Kavita Mam

See discussions, stats, and author profiles for this publication at: https://www.researchgate.
net/publication/263551923
The Power of Diversion: Intermediate Sanctions and

Sentencing Disparity under Presumptive Guidelines
Article in Criminology · August 2012

DOI: 10.1111/j.1745-9125.2012.00279.x
CITATIONS READS
36 162
2 authors:
Brian Johnson Stephanie M. Dipietro

maryland University of Iowa
37 PUBLICATIONS 2,819 CITATIONS 17 PUBLICATIONS 317 CITATIONS
SEE PROFILE SEE PROFILE
All content following this page was uploaded by Brian Johnson on 09 February 2021.
The user has requested enhancement of the downloaded file.

THE POWER OF DIVERSION:
INTERMEDIATE SANCTIONS AND
SENTENCING DISPARITY UNDER
PRESUMPTIVE GUIDELINES∗
BRIAN D. JOHNSON
Department of Criminology and Criminal Justice
University of Maryland
STEPHANIE M. DIPIETRO
Department of Criminology and Criminal Justice
University of Missouri–St. Louis
KEYWORDS: sentencing, disparity, intermediate sanctions, judge, court,
context, hierarchical linear modeling (HLM)
Fiscal constraints and shifting political climates in corrections have
recently led to a renewed interest in intermediate punishments. Despite
their growing prevalence, though, relatively little empirical research has
examined the judicial use of alternative sanctions as a sentencing option.
By using 3 years of data from the Pennsylvania Commission on Sentenc-
ing (PCS), this study investigates little-researched questions regarding
the use of sentencing alternatives among offenders and across contexts.
Results indicate that male and minority offenders are the least likely
to receive intermediate sanctions, both as a diversionary jail or prison
sentence and as a substitute for probation. The probability of receiv-
ing an intermediate sanction also varies significantly across judges and
court contexts and is related to county-level funding for these programs,
among other factors. Findings are discussed as they relate to contempo-
rary theoretical perspectives on the perceived suitability of intermediate
punishments and on the unique role that offender agency plays in the
sentencing of these cases. Directions for future research are discussed.
Additional supporting information can be found in the listing for this arti-
cle in the Wiley Online Library at http://onlinelibrary.wiley.com/doi/10.1111/
crim.2012.50.issue-3/issuetoc.
* The authors would like to thank Jeff Hubert and the staff at the Pennsylvania
Commission on Crime and Delinquency (PCCD) for their assistance in providing
unique county-level data on intermediate sanction programs in Pennsylvania. An
earlier version of this article was presented at the 2006 American Society of Crimi-
nology meetings in Los Angeles, CA. Direct correspondence to Brian D. Johnson,
Department of Criminology and Criminal Justice, University of Maryland, College
Park, MD 20743 (e-mail: bjohnson@crim.umd.edu).

C 2012 American Society of Criminology doi: 10.1111/j.1745-9125.2012.00279.x
CRIMINOLOGY Volume 00 Number 0 2012 1
2 JOHNSON & DIPIETRO
In the past three decades, U.S. incarceration rates have more than tripled,
public expenditures on corrections have quintupled, and more new prisons
have been built than in the entire history of the nation (Austin and Irwin,
2001; Blumstein and Beck, 1999; Ruth and Reitz, 2003). The overwhelming
social and economic costs of mass incarceration have prompted a renewed
interest in alternatives to incarceration. At the vanguard of this movement
are intermediate sanctions, championed for being less costly but sufficiently
punitive sentencing alternatives.1 Although intermediate sanctions come
in sundry forms—electronic monitoring, intensive supervision, drug and
alcohol treatments, day fines, etc.—they all share the common goal of
providing a range of punitive sanctions that in theory fall between probation
and incarceration on the punishment severity continuum. The initial goal
of intermediate sanctions was to divert offenders from prison and jail while
providing increased offender accountability and supervision.
In an effort to increase punishment options and assuage fiscal and spatial
constraints on burdened correctional resources, several states, including
Washington, North Carolina, and Pennsylvania, have incorporated inter-
mediate punishments or similar provisions into their sentencing guidelines
(Tonry, 1996). The need for effective noncustodial punishments has never
been greater; yet, despite the burgeoning interest in their use and imple-
mentation, little is known about the decision-making process that filters
some eligible offenders to alternative sanctions and others to jail or prison.
With few exceptions (Engen et al., 2003; Gainey, Steen, and Engen, 2005),
judicial decisions to impose alternative sanctions, or the consequences of
these decisions for social inequality and “net-widening” concerns, have
received little empirical attention. As Ostrom, Ostrom, and Kleiman (2004:
61) recently opined, “The research community has yet to address the essen-
tial complexity a judge faces in deciding to impose one or more non-prison
sanctions.”
The current study investigates judicial decisions to divert offenders to
intermediate sanctions in the state of Pennsylvania, one of the first states to
incorporate intermediate punishments into its presumptive sentencing grid.
Within predetermined cells of the sentencing matrix, Pennsylvania judges
determine the receipt of intermediate punishments carte blanche. These de-
cisions are of special theoretical import because they not only allow for con-
sideration of special offender circumstances, but they may even encourage
it (Engen et al., 2003). Discretionary application of sentencing alternatives
1. Intermediate sanctions are variously referred to as intermediate punishments,

alternative sanctions, alternative punishments, and graduated sanctions. Although
Morris and Tonry (1990) argued these terms carry subtle connotative differences,
we use them interchangeably to represent the range of sanctions that fall between
probation and incarceration on the continuum of criminal punishments.
THE POWER OF DIVERSION 3
invokes a less structured and more substantively rational decision-making

process than traditional sentencing decisions. Use of sentencing alternatives
also may shift the focus to unique judicial considerations like the perceived
suitability of rehabilitative alternatives for different types of offenders and
offenses, as well as the active role that offenders themselves sometimes
play in the sentencing process. Such considerations are important and
undertheorized practical concerns that may be particularly salient to the
discretionary use of intermediate sanctions under sentencing guidelines.
The current research investigates these issues, contributing to extant
work on criminal sentencing in several ways. It examines the use of a
broader class of judicial diversion decisions in a new state (cf. Engen et al.,
2003; Gainey, Steen, and Engen, 2005). It investigates both the judicial deci-
sion to divert offenders from incarceration and the potential net-widening
effects associated with intermediate punishments. It builds on prior work
that has argued for the unique theoretical importance of studying interme-
diate sanctions (Morris and Tonry, 1990) while highlighting commonly en-
countered data limitations that restrict the ability to test extant theory fully,
and finally, it examines the underlying correlates of jurisdictional variation
in the use of sentencing alternatives, comparing interjudge and intercounty
variation in their use and examining unique measures of funding allocation
for intermediate sanctions at the local level.
INTERMEDIATE SANCTIONS
Theoretically, intermediate sanctions fall along a broad continuum of
punishments between traditional probation and incarceration. The severity
of punishment increases along the continuum from more benign penalties,
including community service and monetary fines, to more restrictive punish-
ments, including house arrest, intensive supervision, and boot camp (Morris
and Tonry, 1990). Although programs vary widely in their implementation,
the cornerstone of intermediate sanctions is their basis in the community.
Stemming largely from the rehabilitation efforts of the 1960s and 1970s, the
underlying goals of early “community corrections” (Petersen and Palumbo,
1997) were to maintain nonviolent offenders, who posed minimal threat to
public safety, within the community (MacKenzie, 2001). Housed in treat-
ment facilities and halfway houses instead of prisons, offenders were able
to receive counseling, education, and job training, without sacrificing critical
ties to neighborhood support networks.
Tonry (1996) has suggested that the shift from rehabilitative community
corrections programs in the 1970s to more punitive intermediate sanc-
tions today was prompted by the convergence of three factors. First, the
collective loss of faith in the rehabilitative movement (Martinson, 1974)
resulted in the abandonment of treatment as the principal rationale for
punishment. Second, the emergence of “just deserts” emphasized the need

