. New York, Van Nostrand Reinhold, 1977, pp 24–44
36. Lehmann N, Krupp S: Alcohol-related domestic violence: Clinical implications and interven-
tion strategies. Alcohol Treat Q 1(4):111–115, 1984.
37. Jones A, Schechter S: When Love Goes Wrong: What to Do When You Can’t Do Anything Right.
New York, HarperCollins, 1992.
38. Minnesota Coalition for Battered Women: Safety First: Battered Women Surviving Violence When
Alcohol and Drugs Are Involved. Saint Paul, MN, Minnesota Coalition for Battered Women,
1992.
39. Barnett OW, Fagan RW: Alcohol use in male spouse abusers and their female partners. J Fam
Violence 8(1):1–25, 1993.
40. Berenson AB, Stiglich NJ, Wilkinson GS, Anderson GD: Drug abuse and other risk factors for
physical abuse in pregnancy among white non-Hispanic, black and Hispanic women. Am J
Obstet Gynecol 164:1491–1499, 1991.
41. Erble PA: Alcohol abusers and non-users: A discriminant analysis of differences between two
subgroups of batterers. J Health Soc Behav 23:260–271, 1982.
42. Frieze I, Knoble J: The effects of alcohol on marital violence. Paper presented at annual
meeting of American Psychological Association, Montreal, Quebec, Canada, 1980.
43. Van Hasselt VB, Morrison RL, Bellack AS: Alcohol use in wife abusers and their spouses.
Addict Behav 10:127–135, 1985.
44. Smith PH, Chescheir N: Services Integration for Substance-Abusing Battered Women: Final Report.
Raleigh, NC, North Carolina’s Governor’s Institute for Alcohol and Substance Abuse,
1994.
45. Center for Substance Abuse Treatment: Practical Approaches in the Treatment of Women Who
Abuse Alcohol and Other Drugs (DHHS publication SMA 94-3006). Rockville, MD, US Depart-
ment of Health and Human Services, Public Health Service, Substance Abuse and Mental
Health Services Administration, 1994.
46. Hankin J, Ortay IS: Mental Disorder and Primary Medical Care: An Analytical Review of the
Literature (publication ADM 78-661, series D, no 5). Washington, DC, US Department of
Health and Human Services, 1979.
47. Gomberg ESL: Suicide risk among women with alcohol problems. Am J Public Health 79:1363–
1366, 1989.
48. Haver B: Female alcoholics: IV. The relationship between family violence and outcome 3–10
years after treatment. Acta Psychiatr Scand 77:449–455, 1987.
49. Korlath MJ: Alcoholism in battered women: A report of advocacy services to clients in a
detoxification facility. Victimology Int J 292–299, 1979.
50. Smith PH, Gittelman DF: Psychological consequences of battering: Implications for women’s
health and medical practice. NC Med J 55(9):2–7, 1994.
51. Aramburu B, Leigh BC: For better or worse: Attributions about drunken aggression toward
male and female victims. Violence Vict 6:31–41, 1991.
52. Corenblum B: Reactions to alcohol-related marital violence: Effects of one’s own abuse expe-
rience and alcohol problems on causal attributions. J Stud Alcohol 11:665–674, 1983.
53. Richardson DC, Campbell JL: Alcohol and wife abuse: The effect of alcohol on attributions of
blame for wife abuse. Pers Soc Psychol Bull 6:51–56, 1980.
54. Bennett L, Lawson M: Barriers to cooperation between domestic-violence and substance-
abuse programs. Fam Soc J Contemp Human Services 75:277–286, 1994.
55. Gondolf EW: Alcohol abuse, wife assault, and power needs. Manuscript submitted for pub-
lication, Indiana, PA, Mid-Atlantic Abduction Training Institute, Indiana University of Penn-
sylvania, 1993.
Alcoholism, vol 4: Social Aspects of Alcoholism. New York, Plenum Press, 1976, pp 351–444.
57. Pernanen K: Theoretical aspects of the relationship between alcohol use and crime, in
Collins J (ed): Drinking and Crime: Perspectives on the Relationship between Alcohol Consumption
and Criminal Behavior. New York, Guilford Press, 1981, pp 1–69.
58. Steele CM, Josephs RA: Alcohol myopia: Its prized and dangerous effects. Am Psychol 45:921–
933, 1990.
59. Frieze IH, Browne A: Violence in marriage, in Ohlin L, Tonry M (eds): Family Violence.
Chicago, University of Chicago press, 1989, pp 163–218.
60. McClelland D: Power: The Inner Experience. New York, Wiley, 1975.
61. McClelland DC, Davis WN: The influence of unrestrained power concerns on drinking in
working-class men, in McClelland DC, Davis WN, Kalin R, Wanner E (eds): The Drinking
Man. New York, Free Press, 1972, pp 142–161.
62. Bowker L: Beating Wife-Beating. Lexington, MA, Heath, 1983.
63. Dobash RE, Dobash RP: Violence Against Wives. New York, Free Press, 1979.
64. Dobash RE, Dobash RP: Women, Violence, and Social Change. New York, Routledge, 1992.
65. Collins JJ: Alcohol and crime aetiology, in Russell J (ed): Alcohol and Crime: Proceedings of a
Mental Health Foundation Conference. London, England, Mental Health Foundation, 1993, pp
22–32.
67. Spieker G: What is the linkage between alcohol abuse and violence? In Gottheil E, Druley
KA, Skoloda TE, Waxman HM (eds): Alcohol, Drug Abuse and Aggression. Springfield, IL,
Charles C. Thomas, 1983, pp 125–136.
drinkers. J Abnorm Psychol 84:508–518, 1975.
70. Taylor SP, Gammon CB: Effects of type and dose of alcohol on human physical aggression.
71. MacAndrew C, Edgerton RB: Drunken Comportment: A Social Explanation. Chicago, Aldine,
1969.
72. Leonard KE: Drinking patterns and intoxication in marital violence: Review, critique, and
future directions for research, in Martin SE (ed): Alcohol and Interpersonal Violence: Fostering
Multidisciplinary Perspectives (research monograph 24, NIH publication 93-3496). Rockville,
MD, National Institute on Alcohol Abuse and Alcoholism, 1993, pp 253–280.
73. New York State Office for the Prevention of Domestic Violence: Adult Domestic Violence: The
Alcohol Connection. Rensselaer, NY, New York State Office for the Prevention of Domestic
Violence, 1993.
74. Levy AJ, Brekke JS: Spouse battering and chemical dependency: Dynamics, treatment, and
service delivery, in Potter-Efron RT, Potter-Effron PS (eds): Aggression, Family Violence and
Chemical Dependency. New York, Haworth Press, 1990, pp 81–97:
75. Chiauzzi EJ, Liljegren S: Taboo topics in addiction treatment: An empirical review of clinical
folklore. J Subst Abuse Treat 10:303–316, 1993.
76. Institute of Medicine: Broadening the Base of Treatment for Alcohol Problems. Washington, DC,
National Academy Press, 1990.
77. Roman PM, Blum TC: The medicalized conception of alcohol-related problems: Some social
sources and some social consequences of murkiness and confusion, in Pittman DJ, White HR
(eds): Society, Culture, and Drinking Patterns Reexamined. New Brunswick, NJ, Rutgers Center
of Alcohol Studies, Publications Division. 1991, pp 753–774.
78. Jellinek EM: The Disease Concept of Alcoholism. New Haven, CT, College and University Press,
1960.
79. Jellinek EM: Phases of alcohol addiction,in Pittman DJ, White HR (eds): Society, Culture, and
Drinking Patterns Reexamined. New Brunswick, NJ, Rutgers Center of Alcohol Studies, Pub-
lications Division, 1991, pp 403–416.
80. Rogan A: Domestic violence and alcohol: Barriers to cooperation. Alcohol Health Res World
10:22–27, 1985/86.
81. Harner IC: The alcoholism treatment client and domestic violence. Alcohol Health Res World
12:150–152, 160, 1987/88.
82. Selzer M: The Michigan Alcoholism Screening Test: The quest for a new diagnostic instru-
ment. Am J Psychiatry 127:1653–1658, 1971.
83. Burkins M: Informational Packet on Individualized Care. Massillon, OH, Longford Health Sources
at Massillon Community Hospital, 1995.
84. Lindquist CU: Battered women as coalcoholics: Treatment implications and case study. Psy-
chotherapy 23:622–628, 1986.
85. Agranoff R: Human services integration: Past and present challenges in public administra-
tion. Public Admin Rev 51:533–542, 1991.
86. US Department of Health and Human Services: Services Integration: A Twenty-Year Retrospec-
tive. Washington, DC, Office of the Inspector General, 1991.
87. Schlenger WE, Kroutil LA, Roland EJ, et al: National Eualuation of Models for Linking Drug Abuse
Treatment and Primary Care: Linking Drug Treatment and Primary Care: A Casebook (report pre-
pared for Clinical Medicine Branch, Division of Clinical Research, National Institute on Drug
Abuse, contract 283-90-0001). Research Triangle Park, NC, Research Triangle Institute, 1992.
88. Schlenger WE, Etheridge RM, Hansen DJ, et al: Evaluation of state efforts to improve systems
of care for children and adolescents with severe emotional disturbances: The CASSP initial
cohort study. J Ment Health Admin 19:131–142, 1992.
89. Schlenger WE, Kroutil LA, Roland EJ, Dennis ML: National Evaluation of Models for Linking
Drug Abuse Treatment and Primary Care: Descriptive Report of Phase 1 Findings (report prepared
for Clinical Medicine Branch, Division of Clinical Research, National Institute on Drug Abuse
contract 283-90-0001). Research Triangle Park, NC, Research Triangle Institute, 1992.
90. Erickson JR, Chong J, Anderson C, Stevens S: Service linkages: Understanding what fosters
and what deters from service coordination for homeless adult drug users. Contemp Drug Probl
22:343–362,1995.
91. Cayouette S: The Addicted or Alcoholic Batterer. Boston, EMERGE, 1990.
Contents of Previous Volumes
Volume 1
I. The Role of Genetics in the Expression of Alcoholism Henri Begleiter, Section