for greater proportionality in punishment (MacKenzie, 2001; von Hirsch,
1976). As Morris and Tonry (1990: 54) noted, “Offenders range from the
mildest milequetoasts to the most brutal blackguards,” and the continuum
of intermediate sanctions offered greater opportunity for judges to assign
punishments that more appropriately fit the crime. Third, as crime control
policy became an increasingly important political platform, proponents of
“law and order” lobbied for less reliance on probationary sentences. As a
result, incarceration grew exponentially during the 1980s, particularly for
drug crimes (Blumstein and Beck, 1999). Ironically, one consequence of
the staggering growth in prison populations, and their associated fiscal and
spatial correctional constraints, has been the subsequent need for puni-
tive community-based alternatives to incarceration. As Tonry (1996: 101)
stated, “Intermediate sanctions have been seen as a way both to reduce the
need for prison beds and to provide a continuum of sanctions that satisfied
the just deserts concern for proportionality.”
Recently, the increasing shift toward intermediate sanctions has resulted
in at least three emerging corollary concerns. First, it represents a return to
a discretionary decision-making regime that may be associated with unwar-
ranted differences in punishments among similarly situated offenders. Most
state sentencing systems govern prison sentences exclusively, and those
that do incorporate other sentencing options provide little or no guidance
on how and when to implement them. Second, it introduces opportunities
for more severe punishments to be applied to offenders who would have
otherwise received probation terms. This process, pejoratively referred to
as “net-widening,” often is criticized because it jeopardizes the intended
respite from incarceration proponents of intermediate punishments cham-
pion. And third, because intermediate sanctions are uniquely local pun-
ishments, with considerable variation in program availability and funding
across jurisdictions, the likelihood of similar offenders receiving different
punitive outcomes may be tied to geographical variation in the resources
devoted to sentencing alternatives. Relatively little empirical work has
explicitly examined these issues.
PRIOR RESEARCH ON INTERMEDIATE

PUNISHMENTS
Since the implementation of sentencing guidelines in the 1980s, sentenc-
ing discourse has largely focused on the efficacy of guidelines systems in
eliminating idiosyncratic sentencing practices, increasing uniformity, and
reducing unwarranted disparities. The effects of race, gender, and ethnicity
on traditional decisions regarding the likelihood and length of incarceration
have been well documented (e.g., Blumstein et al., 1983; Mitchell, 2005;
Spohn, 2000; Zatz, 2000). Examinations of the judicial use of intermediate

punishments, however, remain rare. We are aware of only three prior stud-
ies that explicitly focus on the use of intermediate punishments in criminal
court.
Wooldredge and Gordon (1997) collected survey data from a national
sample of 181 chief trial court judges to examine the likelihood of judges
using alternative sanctions. Their findings suggested that the use of alterna-
tive sentences varied according to state sentencing policies, court size, and
the prevalence of plea bargaining in the court. More structured sentencing
contexts and larger courts were associated with lower estimates of the use
of sentencing alternatives, whereas higher rates of plea bargaining were
associated with greater use. This study identifies important jurisdictional
variations in the use of alternative sanctions based on survey data, although
it did not examine data on actual sentencing behavior.
More recently, two studies have examined the use of “structured sentenc-
ing alternatives” in Washington State. Engen et al. (2003) examined the
use of alternative sentencing options as mechanisms for departing from the
Washington sentencing guidelines. In addition to traditional guidelines de-
partures, they examined the probability of offenders receiving three specific
types of alternative sentences that were designed for specific offender types.
For offenders sentenced to 12 months or less, judges could convert jail sen-
tences to “partial confinement” by substituting intermediate punishments.
For minor offenders with no prior convictions, judges could invoke a “first
time offender waiver, . . . which resembles traditional probation in several
ways” but also may include community service or participation in treatment
programs (Engen et al., 2003: 106). Finally, for nonviolent sex offenders,
judges could invoke a special sex-offender alternative that suspends the
standard sentencing range and mandates treatment. In examining the use of
these structured alternatives, Engen et al. (2003) found that they occurred
frequently and that they generally benefited female and White defendants.
Blacks, Hispanics, and males were less likely to receive any of the structured
alternatives.
Similarly, Gainey, Steen, and Engen (2005) examined the use of struc-
tured sentencing alternatives for felony drug offenders in Washington State.
They examined additional sentences to drug offender alternatives and work
ethic camps. The drug sentence alternative applied to relatively minor drug
offenders who had no prior felony convictions and resulted in a 50 percent
discount in prison time in conjunction with mandated treatment. Sentences
to work ethic camp applied only to offenders with presumptive sentences
between 16 and 36 months with no prior violent or sex convictions. Suc-
cessful completion of camp could earn up to a two-thirds discount on one’s
sentence. Findings indicated that Hispanics were particularly unlikely to
receive all types of alternative sanctions. Black and male offenders were
disadvantaged for some but not all alternatives. Importantly, this study also
concluded that the use of intermediate sanctions varied by court and urged
future research on the sources of jurisdictional variations in their use across
contexts.
INTERMEDIATE PUNISHMENTS UNDER THE

PENNSYLVANIA GUIDELINES
Pennsylvania was one of the first states to implement presumptive guide-
lines, and in 1994, as part of a bill intended to reduce the state’s reliance
on incarceration, it became one of the first states to incorporate inter-
mediate punishments into its sentencing grid. Pennsylvania opted for a
simple bifurcated system of punishment interchangeability that provided
for two categories of sentencing alternatives—restorative sanctions (RS)
that included nonconfinement punishment options such as drug testing,
community service, and certain outpatient programs, and restrictive in-
termediate punishments (RIP) that encompassed alternative sanctions re-
quiring formal supervision in the community, such as house arrest with or
without electronic monitoring, halfway houses, and drug and alcohol in-
patient programs along with other sanctions deemed interchangeable with
incarceration. In 1997, the sentencing commission revised the guidelines
to extend the use of intermediate punishments to additional cells in the
sentencing matrix including those recommending state imprisonment (see
table S1 in the online supporting information2 ).
The guidelines are divided into five sentencing levels that determine the
presumptive type of punishment as well as the availability of alternative
sanctions. Level 1 is reserved for the least severe offenders and recom-
mends restorative sanctions such as restitution and fines. Level 2 has a
lower presumptive sentence of probation but also allows judges to sentence
offenders to county incarceration or to apply intermediate punishments.
Level 3 of the guidelines presumes a sentence of incarceration in county
jail, but it provides judges with the discretionary power to divert offenders
from jail sentences to intermediate sanctions. Level 4 of the sentencing
grid presumes an offender will be incarcerated in a state prison but also
allows judges to divert these offenders to an intermediate sanction program.
An equal exchange is provided for alternative sanctions to be substituted
2. Additional supporting information can be found in the listing for this ar-
ticle in the Wiley Online Library at http://onlinelibrary.wiley.com/doi/10.
1111/crim.2012.50.issue-3/issuetoc.
for jail or prison at the discretion of the sentencing judge.3 Level 5 of

the sentencing matrix requires offenders to be incarcerated in state prison
unless they are eligible for boot camp or receive a departure from the
sentencing guidelines.
Several scholars have called for the integration of interchangeable
punishments within the framework of sentencing guidelines (Morris and
Tonry, 1990; Tonry, 1996). Pennsylvania’s guidelines provide for this inter-
changeability by allowing judges to substitute intermediate sanctions within
specified guidelines levels. Although the availability of intermediate pun-
ishments is coarsely defined by the partitioning of the guidelines into levels,
the sentencing guidelines themselves “provide no guidance as to the choice
to be made when more than one sanction type is allowed in a given cell”
(Frase, 2000: 441).
THEORETICAL IMPORT OF INTERMEDIATE

PUNISHMENTS
According to Engen et al. (2003), research on intermediate sanctions
holds special theoretical import in the study of sentencing for several rea-
sons (see also Gainey, Steen, and Engen, 2005). First, the decision to apply
an intermediate punishment is largely discretionary and relatively immune
from organizational oversight. Second, intermediate sanctions explicitly
reintroduce individualized decision-making criteria, or substantively ratio-
nal sentencing considerations, that encourage the use of individual offender
considerations in sentencing. Third, unlike traditional sentencing outcomes,
intermediate sanctions may involve an important degree of individual of-
fender agency. For instance, an offender’s willingness to participate in
intermediate punishment programs may exert important influences on his
or her likelihood of receiving these sentences (Gainey, Steen, and Engen,
2005; Wood and May, 2003).
Moreover, because the goals of intermediate punishments are both to
divert offenders from incarceration and to create a more dynamic sentenc-
ing continuum, they represent a unique vehicle for net-widening under
guidelines. Net-widening occurs when intermediate sanctions are given
to presumptive probationers, or to offenders who would have otherwise
received probation. As Morris and Tonry (1990: 159) observed, they “are
devised to fill in the center gap in a bipolar, bimodal punishment system,”
so “it should come as no surprise” if “those sentenced to the newly avail-
able punishment come from both sides of the scale.” As outlined in the
3. Although the use of alternative sanctions within specified cells of the sentencing
guidelines is wholly unstructured, judges are limited by statutory limitations that
prevent their application to certain serious violent offenders (Pa. C.S.A. § 2154.1).
subsequent discussion, examination of these net-widening processes can

offer unique theoretical insights into the underlying judicial rationales for
diverting some offenders and not others to sentencing alternatives.
Before detailing our theoretical predictions, though, an important caveat
should be noted. Although contemporary theoretical perspectives provide
a useful framework for hypothesizing the effects of a broad range of in-
dividual and contextual factors in sentencing, data limitations often pre-
clude precise testing of the underlying theoretical constructs responsible
for observed relationships. This is true in the current study as well as
in most research on criminal sentencing (Wellford, 2007). First, although
we include the standard correlates of sentencing, we lack data on several
potentially important offender characteristics that may be related to judicial
assessments of amenability to alternative sanctions. These include such
considerations as offender employment history, drug abuse history, marital
status, and educational attainment as well as information on underlying
judicial rationales for invoking intermediate punishments in some cases and
not in others. Although we develop and test several interesting theoretical
predictions, we lack some of the detailed measures necessary to capture
fully specific elements of our theoretical arguments. As such, future re-
search replicating and extending this work will be needed—an important
point that we return to in the conclusion of the article.
SUBSTANTIVE RATIONALITY, ATTRIBUTION, AND FOCAL

CONCERNS
Contemporary theoretical perspectives on criminal court decision mak-
ing provide a useful foundation for investigating intermediate sanctions as
a locus of individual social inequality as well as a potential contributor to
jurisdictional variation in criminal punishments under sentencing guide-
lines. According to Savelsberg (1992), sentencing guidelines represent a
neoclassical reform movement designed to reemphasize formal rationality
over substantive rationality (Weber, 1968). Substantive rationality, in this
context, refers to “the intrusion of economic, sociological, and ethical crite-
ria upon formal-rational reasoning and decision making” (Savelsberg, 1992:
1346–7), whereas formal rationality comprises “a deductive, logical and
gapless system of rules” in legal decision making (Savelsberg, 1992: 1350).
Under guidelines, intermediate punishments represent an important “gap”
in the “gapless system of rules.” Judges in conjunction with other court
actors maintain substantial individual discretion to decide which offenders
get diverted to alternative punishments, and unlike other mechanisms of
guidelines circumvention (e.g., guidelines departures), these sentence ex-
changes are not subject to appeal. Intermediate sanctions, then, represent
“a qualitatively different kind of decision based on substantive rather than
formal legal criteria” (Engen et al., 2003: 108). That is, judges and other
court actors are charged with identifying individual offenders who are par-
ticularly well suited for alternative sanctions based on substantively rational
sentencing considerations.
Contemporary theoretical perspectives suggest that sentencing involves
an attributional decision-making process necessitated by organizational
time and information constraints (Albonetti, 1991). Sentencing decisions
occur in public forums where judges and other court actors decide the fate
of offenders in the face of great uncertainty about their future criminal
behavior. To lessen this uncertainty, court actors engage in a decision-
making process that is the product of past experience, established habit, and
normative courtroom workgroup expectations (Albonetti, 1991; Eisenstein
and Jacob, 1977), which may involve “stereotypes that link race, gender,
and outcomes from earlier processing stages to the likelihood of future
criminal activity” (Albonetti, 1991: 250).
Attributions apply not only to predictions of future behavior but also
to other relevant sentencing considerations. Steffensmeier, Ulmer, and
Kramer (1998) argued that court actors consider three “focal concerns” that
include offender blameworthiness, community protection, and individual
and organizational constraints. The same attribution processes described
by Albonetti (1991) are thought to characterize assessments of all three
types of sentencing concerns. For instance, young Black males may be
“more likely to be perceived as dangerous, committed to street life, and
less reformable than women and older offenders” (Steffensmeier, Ulmer,
and Kramer, 1998: 787) and young Latino males may be more likely to
be perceived as “a gun-wielding, drug-selling gang banger unless proven
otherwise” (Portillos, 1998: 156). This argument suggests that young, male,
and minority offenders are more likely to have negative attributions involv-
ing increased culpability, increased dangerousness, and reduced potential
for reform. Such associations may inhibit judges from using intermediate
sanctions in place of jail and prison for these offenders. Hence, we expect
the following:
Hypothesis 1: Male and minority offenders will be less likely to re-

ceive intermediate punishments relative to jail and prison sentences.
Similar processes may characterize judicial decisions regarding net-

widening, or the use of intermediate sanctions in lieu of probation. If judges
view alternatives as relatively more severe than probation, and if negative
social attributions attach to male and minority offenders, then one might
expect these defendants to be more likely to receive intermediate pun-
ishments than probation. However, as we will argue, there are persuasive
reasons to anticipate an alternative pattern of findings.
SUITABILITY OF PUNISHMENT AND INDIVIDUAL OFFENDER

AGENCY
Intermediate punishments are somewhat unique in that they often in-
clude a greater emphasis on offender rehabilitation and reintegration. To
the extent that negative attributions of male and minority defendants re-
sult in judicial perceptions of these offenders as being less amenable to
treatment, less likely to successfully complete programs, or less “deserving
offenders” (Gainey, Steen, and Engen, 2005: 491), judges may be unlikely
to apply intermediate sanctions to certain offenders, not only in lieu of
incarceration but also as a substitute for probation. Because availability of
intermediate sanction programs is limited, they are a valuable and scarce
court resource. Negative stereotypical attributions may affect judicial “pre-
sumptions of suitability” (Frase, 1991: 329), painting male and minority
defendants as less suitable targets for intermediate sanctions and resulting
in them being less likely to receive these sentences overall.
Intermediate punishments also are unique in that they tend to involve
some degree of individual offender agency. Most theoretical accounts of
sentencing attribute the punishment decision primarily to the judge (e.g.,
Hogarth, 1971), or more rarely to the court workgroup (e.g., Eisenstein,
Flemming, and Nardulli, 1988). In the case of intermediate sanctions,
though, the offender is likely to wield considerable influence over the
final sentence. An offender’s willingness to participate will color judicial
assessments of the relative suitability of sentencing alternatives. In fact, it
often is a prerequisite for diversion, as “common experience instructs that
motivated people are more successful than unmotivated people . . . [and]
it is a waste of scarce resources to provide opportunities to people who
do not want them” (Tonry, 2004: 11). Importantly, prior research has
demonstrated variation in the perceived severity of alternative sanctions,
with male and minority defendants rating them as more severe and ex-
pressing less willingness to participate in them relative to female and White
defendants (May et al., 2005; Wood and Grasmick, 1999; Wood and May,
2003). This finding suggests that, for at least some offenders, intermediate
sanctions do not neatly fall between probation and incarceration on the
punishment continuum. Although we could not specifically separate the
effects of judicial assessments of suitability from offender agency, both
theoretical arguments suggest the following:
Hypothesis 2: Male and minority offenders will be less likely to re-

ceive intermediate punishments relative to probation.
The perceived suitability of punishment also is likely to be affected by

the type of crime committed (Wooldredge and Gordon, 1997). According
to Gainey, Steen, and Engen (2005: 491–3), intermediate sanctions involve