Editor
Overview Donald Goodwin
Twin Adoption Studies: How Good Is the Evidence for a Genetic Role Robin M.
Murray, Christine A. Clifford, and Hugh M.D. Gurling
Pharmacogenetic Approaches to the Neuropharmacology of Ethanol Dennis R.
Peterson
II. The Behavioral Treatment of Alcoholism Edward Goffheil, Section Editor

Overview Edward Gottheil
How Environments and Persons Combine to Influence Problem Drinking: Current
Research Issues G.N. Barucht
Alcoholism: The Evolution of a Behavioral Perspective William H. George and
G. Alan Marlatt
Behavioral Treatment Methods for Alcoholism Glenn R. Caddy and Trudy Block
Outcome Studies on Techniques in Alcoholism Treatment Gloria K. Litman and
Anne Topham
Contributions to Behavioral Treatment from Studies on Programmed Access to
Alcohol Glenn R. Caddy and Edward Gottheil
Current Status of the Field: Contrasting Perspectives A. The Behavioral
Therapist's View Mark B. Sobell and Linda C. Sobell B. The Future of
Behavioral Interventions S. H. Lovibond C. A Medical Clinician's
Perspective Robert A. Moore D. An Anthropological Perspective on the
Behavior Modification Treatment of Alcoholism David Levinson
III. Social Mediators of Alcohol Problems: Movement toward Prevention

Strategies Alfonso Paredes, Section Editor
Overview Alfonso Paredes
Estimating Alcoholic Prevalence Charles J. Furst
The Role of Alcohol Availability in Alcohol Consumption and Alcohol
Problems Jerome Rabow and Ronald K. Watts
407
408 Contents of Previous Volumes
Price and Income Elasticities and the Demand for Alcoholic Beverages Stanley I.
Ornstein and David Levy
Youth, Alcohol, and Traffic Accidents: Current Status Richard L. Douglass
IV. Current Concepts in the Diagnosis of Alcoholism James A. Halikus, Section

Editor
Overview James A. Halikas
Detection, Assessment, and Diagnosis of Alcoholism: Current
Techniques George R. Jacobson
Types and Phases of Alcohol Dependence Illness Wallace Mandell
Neuropsychology of Alcoholism: Etiology, Phenomenology, Process, and
Outcome Ralph E. Tarter and Christopher M. Ryan
Volume 2
I. Experimental Social and Learning Models of Drinking Alfonso Puredes,

Section Editor
Overview Alfonso Puredes
A Conditioning Model of Alcohol Tolerance Christine L. Melchior and Boris
Tabakoff
Social Models of Drinking Behavior in Animals: The Importance of Individual
Differences Gaylord D. Ellison and Allen D. Potthoff
Social Correlates of Drinking in Contrived Situations Alfonso Paredes and Carolyn
Jenuine Hopper
Alcohol-Ingestive Habits: The Role of Flavor and Effect Jack E. Sherman, Kenneth W.
Rusiniak, and John Garcia
Commentary on the Utility of Experimental Social and Learning Models of
Alcoholism Frank A. Holloway, O.H. Rundell, Pamela S. Kegg, Dick Gregory, and
Thomas Stanitis
II. Alcohol and the Liver: Recent Developments in Preclinical and Clinical
Research Richard A. Deitrich, Section Editor
Overview Charles S. Lieber
Alcohol-Induced Liver Injury: The Role of Oxygen Ronald G. Thurman, Sungchul
Ji, and John J. Lemasters
Hypermetabolic State and Hypoxic Liver Damage Yedy Israel and Hector Orrego
Commentary on the Hypermetabolic State and the Role of Oxygen in Alcohol-
Induced Liver Injury Esteban Mezey
Alcohol-Induced Mitochondrial Changes in the Liver Ellen R. Gordon
Contents of Previous Volumes 409
Effect of Ethanol on Hepatic Secretory Proteins Dean J. Tuma and Michael F.

Sorrell
Use of Colchicine and Steroids in the Treatment of Alcoholic Liver Disease John T.
Galambos and Stan P. Riepe
III. Aging and Alcoholism Edward Gottheil, Section Editor

Neurobiological Relationships between Aging and Alcohol Abuse Gerhard Freund
Alcohol Consumption and Premature Aging: A Critical Review Christopher Ryan
and Nelson Butters
Aging and Alcohol Problems: Opportunities for Socioepidemiological
Research Richard L. Douglass
Life Stressors and Problem Drinking among Older Adults John W. Finney and
Rudolf H. Moos
Cross-Cultural Aspects of Alcoholism in the Elderly Joseph Westermeyer
IV. Contributions from Anthropology to the Study of Alcoholism Linda A.

Bennett, Section Editor
Overview Linda A. Bennett
Ethnohistory and Alcohol Studies Thomas W. Hill
Social-Network Considerations in the Alcohol Field Carl A. Maida
Alcohol Use in the Perspective of Cultural Ecology Andrew J. Gordon
Selected Contexts of Anthropological Studies in the Alcohol Field:
Introduction Dwight B. Heath
Family Research and Alcoholism Joan Ablon
Alcoholism-Treatment-Center-Based Projects Jack O. Waddell
Cross-Cultural Studies of Alcoholism Dwight B. Heath
Volume 3
I. High-Risk Studies of Alcoholism Donald W. Goodwin, Section Editor

Overview Donald W. Goodwin
Behavioral Effects of Alcohol in Sons of Alcoholics Marc A. Shuckit
The EEG in Persons at Risk for Alcoholism Jan Volavka, Vicki Pollock, William F.
Gabrielli, Jr., and Sarnoff A. Mednick
Psychopathology in Adopted-Out Children of Alcoholics: The Stockholm Adoption
Study C. Robert Cloninger, Michael Bohman, Soren Siguardsson, and Anne-Liis von
Knorring
Premorbid Assessment of Young Men at High Risk for Alcoholism Joachim Knop
Minimal Brain Dysfunction and Neuropsychological Test Performance in Offspring

of Alcoholics Victor M. Hellelbrock, James R. Stabenau, and Michie N. Hesselbrock
II. Prostaglandins, Leukotrienes, and Alcohol Richard A. Deitrich, Section Editor

Overview Erik Anggard
Synthesis of Prostaglandins and Leukotrienes: Effects of Ethanol Robert C.
Murphy and Jay Y. Westcott
Biochemical Interactions of Ethanol with the Arachidonic Acid Cascade Sam N.
Pennington
Brain Arachidonic Acid Metabolites: Functions and Interactions with Ethanol Jay Y.
Westcott and Alan C. Collins
III. Cardiovascular Effects of Alcohol Abuse David H. Van Thiel, Section Editor
Overview David H. Van Thiel
Alcohol, Coronary Heart Disease, and Total Mortality Ronald E. LaPorte, Jane A.
Cauley, Lewis H. Kuller, Katherine Flegal, and David Van Thiel
Alcohol Consumption and Cardiovascular Risk Factors Katherine M. Flegal and
Jane A. Cauley
Myocardial Effects of Alcohol Abuse: Clinical and Physiologic
Consequences David H. Van Thiel and Judith S. Gavaler
Biochemical Mechanisms Responsible for Alcohol-Associated
Myocardiopathy David H. Van Thiel, J.S. Gavaler, and D. Lehotay
IV. Cerebral Functioning in Social Drinkers Elizabeth Parker, Section Editor

Overview Elizabeth Parker
The Continuity Hypothesis: The Relationship of Long-Term Alcoholism to the
Wernicke-Korsakoff Syndrome Nelson Butters and Jason Brandt
The Impact of Fathers’ Drinking on Cognitive Loss among Social
Drinkers Elizabeth S. Parker, Douglas A. Parker, and Jacob A. Brody
Alcohol Use and Cognitive Functioning in Men and Women College
Students Roseann Hannon, Charles P. Butler, Carol Lynn Day, Steven A. Khan,
Lupo A. Quitoriana, Annette M. Butler, and Lawrence A. Meredith
CT Demonstration of the Early Effects of Alcohol on the Brain Lesley Ann Cala
Cognitive Deficits and Morphological Cerebral Changes in a Random Sample of
Social Drinkers Hans Bergman
Brain Damage in Social Drinkers? Reasons for Caution Shirley Y. Hill and
Christopher Ryan
Statistical Issues for Research on Social Drinkers Ronald Schoenberg
Functional Brain Imaging Robert M. Kessler
Volume 4
I. Combined Alcohol and Drug Abuse Problems Edward Gottheil, Section Editor
Multiple Drug Use: Epidemiology, Correlates, and Consequences Richard R.
Clayton
Mechanisms of Depressant Drug Action/Interaction Eugene P. Schoener
Sedative Drug Interactions of Clinical Importance Paul Cushman, Jr.
Treating Multiple Substance Abuse Clients Jerome F.X. Carroll
II. Typologies of Alcoholics Thomas F. Babor and Roger E. Meyer, Section Editors
Overview Thomas F. Babor and Roger E. Meyer
Classification and Forms of Inebriety: Historical Antecedents of Alcoholic
Typologies Thomas F. Babor and Richard J. Lauerman
Empirically Derived Classifications of Alcohol-Related Problems Leslie C. Morey
and Harvey A. Skinner
An Examination of Selected Typologies: Hyperactivity, Familial, and Antisocial
Alcoholism Arthur I. Alterman and Ralph E. Tarter
Alcoholic Typologies: A Review of Empirical Evaluations of Common Classification
Schemes Michie N. Hesselbrock
Alcoholic Subtypes Based on Multiple Assessment Domains: Validation against
Treatment Outcome Dennis M. Donovan, Daniel R. Kivlahan, and R. Dale Walker
III. The Alcohol Withdrawal Syndrome Alfonso Paredes, Section Editor

The Alcohol Withdrawal Syndrome: A View from the Laboratory Dora B.
Goldstein
Clinical Neuroendocrinology and Neuropharmacology of Alcohol
Withdrawal Jeffrey N. Wilkins and David A. Gorelick
Clinical Assessment and Pharmacotherapy of the Alcohol Withdrawal
Syndrome Claudio A. Naranjo and Edward M. Sellers
Special Aspects of Human Alcohol Withdrawal David A. Gorelick and Jeffrey N.
Wilkins
IV. Renal and Electrolyte Consequences of Alcohol Abuse David H. Van Thiel,
Section Editor
Disorders of the Serum Electrolytes, Acid-Base Balance, and Renal Function in
Alcoholism Thomas O. Pitts and David H. Van Thiel
Urinary Tract Infections and Renal Papillary Necrosis in Alcoholism Thomas O.