“more utilitarian sentencing objectives such as treatment/rehabilitation,
restitution and restoration, [and] offender reintegration,” which are par-
ticularly well suited to drug offenders. Moreover, some sentencing alter-
natives, such as in-patient clinics, out-patient clinics, and drug and alcohol
treatment programs, directly target drug- or alcohol-involved offenders.
Less serious drug offenders, such as those convicted of possession rather
than of trafficking charges, may be particularly attractive candidates for
diversion from jail and prison to intermediate sanctions. Violent offend-
ers, however, tend to be less attractive candidates for intermediate pun-
ishments. Greater risk is involved with sentencing violent offenders to
community-based punishments, which tend to evoke greater concerns over
dangerousness and community protection among both the court workgroup
and the general public. We therefore expect:
Hypothesis 3: Compared with violent crimes, offenders convicted of

drug or alcohol crimes will be more likely to receive intermediate
punishments relative to jail and prison sentences.
According to organizational efficiency theory, efficient case disposition

emerges as the operational goal that maintains a stable and successful
court system (Dixon, 1995). Therefore, a premium is placed on obtaining
guilty pleas. Some prior research has suggested that intermediate sanctions
may provide one mechanism through which offenders are rewarded for
pleading guilty (Wooldredge and Gordon, 1997). Sentencing alternatives
may be used to induce offenders to plead guilty because trials entail greater
time, cost, and uncertainty for the workgroup (LaFree, 1985). We therefore
expect the following:
Hypothesis 4: Offenders convicted through trials will be less likely to

receive intermediate punishments relative to jail and prison sentences.
COURT COMMUNITIES, LOCAL RESOURCES,

AND CONTEXTUAL VARIATION
According to the court community perspective, punitive decision-making
outcomes are locally determined and vary with the cultural expectations,
organizational structures, and larger community environments (Eisenstein,
Flemming, and Nardulli, 1988; Eisenstein and Jacob, 1977; Ulmer, 1997).
Such court-level factors as its size, departure rate, caseload pressure, local
political milieu, and the availability of criminal justice resources can all
affect the dominant case-processing strategies that emerge. Prior research
has suggested that the use of sentencing alternatives may vary across courts
(Gainey, Steen, and Engen, 2005), but relatively little is known about
the sources of this variation (see Wooldredge and Gordon, 1997). Some
research has suggested that larger courts tend to sentence with relative
leniency (e.g., Ulmer and Johnson, 2004), which could result in the greater
use of sentencing alternatives as viable substitutes for incarceration. Simi-
larly, high rates of guidelines departures may signal a cultural willingness to
individualize punishments and to deviate from traditional guideline expec-
tations (Johnson, 2006). The caseload pressure of the court also can affect
sentencing patterns (e.g., Wooldredge, 1989). Because courts with greater
caseload pressure are likely to place a higher premium on efficient case
disposition, one might expect them to be less inclined to use intermediate
sanctions, which tend to involve a more individualized and specialized sen-
tencing process. Prior theorizing also has argued that punishment is a highly
political process (Helms and Jacobs, 2002), so more conservative court
environments may be less receptive to the use of intermediate sanctions
as substitutes for incarceration.
Alternatively, intercourt variation in punishments also may reflect dif-
ferences in the characteristics of organizational decision makers across
courts (Hogarth, 1971). Extant theorizing, for instance, has suggested that
judicial differences in punitive attitudes, role orientations, and sentencing
philosophies may be tied to sentencing decisions (Gibson, 1978; Johnson,
2006). Prior work in this vein has been limited and has focused primarily on
judicial race (e.g., Spohn, 1990) or gender (e.g., Steffensmeier and Hebert,
1999), arguing that female and minority judges tend to be more liberal and
sympathetic and therefore less punitive at sentencing (Gruhl, Spohn, and
Welch, 1981; Johnson, 2006; Welch, Combs, and Gruhl, 1988). Although
empirical support for these expectations is somewhat limited (Zatz, 2000),
it remains important to control for potential judge-level variations in sen-
tencing when examining the correlates of interjurisdictional variations in
punishment (Johnson, 2006). We therefore include controls for judicial
variation when assessing our measures of court context, with the following
theoretical expectations:
Hypothesis 5: The use of intermediate punishments will vary signifi-

cantly across both judges and courts.
Hypothesis 6: Net of individual- and judge-level controls, offenders
sentenced in courts that are larger, have higher departure rates, lower
caseload pressure, and less conservative political environments will be
more likely to receive intermediate punishments relative to jail and
prison sentences.
Finally, intermediate sanctions are of additional theoretical import be-

cause they are uniquely localized punishments. From a court community
perspective, different courts are likely to develop norms regarding the
appropriateness of substituting intermediate sanctions for incarceration as
well as the acceptability of engaging in some degree of net-widening. In line
with organizational efficiency theory, courts with greater financial invest-
ment in community-based sanctions should be more likely to exercise these
sentencing options as they entail less constraint on courtroom resources.
To obtain state funding, local courts had to develop and submit proposals
for alternative sanction programs. Additional state monies were set aside
specifically for drug and alcohol programming at the county level. Notably,
total funding for intermediate sanctions and specific support for drug and
alcohol programs both vary considerably across jurisdictions, with some
counties investing significantly more resources than others. This variation
may reflect cultural differences in the value placed on community punish-
ments at the local level, varying emphases on the importance of diverting
offenders to more cost-effective alternatives, or even differences in the
skill and motivation of different courts in obtaining funds. Although it is
not possible to adjudicate among the underlying sources of intercounty
variation in funding for intermediate sanctions, it is possible to examine
their association with the use alternative sanctions as a sentencing option in
individual cases. To do so, we investigate the following predictions:
Hypothesis 7: Offenders sentenced in courts with greater total fund-

ing for intermediate sanctions will be more likely to receive inter-
mediate punishments relative to both probation and jail and prison
sentences.
Hypothesis 8: Offenders sentenced in courts with greater funding for
drug and alcohol programs will be more likely to receive intermediate
punishments relative to both probation and jail and prison sentences.
DATA AND METHODOLOGY

The data for the current study come from the Pennsylvania Commission
on Sentencing (PCS) and represent all felony and misdemeanor criminal
cases sentenced in the state from 1998 through 2000. Pennsylvania judges
are legally mandated to submit guideline forms to the Commission for
every criminal case sentenced. The PCS data include detailed information
on offender and offense characteristics, case-processing factors, and county
of adjudication, as well as crucial information on the judicial use of interme-
diate punishments. The individual case-level data were extended to include
data on individual judges and local court sentencing environments, making

them particularly well suited for examining both individual and contextual
influences in the punishment process. The final data are limited to cases
with requisite sentencing information and to those sentenced under the
1997 Guidelines, resulting in a total of 200,982 cases sentenced over the
3-year period.4
DEPENDENT VARIABLE
The dependent variable measures whether an offender is sentenced to
an intermediate punishment. We capture this outcome with a multinomial
dependent variable separating sentences to probation, intermediate punish-
ments, jail, and prison. As Ostrom, Ostrom, and Kleiman (2004: 61) argued,
“Intermediate sanctions are the greatest challenge to the empirical study
of sentencing” in part because they “are often packaged together to meet
different combinations of offender risk and need.” It is not unusual for
judges to combine different sentencing alternatives in a single sentence. To
capture these diverse variations, we examined the data two ways. First, we
employed a hierarchical coding rule that captured the most serious sentence
type (e.g., any sentence involving a term of imprisonment was coded as
prison). Second, we analyzed the data coding any sentence that involved an
alternative sanction in the intermediate punishment category. In total, 2,979
cases involved an intermediate sanction in conjunction with jail or prison.
The overall findings were similar for both sets of analyses so we report only
the latter.
The probation category includes sentences to fines, restitution, probation,
and time served. The intermediate sanction category includes community-
based restorative sanctions like community service, drug testing, and drug
and alcohol outpatient programs, as well as more restrictive intermediate
punishments such as house arrest, electronic monitoring, halfway house,
drug and alcohol inpatient programs, intensive supervision, boot camp, and
work release. We include the various intermediate punishments in a single
category to capture their overall prevalence and judicial use as a sentencing
4. In total, 5,889 cases (2.9 percent of the original sample) were dropped from the
analysis because data for one or more of the following variables were missing from
each case: race (3,742); age (53); sex (179); offense severity/prior record (157); race
and sex (1,747); race and age (2); race and severity/priors (3); race, age, and sex (5);
and race, age, and severity/priors (1). A total of 3,502 cases were dropped because
requisite information on the final sentence decision was not recorded. Data on
sentencing judges were collected from judicial biographies of Pennsylvania judges,
and data on court contexts were obtained through several sources including the
U.S. Census, Administrative Office of Pennsylvania Courts, and the Pennsylvania
Commission on Crime and Delinquency.
option, although it is important for future research also to examine indi-

vidual treatment and punishment modalities. The use of a four-category
dependent variable also addresses concerns raised by Holleran and Spohn
(2004) that important differences might characterize jail and prison sen-
tences. The intermediate sanctions category serves as the reference group,
allowing us to assess the overall use of alternative sanctions relative to both
incarcerative and probationary sentences.
INDEPENDENT VARIABLES
A broad array of predictors is used to examine the judicial use of inter-
mediate sanctions. Offender characteristics include the age of the offender
at time of sentencing, measured in years; the gender of the offender, with
females the referent; and the race of the offender, coded as Black, Hispanic,
and other race, with Whites the reference category. We also include several
legal and case-processing variables. The severity of the current offense
is captured by the offense gravity score (OGS) under the Pennsylvania
guidelines, which ranks the crime seriousness from 1 to 14. The prior record
of the offender is captured with the prior record score (PRS) under the
guidelines, which contains seven categories ranking prior offending from
least to most severe (see table S1 in the online supporting information).
We also include an interaction between offense severity and prior record
to account for possible nonlinearities in the effects of the guidelines struc-
ture (Ulmer, 2000).5 The type of offense is further controlled by using
a series of dummy variables for violent, property, drug possession, drug
trafficking, driving under the influence (DUI), and other offenses. Violent
crimes serve as the reference category. Finally, cases involving mandatory
minimum penalties are controlled with an additional dummy variable.6 To
5. In the current analysis, this strategy is preferred over alternatives such as inclusion
of the presumptive sentence (Engen and Gainey, 2000) because initial analyses
were conducted within subsets of the sentencing guidelines where the presump-
tive sentence would have been uniform. For instance, Level 2 of the guidelines
recommends a minimum sentence of restorative sanctions (RS), which involves
no term of incarceration (see the online supporting information).
6. One reviewer questioned whether cases that involve mandatory minimums should
be included in the analysis. We do so because more than 7,000 cases involved the
coincidence of both a mandatory minimum sentence and an intermediate punish-
ment, indicating that the two are not mutually exclusive in these data. Although
this approach may seem counterintuitive, it makes sense under the Pennsylvania
guidelines for two reasons. First, because the guidelines govern misdemeanors
in addition to felonies, some mandatory minimums involve nonincarcerative
mandatory penalties (such as minimum fines) that are often combined with other
sanctions, including intermediate punishments. Second, because the Pennsylvania
guidelines define intermediate sanctions as equivalent and interchangeable with
capture case-processing effects, the mode of conviction also is included,

which compares non-negotiated pleas with negotiated pleas and with trials.
Because 12.5 percent of cases were missing information on the mode of
conviction, we also include a dummy variable for missing disposition type in
the analysis, which allows us to retain these cases without biasing the other
coefficients. Lastly, the sentencing year is included to control for possible
time trends, with 1999 the omitted reference year.
In addition to individual-level predictors, select judicial characteristics
also are included as control variables. These characteristics include the
gender and race of the judge. Female judges are coded 1, and male judges
are coded 0. Racial and ethnic minority judges are coded 1, and White
judges are coded 0. Ideally, further racial distinctions would be used, but
this level of detail was not available. The age of the judge at the time of
sentencing is captured with a continuous variable in years, as is the judge’s
tenure on the bench, which represents the number of prior years of judicial
experience serving in Pennsylvania courts. Together, these represent the
most commonly examined judge characteristics in prior work (Zatz, 2000).
Several measures of county court context also are examined.7 Court size
is measured as the number of sentencing judges in the county. Caseload
pressure is captured by dividing the total number of cases in a county by
the number of sentencing judges. This ratio is subsequently divided by 100
for ease of interpretability. We also include a measure of the overall de-
parture rate of the court because prior research has suggested this captures
deference to the guidelines as well as to localized sentencing standards and
alternative mechanisms for subverting incarcerative sentences (Johnson,
2006). Data on funding for intermediate sanction programs were graciously
provided by the Pennsylvania Commission on Crime and Delinquency
(PCCD). The state initially provided approximately $10 million to support
the introduction of restrictive intermediate punishment options in Penn-
sylvania counties. For counties to be eligible for this funding, they had to
develop an intermediate punishment program plan that would evaluate cor-
rectional resource needs and establish concrete goals for the implementa-
tion of alternative punishment options. Between 1997 and 2000, all but four
judicial districts (Mifflin, McKean, Juniata/Perry, and Huntingdon) were
jail and prison terms, mandatory minimums that require some minimum term of
confinement still allow for the substitution of an intermediate punishment in place
of some or all of the term of incarceration, as long as the total term of punishment
meets the requirement under the mandatory provision.
7. Details regarding the formal specification of the level 3 model, along with sup-
plemental results for an alternative specification, including jail capacity, an ur-
ban court scale, and a conservatism scale, are available in the online supporting
information.
granted state monies for this purpose. Approximately one third of counties
also received separate funds specifically for drug and alcohol programs.
Funding for intermediate sanctions is captured with two variables. The first
is the mean dollars per year for intermediate sanctions, which is divided by
the mean number of cases sentenced per year to provide a relative indicator
of the dollars per case allocated to counties for intermediate punishments.
This ratio is standardized so that a unit represents a 1 standard deviation
change. The second is the number of years that a county received specific
funds earmarked for drug and alcohol programming. Some counties did not
receive any drug and alcohol funding, whereas others received drug and
alcohol funds in multiple years during the study period.8 A summary of all
variables is provided in table 1.
The political and socioeconomic characteristics of the surrounding com-
munity also are incorporated as additional county-level measures. Socioe-
conomic status is captured with a scale combining inverse median income,
the unemployment rate, and the percentage of the population living below
the poverty level (α = .89). Higher scores on this scale represent greater
disadvantage in the county. Our measure of conservatism is the percent-
age voting for the Republican candidate in the 2000 Presidential election.
Counties with higher scores are politically more conservative.
ANALYTIC APPROACH
To test our hypotheses, several analyses are conducted using three-level
hierarchical generalized linear models (HGLMs). This approach simultane-
ously estimates the effects of individual-, judge-, and court-level predictors.
The basic model we estimate is the hierarchical multinomial model summa-
rized as follows:
φmijk
Level 1 ηmijk = log = π0jk(m) + πijk(m) (Zijk )
φMijk

Level 2 π0jk(m) = β00k(m) + βjk(m) (Xjk ) + r0jk(m)

πijk(m) = β10k(m) + r1jk(m) (1)

Level 3 β00k(m) = γ000(m) + γk(m) (Wk ) + u00k(m)
β10k(m) = γ100(m) + u01k(m)
8. Data on exact dollar amounts for additional funds received specifically for drug
and alcohol programs for all years were not available, necessitating the use of this
yearly measure, which is preferable to a simple funding dichotomy because several
counties received funds in multiple years. This measure includes data on funding
in 1997 because funding in the previous year is likely to affect use of intermediate
sanctions in the following year.
18
Table 1. Summary of Dependent and Independent Variables, PCS Data 1998–2000

Variables Coding Scheme Description
Dependent Variable
Sentencing Outcome Multinomial Variable Four-Category Variable for Prison, Jail, Intermediate Sanction, and Probation
Independent Variables
Individual-Level Variables
Demographic Variables
Offender age Years Age of Offender at Time of Sentencing
Offender gender 1 = Male Dummy Indicator for Male Offenders
Offender race 3 Dummy Variables Dummy Indicators for Black, Hispanic, and Other Offenders with White the Omitted Category
Legal/Case Processing Variables
Sentence year 2 Dummy Variables Dummy Indicators for Sentence Year, with 1999 the Omitted Category
Offense gravity PCS Scale Pennsylvania Commission on Sentencing Scale Rating Offense Gravity from 1 to 14
Criminal history PCS Scale Pennsylvania Commission on Sentencing Scale Rating Criminal History from 0 to 8
Mandatory sentence 1 = Mandatory Dummy Indicator for Mandatory Sentence Applied
Offense type 5 Dummy Variables Dummy Indicators for Violent, Property, Drug Possession, Trafficking, DUIs, and “Other” Offenses
Mode of conviction 3 Dummy Variables Dummy Indicators for Negotiated Plea and Trial with Non-Negotiated Plea the Omitted Category
Judge-Level Variables
Judge’s Gender 1 = Female Dummy Indicator for Female Judges
Judge’s Race 1 = Non-White Dummy Indicator for Racial or Ethnic Minority Judges
Judge’s Age Years Age of Judge at Time of Sentencing
Years on Bench Years Number of Years of Prior Experience Serving as a Judge in Pennsylvania Courts
County-Level Variables
JOHNSON & DIPIETRO
Caseload Pressure Ratio Number of Criminal Cases Sentenced/Average Number of Authorized Judgeships in County/100
Departure Rate Percentage Percent of Cases Sentenced Outside Recommended Guidelines Range
Court Size Count Number of Authorized Judgeships in County
Total Funding Dollars Standardized Value Standardized Value for Mean Funding Dollars for Intermediate Sanctions/Mean Number of Cases
Drug and Alcohol Funding Count Number of Years County Received State Funds for Drug and Alcohol Programs, 1997 to 2000
Percent Republican Percentage Percent of County Population Voting Republican in 2000 Election
SES Disadvantage Scale Standardized Scale Combining Inverse Median Income, Unemployment Rate, and% Below Poverty
ABBREVIATIONS: DUI = driving under the influence; SES = socioeconomic status.
The outcome (ηmijk ) represents the log odds of membership, m, in one of