Pitts and David H. Van Thiel
Disorders of Divalent Ions and Vitamin D. Metabolism in Chronic
Alcoholism Thomas O. Pitts and David H. Van Thiel
The Pathogenesis of Renal Sodium Retention and Ascites Formation in Laennec's
Cirrhosis Thomas O. Pitts
Volume 5
I. Alcohol and Memory Henri Begleiter, Section Editor

Overview Henri Begleiter
The Chronic Effects of Alcohol on Memory: A Contrast between a Unitary and Dual
System Approach D. Adrian Wilkinson and Constantine X. Poulos
The Etiology and Neuropathology of Alcoholic Korsakoff's Syndrome: Some
Evidence for the Role of the Basal Forebrain David P. Salmon and Nelson Butters
Cognitive Deficits Related to Memory Impairments in Alcoholism Marlene Oscar-
Berman and Ronald J. Ellis
Specificity of Memory Deficits in Alcoholism Walter H. Riege
Ethanol Intoxication and Memory: Recent Developments and New
Directions Richard G. Lister, Michael J. Eckardt, and Herbert Weingartner
II. Alcohol Treatment and Society Robin Room, Section Editor

Overview Robin Room
Inebriety, Doctors, and the State: Alcoholism Treatment Institutions before
1940 Jim Baumohl and Robin Room
Sociological Perspectives on the Alcoholism Treatment Literature since
1940 Normal Giesbrecht and Kai Pernanen
The Social Ecology of Alcohol Treatment in the United States Connie Weisner
The Great Controlled-Drinking Controversy Ron Roizen
III. The Effects of Ethanol on Ion Channels Richard A. Deitrich, Section Editor
Overview Richard A. Deitrich
Calcium Channels: Interactions with Ethanol and Other Sedative-Hypnotic
Drugs Steven W. Leslie
Effects of Ethanol on the Functional Properties of Sodium Channels in Brain
Synaptosomes Michael J. Mullin and Walter A. Hunt
Involvement of Neuronal Chloride Channels in Ethanol Intoxication, Tolerance, and
Dependence Andrea M. Allan and R. Adron Harris
The Effects of Ethanol on the Electrophysiology of Calcium Channels R.S. Pozos

and S.G. Oakes
The Electrophysiology of Potassium Channels Peter L. Carlen
IV. Hazardous and Early Problem Drinking Alfonso Paredes, Section Editor
Studying Drinking Problems Rather than Alcoholism Dan Cahalan
Social Drinking as a Health and Psychosocial Risk Factor: Anstie’s Limit
Revisited Thomas F. Babor, Henry R. Kranzler, and Richard J. Lauerman
Methods of Intervention to Modify Drinking Patterns in Heavy Drinkers Hans
Kristenson
Techniques to Modify Hazardous Drinking Patterns William R. Miller
Alcohol-Related Hazardous Behavior among College Students Jerome Rabow,
Carole A. Neuman, Ronald K. Watts, and Anthony C.R. Hernandez
Volume 6
I. Substance Abuse and Posttraumatic Stress Disorder Edward Gottheil, Section

Editor
Posttraumatic Stress Disorder and Substance Abuse: Clinical Issues Edgar P.
Nace
The Interrelationship of Substance Abuse and Posttraumatic Stress Disorder:
Epidemiological and Clinical Complications Terence M. Keane, Robert J. Gerardi,
Judith A. Lyons, and Jessica Wolfe
Biological Mechanisms in Posttraumatic Stress Disorder: Relevance for Substance
Abuse Thomas R. Kosten and John Krystal
Coping and Defending Styles among Vietnam Combat Veterans Seeking Treatment
for Posttraumatic Stress Disorder and Substance Use Disorder Walter E. Penk,
Robert F. Peck, Ralph Robinowitz, William Bell, and Dolores Little
Posttraumatic Stress Disorder in World War II and Korean Combat Veterans with
Alcohol Dependency Keith A. Druley and Steven Pashko
II. Alcohol and Its Management in the Workplace Paul M. Roman, Section Editor
Overview Paul M. Roman
The Epidemiology of Alcohol Abuse among Employed Men and Women Douglas A.
Parker and Gail C. Farmer
Growth and Transformation in Workplace Alcoholism Programming Paul M.
Roman
Constructive Confrontation and Other Referral Processes Harrison M. Trice and

Willim J. Sonnenstuhl
Identification of Alcoholics in the Workplace Walter Reichman, Douglas W. Young,
and Lynn Gracin
Monitoring the Process of Recovery: Using Electronic Pagers as a Treatment
Intervention William J. Filstead
Posttreatment Follow-up, Aftercare, and Worksite Reentry of the Recovering
Alcoholic Employee Andrea Foote and John C. Erfurt
New Occupations and the Division of Labor in Workplace Alcoholism
Programs Terry C. Blum
III. Consequences of Alcohol Abuse Unique to Women David H. Van Thiel,

Section Editor
Effects of Moderate Consumption of Alcoholic Beverages on Endocrine Function in
Postmenopausal Women: Bases for Hypotheses Judith S. Gavaler
Effects of Alcohol Abuse on Reproductive Function in Women Nancy K. Mello
Maternal Ethanol Use and Selective Fetal Malnutrition Stanley E. Fisher and Peter I.
Karl
Ethanol Metabolism and Hepatotoxicity: Does Sex Make a Difference? David H.
Van Thiel and Judith S. Gavaler
IV. Markers for Risk of Alcoholism and Alcohol Intake Richard A. Deitrich,
Section Editor
Physiological and Psychological Factors as Predictors of Alcoholism Risk Marc A.
Schuckit
Brain Evoked Potentials as Predictors of Risk Robert Freedman and Herbert
Nagamoto
Molecular Markers for Linkage of Genetic Loci Contributing to
Alcoholism David Goldman
Blood Markers of Alcoholic Liver Disease Charles S. Lieber
Discriminant Function Analysis of Clinical Laboratory Data: Use in Alcohol
Research Zelig S. Dolinsky and Jerome M. Schnitt
Acetaldehyde and Its Condensation Products as Markers in Alcoholism Michael A.
Collins
Volume 7
I. Alcoholics Anonymous: Emerging Concepts Chard D. Emrick, Section Editor

Overview Chad D. Emrick
A Sociocultural History of Alcoholics Anonymous Harrison M. Trice and William J.
Staudenmeier, Jr,
Alcoholics Anonymous: Membership Characteristics and Effectiveness as

Treatment Chad D. Emrick
Some Limitations of Alcoholics Anonymous Alan C. Ogborne
Alcoholics Anonymous and Contemporary Psychodynamic Theory Edward J.
Khantzian and John E. Mack
Al-Anon and Recovery Timmen L. Cermak
11. Family Systems and Family Therapy in Alcoholism Edward Gottheil, Section
Editor

Family, Alcohol, and Culture Linda A. Bennett
Alcoholism and Family Interaction Theodore Jacob and Ruth Ann Seilhamer
Alcoholism and Family Factors: A Critical Review Jane Jacobs and Steven J. Wolin
Outcomes of Family-Involved Alcoholism Treatment Barbara S. McCrady
III. Serotonin and Alcohol Preference Richard A. Deitrich, Section Editor

Serotonin and Ethanol Preference William J. McBride, James M. Murphy, Lawrence
Lumeng, and Ting-Kai Li
Use of Serotonin-Active Drugs in Alcohol Preference Studies Joseph E. Zabik
Serotonin Uptake Blockers and Voluntary Alcohol Consumption: A Review of
Recent Studies Kathryn Gill and Z. Amit
IV. Clinical Pharmacology in the Treatment of Alcohol Dependence: Manipulation

of Neurobehavioral Mechanisms of Drinking Alfonso Paredes, Section Editor

Serotonin Uptake Inhibitors Attentuate Ethanol Intake in Problem
Drinkers Claudio A. Naranjo and Edward M. Sellers
Serotonin Uptake Blockers and the Treatment of Alcoholism David A. Gorelick
Benzodiazepines in the Treatment of Alcoholism David Nutt, Bryon Adinoff, and
Markku Linnoila
Does Lithium Carbonate Therapy for Alcoholism Deter Relapse Drinking? David C.
Clark and Jan Fawcett
Treatment of Chronic Organic Mental Disorders Associated with
Alcoholism Peter R. Martin, Michael J. Eckardt, and Markku Linnoila
Methodological and Ethical Issues in Alcohol Research Alfonso Paredes
Volume 8
I. The Nature of the Syndrome Thomas F. Babor, Section Editor

The Behavioral Pharmacology of Alcohol and Other Drugs: Emerging
Issues Marilyn E. Carroll, Maxine L. Stitzer, Eric Strain, and Richard A. Meisch
The Dependence Syndrome Concept as Applied to Alcohol and Other Substances
of Abuse Therese A. Kosten and Thomas R. Kosten
Operationalization of Alcohol and Drug Dependence Criteria by Means of a
Structured Interview Linda B. Cottler and Susan K. Keating
From Basic Concepts to Clinical Reality: Unresolved Issues in the Diagnosis of
Dependence Thomas F. Babor, Barbara Orrok, Neil Liebowitz, Ronald Salomon, and
Joseph Brown
II. Social Deviancy and Alcohol Dependence Alfonso Paredes, Section Editor
A Review of Correlates of Alcohol Use and Alcohol Problems in
Adolescence Kathleen K. Bucholz
Drug Use and Its Social Covariates from the Period of Adolescence to Young
Adulthood: Some Implications from Longitudinal Studies Kazuo Yamaguchi
Longitudinal Patterns of Alcohol Use by Narcotics Addicts Yih-Ing Hser,
M. Douglas Anglin, and Keiko Powers
Problem Drinking and Alcohol Problems: Widening the Circle of
Covariation Stanley W. Sadava
III. Biological Issues: Ethanol-Drug Interactions Richard A. Deitrich, Section Editor

Behavioral Aspects of Alcohol-Tobacco Interactions James P. Zacny
Interactions of Ethanol and Nicotine at the Receptor Level Allan C. Collins
Interactions between Alcohol and Benzodiazepines Leo E . Hollister
IV. Emerging Clinical issues in the Treatment of Alcohol and/or Other Drugs of
Abuse Edward Gottheil, Section Editor
Cultural Factors in the Choice of Drugs Dwight B. Heath
Self-Regulation and Self-Medication Factors in Alcoholism and the Addictions:
Similarities and Differences E.J. Khantzian
Treating Combined Alcohol and Drug Abuse in Community-Based
Programs Robert L. Hubbard
Structured Outpatient Treatment of Alcohol vs. Drug Dependencies Arnold M.
Washton
Behavioral Treatment of Alcohol and Drug Abuse: What Do We Know and Where
Shall We Go? Reid K. Hester, Ted D. Nirenberg, and Ann M. Begin
Volume 9
I. Genetic Predisposition to Alcoholism Henri Begleiter, Section Editor

Overview Henri Begleiter
A Longitudinal Study of Children of Alcoholics Marc A. Schuckit
Neuropsychological Factors in Individuals at High Risk for Alcoholism Victor
Hesselbrock, Lance O. Bauer, Michie N. Hesselbrock, and Robert Gillen
Potential Biochemical Markers for the Predisposition toward Alcoholism Robert
Eskay and Markku Linnoila
Neurophysiological Factors in Individuals at Risk for Alcoholism Bernice Porjesz
and Henri Begleiter
Developmental Behavior-Genetic Perspective of Alcoholism Etiology Ralph E.
Tarter
II. Fetal Alcohol Syndrome Donald M. Gallant, Section Editor

Overview Donald M. Galant
Basic Animal Research Stata Norton and Lois A. Kotkoskie
A Revised Estimate of the Economic Impact of Fetal Alcohol Syndrome Ernest L.
Abel and Robert J. Sokol
Clinical Correlations between Ethanol Intake and Fetal Alcohol Syndrome Claire B.
Ernhart
The Effects of Prenatal Alcohol on the Central Nervous System Kathy J. Smith
and Michael J. Eckardt
Multilevel Intervention for Prevention of Fetal Alcohol Syndrome and Effects of
Prenatal Alcohol Exposure Iris E. Smith and Claire D. Coles
III. Vulnerability to Disease in Relatives of Alcoholics David H. Van Thiel,

Section Editor
Vulnerability to Alcoholic Liver Disease Amelia M. Arria, Ralph E. Tarter, and
David H. Van Thiel
Hepatic Encephalopathy Coexistent with Alcoholism Ralph E. Tarter, Amelia M.
Arria, and David H. Van Thiel
Vulnerability to Cardiac Disease Samir Zakhari
IV. Social and Environmental Issues Edward Gottheil and Jeannette L. Johnson,
Section Editors
Overview Edward Gottheil and Jeannette L. Johnson
Adult Children of Alcoholics: The History of a Social Movement and Its Impact on
Clinical Theory and Practice Stephanie Brown
Epidemiological Perspectives on Children of Alcoholics Stephen H. Dinwiddie and
Theodore Reich
Psychological Characteristics of Children of Alcoholics: Overview of Research

Methods and Findings Kenneth J. Sker
From Prevention to Treatment: Issues for School-Aged Children of
Alcoholics James G. Emshoff and Lisa L. Anyan
Treating Adults Raised by Alcoholic Parents Jeannette L. Johnson and Stuart Tiegel
Volume 10
I. Clinical Pathology Alfonso Paredes, Section Editor

The Role of Alcohol in Cocaine Dependence Hari Khalsa, Alfonso Paredes, and
M. Douglas Anglin
Alcohol and Cocaine: Clinical and Pharmacological Interactions David A. Gorelick
Hypothalamic-Pituitary Function during Alcohol Exposure and Withdrawal and
Cocaine Exposure Jeffrey N. Wilkins, David A. Gorelick, Koonlawee Nademanee,
Anna Taylor, and David S. Herzberg
Emergency Room Evaluation of Cocaine-Associated Neuropsychiatric
Disorders Ricardo Mendoza, Bruce L. Miller, and lsmael Mena
Dual-Diagnosis Empirical and Developmental-Humanistic Approaches Alina M.
McKenna and Alfonso Paredes
II. Psychosocial Factors and Treatment Edward Gottheil, Section Editor

Alcohol and Cocaine Abuse: A Comparison of Epidemiology and Clinical
Characteristics Mary H. Closser and Thomas R. Kosten
Prohibition or Liberalization of Alcohol and Drugs? A Sociocultural
Perspective Dwight B. Heath
A Comparison of Drug Conditioning and Craving for Alcohol and Cocaine David B.
Newlin
Psychotherapy and Patient Needs in the Treatment of Alcohol and Cocaine
Abuse Robert J. Schneider and Edward J. Kkantzian
Acute Treatment of Alcohol and Cocaine Emergencies Wanda A. Taylor and
Andrew E. Slaby
III. Pharmacology and Biochemistry Donald M. Gallant, Section Editor

Overview Donald M. Gallant
Neuropharmacology of Cocaine and Ethanol Dependence George F. Koob and
Friedbert Weiss
Recent Advances in Pharmacological Research on Alcohol: Possible Relations with
Cocaine Krystyna M. Wozniak and Markku Linnoila
Molecular Mechanisms Associated with Cocaine Effects: Possible Relationships with
Effects of Ethanol Mary C. Ritz, Michael J. Kuhar, and Frank R. George
Developing and Evaluating New Treatments for Alcoholism and Cocaine

Dependence Charles P. O’Brien, Arthur Alterman, Anna Rose Childress, and
A. Thomas McLellan
IV. Medical Complications of Alcohol and Cocaine Abuse David H. Van Thiel,
Section Editor
Gastrointestinal Complications of Cocaine Abuse David H. Van Thiel and Joshua A.
Perper
Hepatotoxicity Associated with Cocaine Abuse David H. Van Thiel and Joshua A.
Perper
Cardiovascular Complications of Cocaine Abuse Joshua A. Perper and David H.
Van Thiel
Respiratory Complications of Cocaine Abuse Joshua A. Perper and David H. Van
Thiel
Volume 11
I. Social and Cultural Perspectives Dwight B. Heath, Section Editor

Overview Dwight B. Heath
Sociology Helene Raskin White
Anthropology Dwight B. Heath
Psychology Marsha E. Bates
Children and Alcohol Alan R. Lang and Werner G.K. Stritzke
Family Systems Linda A. Bennett and Michael LaBonte
Gender Issues Edith S. Lisansky Gombert
Ethnicity Howard T. Blane
The Workplace Craig R. Janes and Genevieve M. Ames
Public Drinking Eric Single
Substance and Symbol Howard F. Stein
11. Physiology and Biochemistry Richard A. Deitrich, Section Editor

Neuronal Ion Channels Enrico Sanna and R. Adron Harris
Opiod Peptides Janice C. Froehlich and T.K. Li
The Liver David W. Crabb
Genetic Transmission David Goldman
Immunological Aspects Rodney C. Baker and Thomas R. Jerrells
III. Clinical Pathology Alfonso Paredes, Section Editor

Biobehavioral Correlates David V. Gauvin, Eme Y. Cheng, and Frank A. Holloway
Typologies in Women Sara Jo Nixon
Reducing the Desire to Drink: Pharmacology and Neurobiology Ray Z. Litten and
John P. Allen
Molecular Biology and Behavior Ernest P. Noble and Alfonso Paredes
IV. Trends in Treatment Edward Gottheil, Section Editor

Developments in Alcoholism Treatment Laura Schmidt and Constance Weisner
Behavioral Treatment Dennis M. Donovan and G. Alan Marlatt
Pharmacological Treatment David A. Gorelick
Inpatient Treatment Edward P. Nace
Psychodynamic Approaches Nancy Brehm, E.J. Khantzianj, and Lance M. Dodes
Dealing with Alcohol Problems in the Workplace Paul M. Roman and Terry C. Blum
Volume 12
I. Epidemiology Mary C. Dufour and Richard K. Fuller, Section Editors

Overview May C. Dufour and Richard K. Fuller
Vulnerability to Alcoholism in Women: Genetic and Cultural Factors Shirley Y.
Hill
Drinking and Problem Drinking in US Women: Patterns and Recent
Trends Sharon C. Wilsnack and Richard W. Wilsnack
Older Women and Alcohol: Use and Abuse Edith S. Lisansky Gomberg
Violent Victimization among Women with Alcohol Problems Brenda A. Miller
and William R. Downs
Women, Alcohol, and Driving Patricia F. Waller and Frederic C. Blow
Employed Women with Alcohol Problems Who Seek Help from Employee
Assistance Programs: Description and Comparisons Terry C. Blum, Paul M.
Roman, and Eileen M. Harwood
II. Physiology Mary C. Dufour and Richard K. Fuller, Section Editors

Overview Mary C. Dufour and Richard K. Fuller
Gender Differences in Alcohol Metabolism: Physiological Responses to
Ethanol Holly S. Thomasson
Mental and Physical Health Consequences of Alcohol Use in Women Shirley Y.
Hill
Alcohol Effects on Hormone Levels in Normal Postmenopausal Women and in

Postmenopausal Women with Alcohol-Induced Cirrhosis Judith S. Gavaler
Gender Differences in Animal Studies: Implications for the Study of Human
Alcoholism Francine E. Lancaster
Sex Distinctiveness in Effective Genotype Gerald E. McClearn
Sex Differences in Ethanol-Related Behaviors in Genetically Defined Murine
Stocks Byron C. Jones and Keith E. Whitfield
Sex Differences in Mesolimbic Dopamine Responses to Ethanol and Relationship to
Ethanol Intake in Rats Betty A. Blanchard and Stanley D. Glick
Anxiolytic Effects of Steroid Hormones during Estrous Cycle: Interactions with
Ethanol Michelle D. Brot, George F. Koob, and Karen T. Britton
III. Behavior and Treatment Issues Alfonso Paredes, Section Editor