the M categories of the dependent variable for the ith offender, sentenced
by the jth judge in the kth court. The m subscripts indicate terms that are
specific to the comparison between the reference category, intermediate
sanctions, and one of the other categories of the dependent variable. The
model therefore estimates the probability of receiving an intermediate
sanction relative to 1) prison, 2) jail, and 3) probation. The effects of
the vector of individual-level predictors (Z ijk ) are given by the vector of
(π ijk(m) ) coefficients. The model intercept (π0jk(m) ) represents the expected
likelihood of an alternative sanction for the “average” offender sentenced
by the “average” judge in the “average” court. This intercept is subse-
quently modeled at level 2 of the analysis with a vector of judge covariates
(X jk ), and at level 3 with a vector of county court covariates (W k ). The
effects of the former are given by the judge-level (β jk(m) ) coefficients and
the latter by the court-level (γ k(m) ) coefficients. Unique error terms are
included to capture the between-judge variance (r0jk(m) ) and the between-
district variance (u00k(m) ) as well as the variation in the effects of individual
predictors across judges (r1jk(m) ) and court contexts (u01k(m) ).9 All variables
are grand-mean centered.10
One advantage of the hierarchical model is that it provides an explicit
assessment of interjudge and intercounty variation in the use of intermedi-
ate punishments. Moreover, because criminal cases cluster within judges
and within county courts, residual errors are likely to be correlated, vi-
olating a key assumption of ordinary least-squares (OLS) regression and
producing misestimated standard errors. The multilevel model surmounts
9. Hierarchical models allow the researcher to specify individual-level predictors

as either fixed or random effects. Fixed effects assume that the influence of the
individual-level predictor is uniform, whereas random effects allow their influence
to vary across level 2 and/or level 3 units. Random effects were specified where
theoretically relevant and then empirically tested to ensure they demonstrated
significant variation. The final model included random coefficients for offender
age, gender, race, and ethnicity, as well as mode of conviction and type of crime.
These effects were allowed to vary randomly across both judges and courts. In our
tables, we refer to these models as “mixed models” because they include a mixture
of both “randomly varying” and “fixed” level 1 coefficients.
10. Given our substantive interest in the effects of court context measures, we employ
grand-mean centering, which is well suited for examining higher order relation-
ships while controlling for individual differences in level 1 covariates (Enders and
Tofighi, 2007). However, as Raudenbush and Bryk (2002: 138) noted, using grand-
mean centering can introduce bias into level 1 estimates because these coefficients
capture both within- and between-group associations between the individual-level
predictor and the outcome. To ensure that our conclusions for individual-level
effects were not driven by our centering convention, we reestimated our level 1
model using group-mean centering. The pattern of results for this analysis repli-
cates our reported findings and helps ensure that our conclusions for level 1
predictors are not the result of bias introduced by our centering convention.
this problem by estimating a unique random effect for each organizational

unit. Statistical significance tests also are adjusted to the correct degrees of
freedom at each level of analysis. These and other advantages are discussed
in detail in prior sentencing research, so we do not elaborate on them
in further detail here (e.g., Britt, 2000; Johnson, 2006, 2010; Wooldredge,
2007). Our analysis begins with a brief discussion of descriptive findings
before turning to the full results of our multilevel models.
FINDINGS
DESCRIPTIVE STATISTICS
Table 2 presents summary statistics for our independent variables. Ju-
dicial use of intermediate punishments in Pennsylvania is relatively rare.
Survey data from Wooldredge and Gordon (1997) estimated that judges use
incarceration alternatives approximately 42 percent of the time; however,
this work included probation in its estimate. Our results indicate that a
total of 18,766 offenders, or approximately 9.3 percent of the sample, was
sentenced to an intermediate punishment. Intermediate sanctions were
given for 11.6 percent of cases in level 1 of the guidelines matrix, for
6.4 percent of cases in level 2, 8.3 percent of cases in level 3, 6.2 percent
of cases in level 4, and for less than 1 percent of cases in level 5 (see table S1
in the online supporting information). Because intermediate sanctions are
rarely used and not explicitly provided for in level 5 of the guidelines, these
cells of the guidelines are not included in subsequent analyses.
Most offenders have relatively minor criminal histories, and the average
offense severity score is 3.5 on the 14-point scale. White offenders represent
almost two thirds of all offenders with Blacks and Hispanics composing
28 percent and 6 percent of the sample, respectively. Not surprisingly, more
than four out of five offenders are male and the average age is slightly older
than 30 years. The largest proportion of offenders commits property crimes
and less than 5 percent are convicted at trial. Examination of judicial control
variables indicates that most Pennsylvania judges are male and White—
only 15 percent of judges are female, and only 7 percent are racial and
ethnic minorities. On average, judges are approximately 56 years old and
have served around 12 years on the bench.
Pennsylvania courts vary considerably in their organizational and demo-
graphic constitutions. Whereas average county courts have approximately
7 judges, some districts only have 1 judge and others exceed 70 judges.
Average caseloads vary between 47 cases per judge and 359 cases per
judge. Some courts almost never deviate from the sentencing guidelines,
with departure rates of less than 2 percent, whereas others depart nearly a
quarter of the time. Some districts received no state funds for intermediate
sanction programs, whereas others averaged more than $100 per case. Drug
Table 2. Descriptive Statistics for Pennsylvania Commission

on Sentencing Data, 1998–2000
Variables n Mean SD Minimum Maximum
Multinomial Dependent Variables
Prison 21,582 .11
Jail 71,958 .36
Intermediate Sanction 18,766 .09
Probation 88,676 .44
Individual-Level Variables
Age 31.40 10.25 12 99
Male .83 .37 0 1
Black .28 .45 0 1
Hispanic .06 .24 0 1
Other/Unknown Race .01 .08 0 1
Year 1998 .28 .45 0 1
Year 2000 .38 .48 0 1
Offense Gravity 3.46 2.16 1 14
Criminal History 1.28 1.83 0 8
Mandatory Sentence .24 .42 0 1
Violent .14 .35 0 1
Property .23 .42 0 1
Drug Possession .09 .28 0 1
Drug Trafficking .11 .31 0 1
DUI .22 .42 0 1
Negotiated Plea .64 .48 0 1
Trial .04 .20 0 1
Judge-Level Variables
Female Judge .15 .36 0 1
Minority Judge .07 .26 0 1
Age of Judge 55.97 6.58 42 75
Years on Bench 11.91 8.45 0 42
County-Level Variables
Caseload Pressure 1.70 .71 .47 3.59
Departure Rate 7.38 4.38 1.2 24.6
Court Size 6.67 10.27 1 76
Total Funding Dollars .00 1.00 −1.48 2.65
Drug and Alcohol Funding 1.02 1.60 0 4
Percent Republican 44.46 8.12 16.0 59.8
SES Disadvantage .00 2.72 −6.8 5.0
NOTES: Level 1 N = 200,982; Level 2 N = 303; Level 3 N = 60.
ABBREVIATIONS: DUI = driving under the influence; SD = standard deviation; SES =
socioeconomic status.
and alcohol funding varies similarly, with some courts receiving funds in
each year of the study and others receiving no funds. The local political
and economic environment of the courts also is diverse. For instance, some
courts are located in predominantly Republican areas, whereas others are
more than 80 percent Democratic. Table 3 reports the correlation matrices
for judge and county court measures and suggests that multicollinearity is
unlikely to be a problem in our analysis.
Table 3. Correlation Matrices for Judge- and County-Level