Women, Alcohol, and Sexuality Linda J. Beckman and Kimberly T. Ackerman
Cognitive Psychosocial Performance and Recovery in Female Alcoholics Sara Jo
Nixon and Susan Wagner Glenn
The Emergence of Problem-Drinking Women as a Special Population in Need of
Treatment Laura Schmidt and Constance Weisner
IV. Social and Cultural Issues Edward Gottheil and Ellen F. Gottheil, Section
Editors
Overview Edward Gottheil and Ellen F. Gottheil
Race/Ethnicity and Other Sociocultural Influences on Alcoholism Treatment for
Women Beatrice A. Rouse, James H. Carter, and Sylvia Rodriguez-Andrew
Patterns of Alcohol Use among Ethnic Minority Adolescent Women Ruth W.
Edwards, Pamela Jumper Thurman, and Fred Beauvais
Alcoholism in the Family: A Multicultural Exploration Andrea G. Barthwell
Gender Differences for the Risk of Alcohol-Related Problems in Mulitple National
Contexts: A Research Synthesis from the Collaborative Alcohol-Related
Longitudinal Project Kaye Middleton Fillmore, Jacqueline M. Golding, Steven
Kniep, E. Victor Leino, Carlisle Shoemaker, Catherine R. Ager, and Heidi P. Ferrer
Index
Abused Women’s Aid in Crisis, 389 Age (cont.)
Accidents, 9, 10 differential risk of alcohol-involved violence
Acculturation, 64–66, 67, 72, 74–75 and, 32
Acquaintance rape, 12–13, 20, 22, 25, 28–29 drinking laws and, 275–276, 277
Across-group correlational studies, 274 emergency room visits and, 106–107, 112
Across-time and -group correlational studies, partner abuse and, 345
275–276 rape and, 19
Across-time correlational studies, 274–275 Aggression
ACTH, 177 active, 230
Active aggression, 230; see also Direct aggres- in adolescents: see Adolescents
sion; Indirect aggression; Physical ag- alcohol effects on, 124
gression; Verbal aggression anxiolytic and inhibition-reducing effects of
Acute alcohol effects, 14, 25, 249 alcohol and, 132–133
child abuse and, 362–363 bursts of, 143–146
correlational studies of, 263 definitions of, 82, 230
Addiction Severity Index, 349 direct, 230
Adolescents, 5, 81–99 physical, 230, 231–232
acute incidents of alcohol-related aggression verbal, 231, 232–233
in, 88-90 early-onset, 176–177
association between alcohol use and aggres- effects on alcohol use, 125–126
sion in, 90–92 GABAA-benzodiazepine receptor complex
cocaine use in, 291–293 and, 121, 154–164
developmental trends in, 92–95 gender differences in, 236–237
explanatory models for alcohol-related ag- increase in due to alcohol, 126–127
gression in, 82–84 indirect, 230
extent of alcohol use and aggression in, physical, 231, 232
84–88 verbal, 231, 233
preventing alcohol abuse and aggression in, individual differences in alcohol effects on,
97–98 141–150
Adoption studies, 174, 176 measurement of, 230–233, 237
Adrenocorticotropin, 209 passive, 230
African Americans predisposition to alcohol consumption and,
adolescent aggression in, 85, 88, 92 131-132
cocaine use in, 287 self-administration of alcohol and, 150–154
matriarchal family structure in, 61 stereotyped, 150–151
partner abuse in, 58–61, 62–64, 71–72, 73, 345 theories of intoxicant-induced, 318
poverty in, 58 Aggression machine paradigm, 45, 231–232
problem drinking in men, 59–60 Alcohol cognitions, 69-71
rape by, 19, 20 Alcohol expectancy effects: see Expectancy ef-
rape of, 23 fects
Age Alcohol Expectancy Questionnaire (AEQ), 319
assault and, 48, 49, 51 Alcoholics Anonymous (AA), 311–312, 392, 399
in causal relationship studies, 268 Alcohol in Human Violence (Pernanen), 12, 30
cocaine use and, 287, 288 Alcohol myopia, 45, 362, 391
423
424 Index
Alcohol outlets, 274 Assault, 9, 13, 23, 41–54, 230, 245

Alcohol self-administration, 150–154 by acquaintances, 48, 49, 50, 52
Alcohol State Monopoly (Sweden), 276 causal relationship in, 274–275, 277
Alcohol withdrawal, 132, 290 correlational studies of, 42–45, 274–275, 277
Allopregnanolone, 159–162 defined, 47
Alphaxalone, 160 dependent variables for severity of, 47
Alprazolam, 155 experimental studies of, 45–46
Amend Program, 394, 397 independent variables for severity of, 48
Amity Settlement Services for Education and by intimates, 48, 50, 51, 52–53
Transition Program, 398 limitation of studies on, 46–53
Amphetamines, 289, 290, 318 location of, 48, 49
animal studies of, 130–131 NCVS on, 42, 46–53, 54
intoxication–victimization effects and, 326 perception of assailant’s alcohol use in, 48,
partner abuse and, 346 49, 50
Anger rape, 26 by strangers, 48, 49, 50–51, 52
Anglos: see Whites Attachment styles, 344
Animal studies, 123–134, 139–164; see also Non- Attention deficit disorder, 174, 179
human primate model Australia, 275
alcohol and increased aggression in, 126–
127 Barbiturates, 163–164, 318
alcohol and recipient of attack in, 127–128 intoxication–victimization effects and, 326
alcohol self-administration and aggression partner abuse and, 346, 349
in, 150–154 Battered-husband syndrome, 339
anxiolytic effects of alcohol in, 132–133, 197 Battered-woman syndrome, 339, 352
causal relationship in, 251–253 Beer sales, 276
GABAA –benzodiazepine receptor complex Beer taxes, 275
in: see GABAA–benzodiazepine recep- Benzodiazepine, 163–164,197
tor complex Benzodiazepine agonists, 154–155
individual differences in alcohol effect on Benzodiazepine receptor antagonists, 155–157
aggression, 141–150 Bicuculline, 160
inhibition-reducing properties of alcohol in, Binge drinking
132–133 in adolescents, 92
predisposition to aggression and alcohol child abuse and, 363
use in, 131–132 emergency room visits and, 112
relationship to human alcohol-aggression intoxication–victimization effects and, 322
phenomena, 133–134 partner abuse and, 63, 64, 67, 71, 73
social stress and alcohol consumption in, Bivariate analysis
128–130, 151–152 of child abuse, 361, 364, 366, 370
stress and substance abuse in, 130–131 of intoxication-victimization effects,
Anomie, 61 324–325, 327, 330–331
Antisocial batterers, 342 Blacks: see African Americans
Antisocial/narcissistic batterers, 341 Black-white box, 132
Antisocial personality disorder, 177 Blood alcohol concentration (BAC)
child abuse and, 364–365 in animal studies, 151, 152, 153
early-onset alcoholism and, 175–176, in causal relationship studies, 255
178–184 in emergency room studies, 106–109
monoamine oxidase and, 181 Borderline personality disorder, 181
nonhuman primate model of, 204–206 Borderline/schizoidal batterers, 341
serotonin and, 178–184 Burglary, 23, 24
Anxiolytic effects Bursts of aggressive acts, 143–146
of alcohol, 132–133, 197 [35S]t-Butylbicy-clophosphorothionate (TBPS),
of benzodiazepine receptor antagonists, 157 159
Approval seekers, 340
Artifacts, in causal relationship studies, Caffeine, 131
256–260 CAGE questionnaire, 177
Index 425
California Child abuse (cont.)

cocaine use in, 285–286 by mothers, 371, 372, 378
emergency room studies in, 108, 109, 112, partner abuse correlated with, 346, 351
113–114, 115 perpetrator’s substance abuse and, 358–365
Cambridge–Somerville Study, 366 reporting of, 373, 376–377
Car theft, 23 Child Abuse, Domestic Violence, Adoption
Catecholamines, 131 and Family Services Act, 373
Caucasians: see Whites Child Abuse Potential Scale, 359, 374
Causal relationship, alcohol/violence, Child and Adolescent Service System Pro-
223–224, 245–278 gram, 397
complexity of demonstrating, 248–251 Child physical abuse, 94, 310–311
correlational studies of, 261-277 as antecedent to later substance abuse, 365–367
across-group, 274 definitions and measurement of, 372–375
across-time, 274–275 father vs. mother as perpetrator, 366–367
across-time and -group, 275–276 parental substance problems and, 358–360
confounding variables in, 267–268, partner abuse correlated with, 345
270–273 Child sexual abuse, 94, 310–311, 359, 363
individual differences in, 272 as antecedent to later substance abuse,
individual-level, 262-273 367–369
macro-level, 273–277 contact, 371
natural experiments, 276 controversies regarding memories of,
experimental studies of, 251–261 375–376
animal, 251–253 definitions and measurement of, 372–375
external validity, 254–255 explanations for link to later substance
human, 253–261 abuse, 369–372
individual differences in, 260 explanations for link to perpetrator’s sub-
potential artifacts in, 256–260 stance abuse, 364
nature of question, 246–248 parental substance problems and, 360–362
Centralized model of domestic violence serv- Chlordiazepoxide, 127, 154–155
ices linkage, 397–398 Chronic alcohol effects, 14, 25, 249
Central nervous system, reduced 5-HIAA child abuse and, 362–363
functioning in correlational studies of, 263, 265
alcohol consumption and, 201–202 Cigarette smoking, 95
antisociallike behavior and, 204–206 Cocaine, 224–225, 283–302, 318
as a long-term trait, 200–201 alcohol interactions with, 296–297
Cerebrospinal fluid, 5-HIAA in: see 5-Hydroxy- alcohol use and, 290–291
indoleacetic acid animal studies of, 130–131
Child abuse, 357–380; see also Child physical criminal activity and, 294–295, 296–297, 299,
abuse; Child sexual abuse 300–301
adolescent aggression and, 94–95 data collection in study of, 288
as antecedent to later substance abuse, intoxication–victimization effects and, 326
365–372 other drug use and, 289–290
considerations for future research in, partner abuse and, 349–350
378–379 patterns of use, 288–289
controversies regarding memories of, routes of administration, 289, 300
375–376 sample in study of, 286–288
definitions and measurement of, 372–375 social deviance and violence in adolescence,
epidemiological issues in, 8, 11–12 291–293
ethical and legal issues in, 376–377 social deviance and violence in adulthood,
explanations for link to perpetrator’s sub- 293–296
stance abuse, 362–365 violence and, 284–286
by fathers, 370, 371, 372, 378 according to amount of alcohol used
history of in women alcoholics, 321 with, 297–298
hypothetical situations in determining, 377 Codependents, 399
426 Index
Cognitive impairment Death of parent, 24–25

child abuse and, 362–363 Decentralized model of domestic violence
intoxication–victimization effects and, services linkage, 397–398
318–319 Defenders, 340
partner abuse and, 390 Definitions
Colorado, 394, 397 of aggression, 82, 230
Common–cause model, 83, 95–98 of child abuse, 372–375
Competitive reaction–time paradigm, 232, of violence, 82
254, 256–258, 316 Dehydroepiandrosterone, 160
Conduct disorder, 174, 179, 284 Dehydroepiandrosterone sulfate, 162
child abuse and, 371 Demographics: see Sociodemographic charac-
early-onset aggressivity and, 177 teristics
monoamine oxidase and, 181, 182 Denmark, 275
serotonin and, 178 Dependent variables, for assault severity, 47
Confessions, 310, 352 Depression
Conflict Tactics Scale (CTS), 44, 349, 359, 360, child physical abuse and, 359
373–374 child sexual abuse and, 369, 370
Contact child sexual abuse, 371 cocaine use and, 288
Contra Costa County, California, 112, 113–114, intoxication–victimization effects and, 321,
115 330
Controllers, 340 serotonin and, 182-183
Control theory, 97 Developmental victimology, 378
Coroner data, 113–114 Deviance disavowal hypothesis, 247, 363
Corporal punishment, 363, 373 Diagnostic and Statistical Manual of Mental Dis-
Correlational studies orders (DSM), 284, 287, 371n
of assault, 42–45, 274–275, 277 Diagnostic Interview Schedule, 349
of causal relationship, 261–277 Diazepam, 132, 154
across-group, 274 Differential association theory, 97
across-time, 274–275 Differential risk of alcohol-involved violence, 32
across-time and -group, 275–276 Direct aggression, 230
confounding variables in, 267–268, 270– physical, 230, 231–232
273 verbal, 231, 232–233
individual differences in, 272 Disease concept of alcoholism, 391, 392, 395
individual-level, 262–273 Disinhibition hypothesis, 363
macro-level, 273–277 Disorganization hypothesis, 362
natural experiments, 276 Divorce, 275
Corticosterone, 130, 131, 132 in cocaine users, 287
Corticosterone-binding globulin (CBG), 130 in partner abusers, 343
Corticotropin-releasing hormone, 130 in rapists’ families, 24–25
Cortisol, 177–178, 180, 199, 209 Divorce mediation, 310,352
Crack cocaine, 285, 286, 289, 290, 299, 300, Domestic violence, 263, 265, 268–270; see also
346 Child abuse; Domestic violence serv-
Crime, 9; see also specific crimes ices linkage; Partner abuse
cocaine use and, 294–295, 296–297, 299, 300– Domestic violence services linkage, 311–313,
301 387–401
instrumental, 13 barriers to, 392–393, 394–396
property, 24, 89, 90, 93 examples of current, 392–394
violent: see Violent crimes expertise needed in, 393
Criminal record status, see also Parolees/proba- financial limitations in, 396
tioners; Prison populations mechanisms and logistics lacking for, 394, 395
in causal relationship studies, 268 models for, 396–400
cocaine use and, 287–288 organizational boundaries and, 396
partner abuse and, 345 philosophy and goals in, 394, 395
Cubans, 63, 66, 68–70 scarcity of resources for, 394, 395–396
Index 427
Domestic violence services linkage (cont.) Epidemiological issues, 7–37

screening in, 393 definitions of violence in, 13–14
services in, 394–396 event-based research in: see Event-based re-
Dominance status: see Social dominance search
Dominicans, 64–65 measurement of alcohol use and problems
Dopamine, 130 in, 14–15
Downers, 289, 290 methodological problems in, 15–18
Drinking age laws, 275–276, 277 six dilemmas in, 7–13
Drinking patterns Epinephrine, 199
emergency room visits and, 111–112 Ethnicity, see also specific ethnic and racial
partner abuse and, 67–68 groups
Driving under the influence of alcohol, 291 alcohol-related adolescent violence and, 5
Drug dealing, 285–286, 296 cocaine use and, 287
Drug use, 10–11; see also Substance abuse partner abuse and, 5, 57–75
in adolescents, 89–90, 91–92, 96 acculturation in, 64–66, 67, 72, 74–75
in animal studies, 130–131 drinking patterns in, 67–68
causal relationship to violence, 267, 268, evidence from community and clinical
269, 271 samples, 71–72
intoxication–victimization effects and, intergroup comparisons of, 62–64
325–327 social structure in, 73–74
partner abuse and, 347–350 sociocultural factors in, 66–71
rape and, 28 rape and, 20
Drug Use Forecasting group, 11 Event-based research, 4, 15–17
Drunken comportment theories, 35 on rape, 16, 18–23
Duluth Domestic Abuse Intervention Project, 351 Expectancy effects, 14, 83, 233
assault and, 46
Early-onset aggressivity, 176–177 in causal relationship studies, 259–260
Early-onset alcoholism ethnicity and, 69–71
antisocial personality disorder and, explained, 234
175–176,178–184 intoxication–victimization effects and, 319–320
early-onset aggressivity and, 176–177 partner abuse and, 69–71, 391
molecular genetic findings in, 183–184 in placebo design studies, 229–230
psychobiology of, 177–178 validity of explanation, 237
serotonin and, 121,177–184 Experimental studies, 229–233
type 2 alcoholism and: see Type 2 alcoholism of assault, 45–46
Educational model, 350–351 of causal relationship, 251–261
Electric shock administration experiments, 45, animal, 251–253
231–232, 254, 256–258 external validity, 254–255
Elevated plus-maze, 132, 160 human, 253–261
Elicitation hypothesis, 34 individual differences in, 260
Emergency room (ER) studies, 5, 16, 43, potential artifacts in, 256–260
105–116 Explosive rage, 124, 178–179, 343
blood alcohol concentration in, 106–109 Exposed rescuers, 340–341
drinking patterns and alcohol-related prob- Externalizing behaviors, 370–371
lems in, 111–112 External validity, of causal relationship stud-
limitations to, 114–116 ies, 254–255
regional comparisons of data, 113–114
self-reports of alcohol use in, 109–111 False memories, 376
Empirical studies, 323–329 Family of origin
Enablers, 398, 399 intoxication–victimization effects and,
England: see Great Britain 320–321, 325, 331–332
Epidemiological Catchment Area (ECA) samples partner abuse and, 339, 340, 342–343, 344
on child abuse, 359, 368, 369 type 2 alcoholism and, 178
on partner abuse, 65 Fatal injuries, 9, 10, 105, 113–114
428 Index
Feminist theory, 317, 338, 339 Heroin, 285, 289, 290

Fenfluramine, 181–182 intoxication–victimization effects and, 326
Finland, 175, 179, 183 partner abuse and, 346, 349
causal relationship studies in, 275 5-HIAA: see 5-Hydroxyindoleacetic acid
early-onset aggressivity studies in, 176–177 Hispanics
emergency room studies in, 43 adolescent aggression in, 92
Fish, 142 cocaine use in, 287
Flumazenil, 157 partner abuse in, 58–59, 61–73, 74–75
Fluoxetine, 204 poverty in, 58
France, 274, 275 problem drinking in men, 61–62
Freebasing, 289 Histrionic/narcissistic batterers, 341
Free-operant paradigm, 232 Homicide, 9 , 23, 230, 245
Frontal cortex, serotonin in, 207–208 adolescents and, 85, 88
F tests, 242 causal relationship in, 274–276, 277
correlational studies of, 42, 274–276, 277
GABAA–benzodiazepine receptor complex, event-based research on, 18, 20
121, 154–164 gang-related, 285–286
neurochemical mechanisms for aggression, instrumental, 13
162–164 intoxication–victimization effects and,
neurosteroids and, 159–162 322–323
Galanin, 130 nonprimary, 276
Gambling, 296 partner abuse and, 63, 124, 388, 390
γ-Aminobutyric acid: see GABA primary, 276
Gang-related homicide, 285–286 rape and, 23
Gender type 2 alcoholism and, 175
assault and, 46, 48, 49 Homosexuals, 125
in causal relationship studies, 268 Human immunodeficiency virus (HIV), 125
differential risk of alcohol-involved violence Human Subjects Subcommittee, 287
and, 32 HVA, 178
emergency room visits and, 106–107, 112 5-Hydroxyindoleacetic acid (5-HIAA),
expectancy effects and, 319–320 178–179, 183, 284
indiscrimination hypothesis on, 35 alcohol consumption and, 201–202
intoxicated aggression differences and, antisociallike behavior and, 205, 206
236–237 environmental influences on low, 208–211
Gender roles, 71–72, 74 genetic influences on low, 208
Generalists in deviance, 95, 96 impulsive behaviors and, 207–208
General population studies, 4, 17–18 reduced functioning as a long-term trait,
of intoxication–victimization effects, 324–327 200–201
of rape, 27–30 social ostracism and early mortality linked
Glucocorticoid receptors, 130 to low, 206–207
Glycine receptors, 162 violence and, 202–204
Great Britain Hypothalamic-pituitary-adrenal (HPA) axis,
causal relationship studies in, 275 130, 131, 178, 197
emergency room studies in, 114–115 Hypothalamic-pituitary-gonadal axis, 178
Group counseling for batterers, 310, 392 Hypothetical situations, 377
Guatemalans, 65 Hysterical personalities, 341
Guided doing, 35
Gun possession, 285, 296 Illinois, 392, 397, 399
Impulse control
Hallucinogens, 289, 290 serotonin and, 178–179, 181, 201, 207–208
Hazelden (Minnesota) model of treatment, type 2 alcoholism and, 192
391–392 Incorporators, 340
Health and Human Development Project Independent variables, for assault severity,
(HHDP), 84, 85, 88, 90, 93 48
Index 429
Indirect aggression, 230 Marital status (cont.)