Variables, PCS Data, 1998–2000
Judge-Level Variables (N = 303) Female Minority Age Years
Female Judge 1.00
Minority Judge .20 1.00
Age of Judge −.33 −.15 1.00
Years on Bench −.22 −.13 .40 1.00
County-Level Variables (N = 60) Caseload Departure Size Funding D&A Repub SES
Caseload Pressure 1.00
Departure Rate .13 1.00
Court Size −.02 .30 1.00
Total Funding Dollars −.20 .04 .08 1.00
Drug and Alcohol Funding .18 .14 .45 .01 1.00
Percent Republican .10 .01 −.49 −.17 −.11 1.00
SES Disadvantage −.37 −.29 −.04 .24 −.23 −.37 1.00
ABBREVIATIONS: D&A = drug and alcohol; Repub = Republican; SES = socioeconomic
status.
MULTILEVEL STATISTICAL MODELS—INDIVIDUAL EFFECTS

Table 4 reports the results of our fully specified, multinomial mixed
models including individual, judge, and county court predictors. Because
level 1 estimates were not substantively altered by the inclusion of level 2
and level 3 predictors, we report only the results of the fully specified three-
level model. Consistent with our first research hypothesis, several offender
characteristics exert significant influences over the use of intermediate pun-
ishments. Male offenders and racial and ethnic minorities are substantially
less likely to receive intermediate punishments relative to both prison and
jail. Specifically, male defendants are .43 times as likely as females to
receive intermediate sanctions in lieu of prison, and .56 times as likely in
lieu of jail. Relative to White defendants, racial and ethnic minorities are
approximately two thirds as likely to receive a sentencing alternative in
place of prison, and they are .69 and .72 times as likely to receive them
as a substitute for jail, respectively.
Not only are male and minority offenders less likely to receive interme-
diate punishments in lieu of prison and jail terms, but they also seem to be
less likely to receive them compared with probation. Relative to probation,
male offenders are .88 times as likely as females to receive an alternative
sanction, whereas Black offenders are .83 times as likely as White offenders.
The coefficient for Hispanic offenders also is negative, although it is only
marginally significant (p = .09). Taken together, this pattern of results sug-
gests that certain offenders are less likely to receive intermediate sanctions
relative to either incarceration or probation sentences—a key finding that
we discuss in further detail in the subsequent discussion.
Table 4. Multinomial HGLM Mixed Models of Intermediate Sanctions versus Prison, Jail, and
Probation
Inter. Sanction vs. Prison Inter. Sanction vs. Jail Inter. Sanction vs. Probation
b SE Exp(b ) b SE Exp(b ) b SE Exp(b )

Constant .21 .17 – 2.63 .17 –∗∗∗ 1.93 .17 –∗∗∗
Individual-Level Variables (N = 191,659)
Age .02 .00 1.02∗∗∗ .02 .00 1.02∗∗∗ .01 .00 1.01∗∗
Male −.84 .09 .43∗∗∗ −.58 .06 .56∗∗∗ −.13 .06 .88∗
Black −.48 .09 .62∗∗∗ −.37 .06 .69∗∗∗ −.18 .06 .83∗∗
Hispanic −.39 .11 .68∗∗∗ −.33 .11 .72∗∗ −.23 .13 .79
Other/Unknown Race −.08 .19 .92 −.02 .14 .98 −.16 .15 .85
Year 1998 .05 .04 1.05 −.01 .03 .99 .09 .03 1.10∗∗
Year 2000 .02 .03 1.02 .06 .02 1.06∗ −.05 .03 .95∗
Offense Gravity −.71 .01 .49∗∗∗ −.21 .01 .81∗∗∗ .34 .01 1.41∗∗∗
Criminal History −.72 .02 .48∗∗∗ −.37 .01 .69∗∗∗ .13 .02 1.14∗∗∗
Gravity × History .04 .00 1.04∗∗∗ .06 .00 1.06∗∗∗ .05 .00 1.05∗∗∗
Mandatory Sentence −3.62 .24 .03∗∗∗ −.85 .25 .43∗∗∗ 1.77 .33 5.89∗∗∗
Property −.07 .10 .93 .12 .08 1.13 −.03 .08 .97
Drug Possession .81 .16 2.26∗∗∗ .32 .13 1.37∗ −.19 .14 .83
Drug Trafficking .18 .17 1.19 .47 .15 1.60∗∗ .91 .16 2.48∗∗∗
DUI 4.14 .36 62.80∗∗∗ .43 .36 1.54 2.56 .42 12.91∗∗∗
Other Offense .07 .06 1.07 .17 .05 1.18∗∗∗ −.04 .05 .96
Negotiated Plea .16 .15 1.17 .17 .13 1.19 −.05 .12 .95
Trial −1.72 .23 .18∗∗∗ −.57 .20 .57∗∗ .06 .21 1.06
Judge-Level Variables (N = 303)
Female Judge .34 .18 1.40 .17 .18 1.19 .14 .19 1.15
Minority Judge .66 .34 1.94∗ .11 .34 1.11 −.05 .34 .95
Age of Judge .00 .01 1.00 .00 .01 1.00 −.01 .01 .99
Years on Bench .01 .01 1.01 .00 .01 1.00 .00 .01 1.00
THE POWER OF DIVERSION
County-Level Variables (N = 60)

Caseload Pressure −.03 .20 .97 −.27 .19 .77 −.67 .20 .51∗∗
Departure Rate .14 .03 1.15∗∗∗ .17 .03 1.18∗∗∗ .15 .03 1.17∗∗∗
Court Size −.05 .01 .95∗∗∗ −.07 .01 .94∗∗∗ −.10 .01 .91∗∗∗
Total Funding Dollars .21 .15 1.23 .31 .14 1.36∗ .43 .14 1.53∗∗
Drug and Alcohol Funding .33 .08 1.39∗∗∗ .37 .08 1.44∗∗∗ .33 .08 1.39∗∗∗
Percent Republican −.04 .02 .96 −.03 .02 .97 −.05 .02 .95∗
SES Disadvantage −.12 .06 .89∗ −.03 .06 .97 −.01 .06 .99
Between-Judge R2 46% 31% 32%
Between-County R2 28% 50% 41%
23
ABBREVIATIONS: DUI = driving under the influence; Inter. = intermediate; SE = standard error; SES = socioeconomic status.
∗
p < .05; ∗∗ p < .01; ∗∗∗ p < .001.
In addition to offender characteristics, several legal considerations also

affect the judicial use of intermediate sanctions. Offenders who commit
more serious offenses and have longer prior records are less likely to
receive intermediate sanctions relative to jail and prison but more likely
to receive them relative to probation. For instance, each increase in of-
fense severity reduces the odds of receiving an intermediate punishment
compared with prison by approximately 50 percent. This effect is consis-
tent with the argument that intermediate punishments are considered less
appropriate for serious and recalcitrant offenders and as offering greater
supervision and control than traditional probation (Morris and Tonry,
1990). Conviction under a mandatory minimum also exerts dramatic effects,
virtually precluding the use of alternatives in place of prison and increasing
the odds of their use relative to probation by a factor greater than 5.
In line with our third hypothesis, offenders convicted of drug and alcohol
offenses tend to be more likely than violent offenders to be diverted from
jail and prison to intermediate sanctions. Drug possession increases the
odds of prison diversion by a factor of 2.26. The effect for DUI cases is
enormous and reflects the fact that although sentencing alternatives often
are used in DUI cases, prison is very rarely used in these cases. Drug
offenses also are more likely to receive sentencing alternatives relative to
jail; for possession and trafficking offenses, the odds of diversion are 1.37
and 1.6 times greater than for violent offenses, respectively. DUI offenders
also are more than 10 times as likely to receive alternatives relative to
probation, which in part reflects the fact that these cases routinely involve
some period of time in a treatment program. Finally, in support of our
fourth hypothesis, offenders convicted at trial are approximately one fifth
as likely to be diverted from prison to intermediate sanctions and they are
.57 times less likely to be diverted from jail.
MULTILEVEL STATISTICAL MODELS—CONTEXTUAL EFFECTS