physical, 231, 232 of partner abusers, 343
verbal, 231, 233 of rapists, 25
Indirect cause, 233, 234, 237–239 Marital violence: see Partner abuse
Indiscrimination hypothesis, 34–36, 44–45 mCPP: see Meta-chlorophenylpiperazine
Individual-level correlational studies, 262– Measurement
273 of aggression, 230–233, 237
Inhibition-reducing properties of alcohol, 132– of child abuse, 372–375
133 Memory of child abuse, 375–376
Injury severity, 106 Men, see also Gender
Instrumental crimes, 13 adolescent aggression in, 88, 94
Intercede Program, 397 childhood sexual abuse in, 368–370
Intoxication–aggression effects, 308–309, 316, problem drinking in African-American,
318 59–60
Intoxication–victimization effects, 308–309, problem drinking in Hispanic, 61–62
315–332 types of violence reported by, 32–33
in clinical populations, 327–329 Meta-analysis
cognitive effects and, 318–319 of alcohol use and aggression, 235, 236
drinking at time of assault, 326 of alcohol use and violent crime, 94
drinking by both parties in, 322–323 in causal relationship studies, 253, 256, 265,
drug use at time of assault, 326–327 268, 272
dynamics of, 329–332 of child abuse, 364
empirical studies of, 323–329 Meta-chlorophenylpiperazine (mCPP), 180
expectancy effects and, 319–320 3-Methoxy-4-hydroxyphenylglycol (MHPG),
family of origin and, 320–321, 325, 331–332 178, 201
personality and, 321 N-Methyl-D-aspartate (NMDA), 154, 162
severity of alcohol abuse in, 328, 329 Mexican Americans, 63, 64, 65, 66, 67, 68, 70,
social dynamics and, 321–322 71–72, 74
Intraindividual theory, 317 Mexico, 108, 110–111
Irish, 59 MHPG: see 3-Methoxy-4-hydroxyphenylglycol
Mice, 126, 127, 129, 131–132, 142–143, 145,
Karolinska Scales of Personality, 178 150
causal relationship in, 253
Large neutral amino acid (LNAA), 180 GABAA –benzodiazepine receptor complex
Late-onset alcoholism, 174, 179–180; see also in, 154–155, 157, 160
Type 1 alcoholism Michigan Alcoholism Screening Test (MAST),
Lesbian relationships, 124 393
Literature search procedures, 235 Military service, 25
Liver cirrhosis, 277 Millon Clinical Multiaxial Inventory, 341, 342–
Los Angeles, 285–286 343
Mineralocorticoid receptors, 130
Macaque monkeys, 142, 150 Minnesota Coalition for Battered Women, 391
Machismo drinking, 62, 64–65, 69, 75 Minnesota (Hazelden) model of treatment,
Macro-level correlational studies, 273–277 391–392
Mandatory arrests for batterers, 310, 352–353 Minnesota Multiphasic Personality Inventory
MAO: see Monoamine oxidase (MMPI), 340, 342
Marijuana, 285, 289, 290 Mississippi, 110, 113–114
adolescent use of, 89–90, 91–92, 95, 96 Monitoring the Future, 91
intoxication–victimization effects and, Monoamine oxidase (MAO), 180–182,
324–325 183–184, 195, 201, 208
partner abuse and, 349 Monoamine oxidase A (MAOA), 180, 183–184
Marital status Monoamine oxidase B (MAOB), 180, 184
assault and, 51 Morphine, 130
of cocaine users, 287 Mortality, early, 206–207
430 Index
Multidimensional Personality Questionnaire, Nonhuman primate model (cont.)

175 difficulty producing alcohol abuse in,
Multivariate analysis 193–194
of assault severity, 47, 52 human parallels to, 194
in causal relationship studies, 268–270, 271, interindividual differences in, 196
272–273 peer-only rearing of, 197–199, 210–211
of intoxication–victimization effects, reasons for studying, 193
325–327, 328, 331 serotonin and, 199–200
of partner abuse, 63–64, 72, 344 alcohol consumption in, 201–202
Murder: see Homicide antisociallike behavior in, 204–206
Muscimol, 159 early mortality and low 5-HIAA in,
206–207
NAFVS: see National Alcohol and Family Vio- etiology of low 5-HIAA in, 208–211
lence Survey impulsive behaviors in, 207–208
Narrative procedures, 235 reduced functioning as long-term trait in,
National Alcohol and Family Violence Survey 200–201
(NAFVS) social ostracism and low 5-HIAA in,
on intoxication–victimization effects, 324, 206–207
325, 326 violence in, 202–204
on partner abuse in Hispanics, 63, 65–71 type 1 alcoholism in, 196–199, 211
National Child Abuse and Neglect Data Sys- type 2 alcoholism in, 199–200, 211
tem, 373 Nonprimary homicide, 276
National Crime Victimization Survey (NCVS), Norepinephrine, 178, 192, 199, 201, 202, 210,
5, 42, 46–53, 54 211
on adolescent crime, 85 North Carolina, 389
analytic procedures in, 48 Norway, 275
findings of, 48–51 NYS: see National Youth Survey
measures in, 47–48
sample in, 46–47 Observational learning, 97
National Family Violence Resurvey (NFVR), Ohio, 397
43–44, 54, 373 Opiates, 318
National Family Violence Survey, 60, 63, 66 Oxazepam, 155
National Institute on Alcohol Abuse and Alco-
holism (NIAAA), 391 Para-chlorophenylan (PCPA), 205
National Research Council Report, 96 Parental protectiveness, 363–364
National Women’s Study, 372 Parental Punitiveness Scale, 374
National Youth Survey (NYS), 85, 89, 96 Parent–child Conflict Tactics Scale (PCCTS),
Natural experiments, 276 374
Natural History Interview, 288 Parolees/probationers, see also Criminal re-
NCVS see National Crime Victimization Sur- cord status; Prison populations
vey child abuse in, 360, 366
Neuroimaging, 184 cocaine use in, 287–288
Neurosteroids, 159–162 partner abuse in, 349
New York City, 285 Partial benzodiazepine agonists, 157–159
NFVR: see National Family Violence Resurvey Partner abuse, 32–33, 124, 125, 245, 307–310,
NMDA: see N-Methyl-D-aspartate 337–353; see also Domestic violence;
Noise delivery experiments, 254, 257 Domestic violence services linkage
Nonexposed altruists, 340–341 blaming the victim of, 317, 339,399
Nonfatal accidents, 9 causes of, 338–339
Nonhuman primate model, 121–122,150–151, characteristics of male batterers in, 339–
191–211 347
biological and behavioral measures of, correlational studies of, 42, 43–44
195–196 drinking by victim in, 43–44; see also Intoxi-
causal relationship in, 252, 253 cation–victimization effects
Index 431
Partner abuse (cont.) Prison populations, 9, 16; see also Criminal re-
drug use and, 347–350 cord status; Parolees/probationers
ethnicity and, 5, 57–75 assault studied in, 42–43
acculturation in, 62–66, 67, 72, 74–75 partner abuse studied in, 343–344
alcohol as temporal antecedent of, 68–69 rape studied in, 23–27
drinking patterns in, 67–68 Probationers: see Parolees/probationers
evidence from community and clinical Prolactin, 180, 181–182
samples, 71–72 Property crimes, 24, 89, 90, 93
intergroup comparisons of, 62-64 Psychodynamic theory, 338, 339, 340, 346–347
social structure in, 73–74 Psychopharmacological model, 82–83
sociocultural factors in, 66–71, 74–75 Psychotic wife assaulters, 340–341
explanation of alcohol relationship to, Public Law 102–295, 373
390–392 Puerto Ricans, 62, 63, 65, 66, 67, 68, 74
murder of batterer in, 322
role of alcohol in, 388–390 Quipazine, 204
substance abuse by victim in, 389–390,
392 Race: see Ethnicity
treatment programs for batterers, 309–310, Rape, 18–30, 230
350–353 by acquaintances, 12–13, 20, 22, 25, 28–29
victimization theories on, 317–318 of adolescents, 88
by women, 320 anger, 26
Passive aggression, 230 causal relationship in, 274, 275
PCP: see Phencyclidine cocaine use and, 301
Peer-only rearing, 197–199, 210–211 convictions for, 19
Pernanen, Kai, 30–36 defined, 18
Personality, intoxication–victimization effects drinking by victim in, 21, 22, 27–28
and, 321 epidemiological issues in, 12–13
Personality behavioral models, 340–345 event-based research of, 16, 18–23
Personality disorders, 340, 344 general population studies of, 27–30
Phencyclidine (PCP), 289, 290, 318, 346 intoxication–victimization effects and, 322, 323
Philadelphia, Pennsylvania, 389 power, 26
Physical aggression prison offender studies of, 23–27
direct, 230, 231–232 rate of, 19
indirect, 231, 232 by relatives, 25, 28–29
Physiological disinhibition, 233–234, 237 sadistic, 26
Pittsburgh Veterans Affairs Medical Center, by spouse, 28–29
397 statutory, 18
Placebo design studies, 229–230 by strangers, 21, 28–29
of causal relationship, 259–260 underreporting of, 19
Platelets victim precipitated, 21–22
monoamine oxidase in, 180–182 Rationing of alcohol, 276
serotonin uptake to, 182–183 Rats, 126, 127, 129–131, 132–133, 141, 151, 152
Posttraumatic stress disorder (PTSD) causal relationship in, 253
in child sexual abuse victims, 371–372 GABAA –benzodiazepine receptor complex
in partner abuse victims, 389 in 154–157
Poverty, 58, 63–64, 74; see also Socioeconomic individual differences in, 142–143, 145,
status 149-150
Power rape, 26 Reinforcement, 97
Preclinical studies: see Animal studies Rhesus monkeys, 151; see also Nonhuman pri-
5β-Pregnan-3α-ol-20-one, 160 mate model
Pregnenolone sulfate, 161–162 Risk factors for violence, 267, 268, 269, 271
Primary alcoholics, 173 Ro15-4513, 157–159
Primary homicide, 276 Robbery, 9, 23, 230–231, 245, 274, 275
Primates: see Nonhuman primate model Rodents, 127, 180,197; see also specific types
432 Index
Sadistic rape, 26 Social status: see Social dominance

Scandinavian countries, 114 Social stress, 128–130, 151–152
SCL-90-R, 349 Social stress deaths, 129
Secobarbital, 89–90 Social structure, 73–74
Secondary alcoholics, 173 Sociocultural theory, 66–71, 74–75, 317
Self-esteem Sociodemographic characteristics
child sexual abuse and, 369 in causal relationship, 267, 268, 269, 271
intoxication–victimization effects and, 321, of cocaine users, 287
330 Socioeconomic status, see also Poverty
partner abuse and, 340 in causal relationship, 268
Self-help model, 350–351 partner abuse and, 58, 64
Self-reports of alcohol use, 14 Sociopathic batterers, 342
in emergency room studies, 109–111 Sociopathic/narcissistic batterers, 342
by rapists, 19 Spain, 110–111
Serotonin (5-HT), 130,154, 284; see also 5-Hy- Specialists in deviance, 95
droxyindoleacetic acid Speedballs, 289, 290
alcohol consumption and, 201–202 Spouse abuse: see Partner abuse
antisociallike behavior and, 204–206 Spurious model, 83, 92, 95; see also
antisocial personality disorder and, 178–184 Common–cause model
early mortality and low 5-HIAA, 206–207 Squirrel monkeys, 127, 141, 142,146–150, 155,
early-onset alcoholism and, 121,177–184 157, 158, 159
etiology of low 5-HIAA, 208–211 Statutory rape, 18
impulsive behaviors and, 207–208 Steady drinking
late-onset alcoholism and, 179–180 intoxication–victimization effects and, 322
nonhuman primate model of: see under partner abuse and, 73
Nonhuman primate model Stereotyped aggression, 150–151
reduced CNS functioning as a long-term Stress, see also Social stress
trait, 200–201 alcohol consumption and, 197–199
social ostracism and low 5-HIAA, 206–207 anxiolytic effects of alcohol and, 132–133
turnover rate of, 178–179 serotonin and, 202
type 2 alcoholism and, 178,179–180, 182, substance abuse and, 130–131
183, 192, 199–200, 201 Substance abuse, see also Drug use
uptake to platelets, 182–183 in animal studies, 130–131
violence and, 202–204 child abuse as antecedent to, 365–372
Sertraline, 202 child abuse linked to, 358–365
Severity hypothesis, 33–34, 44–45, 49 partner abuse and, 347–350
Sex: see Gender Suicide/suicidal behavior, 9, 177
Sex offenses, 9, 245, 345; see also Child sexual causal relationship in, 274
abuse; Rape child sexual abuse and, 369
Shame, 310, 344, 352 serotonin and, 179, 183, 207
Single-photon emission computed tomogra- Sweden, 274–275, 276, 277
phy (SPECT), 184 Systems theory, 317
Social dominance, 149–150, 153, 195, 252
serotonin and, 205–206, 210 Taylor–Johnson Temperament Analysis (T-
Social dynamics, of intoxication–victimization JTA), 344
effects, 321–322 TBPS: see [35S]t-Butylbicy-clophosphorothionate
Social Fridays, 62 Teacher–learner paradigm, 254, 256–258
Social isolation, 151 Testosterone, 126–127, 130, 177, 203–204
Social learning theory, 97 Tetrazole, 160
on partner abuse, 317–318, 340, 345–346 Traffic accidents, 9
Social ostracism, 206–207 Trait anxiety, 177, 192, 197
Social play, 196–197 Tranquilizers, 289, 326
Social psychological theory, 317 Transitional model of domestic violence serv-
Social separation, 198, 201 ices linkage, 397
Index 433
Tridimensional Personality Questionnaire, 176 Victim–offender relationship, 35, 37

Tryptophan, 180, 204 in assault, 46, 48, 49, 52
Tryptophan hydroxylase (TPH), 183, 204–205 in rape, 21
t tests, 242 Victim precipitation
Twin studies, 208 of rape, 21–22
Type 1 alcoholism, 174–175, 200 of violence, 32
brain neuroimaging in, 184 Violence-prone personality, 342
characteristics of, 192 Violent crimes, 24, 31, 227–228, 230
nonhuman primate model of, 196–199, 211 by adolescents, 85, 89, 91–92, 93, 94
serotonin and, 180 antisocial personality disorder and, 176
Type 2 alcoholism, 174–175, 177 causal relationship in, 263, 265, 270
antisocial personality disorder and, 176, cocaine use and, 300
204, 205 drug use and, 11
characteristics of, 192 Visible burrow system (VBS), 129, 131
child abuse and, 365
genetic transmission of, 208 Wales, 275
monoamine oxidase and, 181 Whites
nonhuman primate model of, 199–200, 211 adolescent aggression in, 85, 92
serotonin and, 178, 179–180, 182, 183, 192, cocaine use in, 287
199–200, 201 partner abuse in, 62–64, 66, 71–73, 74–75,
Type B alcoholism, 365 345
Typical batterers, 342 rape by, 19, 20, 22
Tyrants, 340–341 serotonin and, 183
Tzeltal, 62 Women, 16; see also Gender
adolescent aggression in, 85, 88, 94
U3, 265 assault of, 52–53
Unemployment, 275, 345 drinking by, 14
Uniform Crime Reports (UCR), 18, 274, 275 types of violence reported by, 32–33
violent crimes against, 31
Vandalism, 88 Work history
Verbal aggression of cocaine users, 287
direct, 231, 232–233 of rapists, 25
indirect, 231, 233
Vervet monkeys, 151, 197, 204–205 X-linked transmission of alcoholism, 175
Victim blaming, 317, 339, 399
Victimization theories, 317–318 ZK 93426, 157, 158

Richard Fuller, Donald Gallant, Donald W. Goo-Alcoholism & PDF

Uploaded by

Copyright:

Available Formats

Richard Fuller, Donald Gallant, Donald W. Goo-Alcoholism & PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Richard Fuller, Donald Gallant, Donald W. Goo-Alcoholism & PDF

Uploaded by

Copyright:

Available Formats

RECENT DEVELOPMENTS IN

An Official Publication of the American Society of Addiction Medicine

KLUWER ACADEMIC PUBLISHERS

©2002 Kluwer Academic Publishers

Print ©1997 Kluwer Academic / Plenum Publishers

All rights reserved

Created in the United States of America

Visit Kluwer Online at: http://kluweronline.com

Research Society on Alcoholism American Society of

Ronet Bachman, Department of Sociology, University of Delaware, Newark,

James H. Derzon, Vanderbilt Institute for Public Policy Studies, Vanderbilt

William R. Downs, Center for the Study of Adolescence, University of

Annemoon M. M. van Erp, Department of Psychology, Tufts University,

Richard K. Fuller, Division of Clinical and Prevention Research, National

Edward Gottheil, Department of Psychiatry and Human Behavior, Thomas

Ellen F. Gottheil, Department of Psychiatry and Behavioral Sciences, Univer-

J. Andy Henrie, Department of Chemistry, University of Hawaii at Manoa,

J. Dee Higley, Laboratory of Clinical Studies, Primate Unit, National Institute

Glenda Kaufman Kantor, Family Research Laboratory, University of New

W. Vernon Lee, Penn Recovery Systems, Philadelphia, Pennsylvania 19104-1953

Markku Linnoila, Laboratory of Clinical Studies, Division of Intramural Clin-

Mark W. Lipsey, Vanderbilt Institute for Public Policy Studies, Vanderbilt

Eugene Maguin, Research Institute on Addictions, Buffalo, New York 14203

Susan Ehrlich Martin, Prevention Research Branch, Division of Clinical and

Klaus A. Miczek, Departments of Psychology, Psychiatry, and Pharmacol-

Brenda A. Miller, Research Institute on Addictions, Buffalo, New York 14203

MarIee Moore-Gurrera, Health and Social Policy Division, Research Triangle

Judith Roizen, Institute of Population Studies, University of Exeter, Exeter

Walter Tornatzky, Department of Psychology, Tufts University, Medford,

Matti Virkkunen, Department of Psychiatry, Helsinki University Central

Stephen P. Weinstein, Department of Psychiatry and Human Behavior,

David B. Wilson, Vanderbilt Institute for Public Policy Studies, Vanderbilt

Errol Yudko, Department of Psychology, University of Hawaii at Manoa,

From the President of the Research Society on Alcoholism

On behalf of the Research Society on Alcoholism, I am pleased to introduce

Ivan Diamond M.D., Ph.D.

From the President of the American Society of Addiction Medicine

The American Society of Addiction Medicine is honored to continue its co-

David E. Smith, M.D.

Epidemiological Issues in Alcohol-Related Violence

3.3. The Importance of Race and Social Structure . . . . . . . . . . . . . . . 73

3.2. Alcohol Interactions with Neurosteroids . . . . . . . . . . . . . . . . . . . 159

7. Studies of Interindividual Differences . . . . . . . . . . . . . . . . . . . . . . . . . . 196

IV. Family Issues

Contents of Previous Volumes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 407

Richard K. Fuller, Section Editor

It is both appropriate and timely that this volume of Recent Developments in

Epidemiological Issues in Alcohol-

Abstract. Epidemiological research on alcohol and violence exhibits a number of methodologi-

1. Defining the Problem

Table I. Summary of Studies Reporting Alcohol Presencea at the Time

Number of studies Range

is now present in violent behavior, especially criminal behavior, to a degree

It is a conclusion of many who write on alcohol and violence (e.g., Pernanen,2

alcoholism and child abuse. . . . Although from a scientific point of view it

event, and it proved to be a fairly weak predictor. Alcohol use by men is

1.2. Definitions of Violence

The focus of most research on alcohol and violence, especially criminal

1.3. Measurement of Alcohol Use and Alcohol Problems