To investigate our contextual hypotheses, we began by first examining
the unconditional model (not reported), which is useful for providing es-
timates of the amount of variation attributable to each level of analysis.
The unconditional judge-level variances were 1.63, 1.36, and 1.38 for prison,
jail, and probation contrasts, respectively. The equivalent unconditional
county-level variances were 1.76, 2.34, and 2.07, respectively. All vari-
ance components were statistically significant (p ≤ .001), indicating that
significant variation exists between both judges and county courts in the
probability of receiving an intermediate punishment relative to prison,
jail, and probation. Moreover, comparison of the relative magnitude of
the variance components for judges and counties suggests that more of the
total variation in the outcome occurs between jurisdictions than between
judges, particularly for jail and probation contrasts where the county-level
variance is approximately 70 percent larger for jail and 50 percent larger for
probation. This finding indicates that variation in the use of intermediate
sanctions is greater across court contexts relative to individual sentencing
judges.
Examination of the judge characteristics in table 4 suggests that few
of the judge-level control variables were significantly related to the use
of intermediate sanctions in these data. The lone exception occurred for
the prison contrast, where minority judges were almost twice as likely as
White judges to divert offenders from prison to intermediate sanctions. The
coefficient for female judges also approached statistical significance for the
prison contrast (p = .065), suggesting female judges also may be more likely
to substitute intermediate sanctions for imprisonment. None of the other
judge-level control variables, however, demonstrated significant effects.
In contrast, several of the county court measures demonstrated signif-
icant effects on the likelihood of receiving an intermediate punishment.
In particular, courts with greater caseload pressure were less likely to use
intermediate sanctions in place of probation; however, this effect was not
significant for jail or prison outcomes. Courts with higher departure rates,
however, were significantly more likely to use sentencing alternatives. Each
1 percent increase in the departure rate multiplied the odds of receiving a
sentencing alternative by 1.15 relative to prison, by 1.18 relative to jail, and
by 1.17 relative to probation. This effect may reflect underlying differences
in court community cultures, with stricter courts more apt to adhere to
sentencing guidelines and less apt to use intermediate punishments. The
effect for court size also was significant for all three contrasts, with each
additional judge reducing the likelihood of an intermediate sanction by a
factor of .95, .94, and .91, respectively.
State funding for alternative sanction programs was significantly related
to the use of intermediate sanctions for the jail and probation contrasts but
not for the prison contrast. Each standard deviation increase in funding
elevated the odds of an intermediate sanction by 36 percent relative to
jail and by 53 percent relative to probation. Funding for drug and alcohol
programs exerted strong influences for all contrasts; each additional year
of funding for drug and alcohol programs increases the odds of an inter-
mediate punishment by 39 percent compared with prison, by 44 percent
compared with jail, and by 39 percent compared with probation. This effect
clearly indicates that increased funding, particularly for drug and alcohol
programs, is associated with greater use of intermediate sanctions. In ad-
dition to translating into prison and jail diversion, though, it also seems to
be associated with some degree of net-widening, or the discretionary use of
intermediate sanctions in lieu of probation sentences.11
11. To investigate further the relationship between drug and alcohol funding and
jail and prison diversion specifically for drug and alcohol offenders, we also
Also, partial support was given for the broader impact of surrounding
community-level influences. Percent Republican was negatively but mod-
estly associated with the use of intermediate sanctions relative to probation,
and it approached statistical significance for the prison contrast as well (p =
.053). Socioeconomic disadvantage also was associated with the use of inter-
mediate sanctions relative to prison, reducing the odds of prison diversion
by a factor of .89. Collectively, the model explained between 31 percent and
46 percent of the between-judge variation across contrasts, and it explained
between 28 percent and 50 percent of the between-county variation. It is
perhaps not surprising that the lowest explained variance for court contexts
occurred for prison given they are state not local institutions.
DISCUSSION
In his critique of contemporary sentencing research, Mears (1998: 672)
argued that “[r]arely do sentencing models consider the existence of mul-
tiple sentencing outcomes or the contexts within which decisions among
available outcomes are made” (emphasis added). He went on to point out
that “with the advent of various . . . alternatives to incarceration . . . the need
to redress this situation will become all the more critical in coming years”
(Mears, 1998: 672). The current study attempts to answer the call for more
research on sentencing alternatives and on social contexts by analyzing
the judicial use of intermediate sanctions across county courts. Such an
endeavor is important because, as Gainey, Steen, and Engen (2005: 489)
argued, contemporary sentencing research has “virtually ignored factors
affecting the application of qualitatively different kinds of sentences like
alternative or intermediate sanctions.”
The current results suggest that, in line with hypothesis 1, Black or
Hispanic and male offenders are significantly less likely to be diverted
from prison and jail to intermediate punishments. These results are con-
sistent with theoretical perspectives that emphasize increased perceptions
of culpability and dangerousness for young, male, and minority offenders
(Albonetti, 1991; Steffensmeier, Ulmer, and Kramer, 1998; Ulmer and
Johnson, 2004) and they echo the concerns of Morris and Tonry (1990: 7)
“that these ‘lesser’ punishments” may be “applied to privileged criminals
estimated supplemental models including cross-level interaction terms between

drug offenders and drug and alcohol funding. For prison diversion, this interaction
was highly significant for drug trafficking (b = .40; standard error [SE] = .07),
for jail diversion it was significant for both drug possession (b = .12; SE = .05)
and drug trafficking (b = .39; SE = .06), and for probation it was significant again
for drug trafficking (b = .38; SE = .07). This finding suggests that the odds of a drug
offender being diverted to intermediate sanctions is specifically tied to funding for
drug programs at the county level, especially for those defendants convicted of
trafficking offenses.
while the underclass, particularly black and Hispanic underclass, [go] to

prison.”
Although these findings are important, they only tell part of the story.
One reason that intermediate sanctions are theoretically important is be-
cause they tend to emphasize more substantive, individualized sentencing
goals, like rehabilitation, which encourage consideration of individual of-
fender qualities such as amenability to treatment, offender remorse, and
willingness to participate in alternative programming (Engen et al., 2003).
If observed racial and gender disparities in jail and prison diversion result
solely from judicial bias against these offenders, then one also should expect
them to be more likely to receive intermediate sanctions compared with
probation. However, this is not what our results indicate. As suggested
in hypothesis 2, we find that male and Black offenders are uniquely less
likely to receive intermediate sanctions compared with all other sentencing
options—not only are they less likely to be diverted from jail or prison,
but also they are less likely to receive alternative sanctions relative to
probation. The direction of the probation coefficient for Hispanics also was
consistent with this pattern of findings, although marginally significant.
One possible explanation for this pattern of results is that male and mi-
nority offenders may receive less favorable assessments of their suitability
for intermediate punishment programs relative to other offenders. Judges
may view these offenders as less amenable to treatment, less likely to
complete programs successfully, or less deserving of scarce court resources
involving rehabilitative treatment programs. Select qualitative research, for
instance, has suggested that Hispanic offenders may be viewed by court
actors as less suitable targets for rehabilitative alternatives (Gainey, Steen,
and Engen, 2005: 508). In the case of sentencing alternatives, then, judi-
cial social attributions may work to decrease the perceived suitability of
certain offenders for community-based intermediate sanctions. Such an in-
terpretation, however, remains exploratory in nature. Like most sentencing
research, direct measures of judicial attributions and stereotypes are not
available in the current work. An essential priority for future research, then,
is to develop better measures of these underlying theoretical constructs,
incorporating more specific information on judicial perceptions of offender
culpability, dangerousness, and amenability to treatment into large-scale
statistical analyses. Without more detailed theoretical measures, scholarly
interpretation of individual sentencing disparity will remain at least par-
tially a speculative enterprise.
A second and complementary explanation for why male and minority
offenders may be less likely to receive intermediate punishments is because
these sentencing options typically involve a unique degree of offender
agency. Outside of specific exceptions, such as compulsory drug treatment,
most intermediate punishments require at least tacit compliance on the part
of the offender, and often, they involve the active participation of the of-
fender in the sentencing process. Willingness to participate in intermediate
punishment programs, or in some cases ability to pay, can serve as an impor-
tant prerequisite for the judicial use of alternative sentences. Unlike more
traditional punishments, then, the offender can exert important influence
over the use of intermediate sanctions in criminal courts. As such, we agree
with Gainey, Steen, and Engen (2005: 516) that “[b]ringing the offender
into the equation, at least for studying alternative sanctions, would appear
to make a great deal of sense.” Further theoretical development and future
empirical research is needed to develop the unique role that individual
offender agency plays in the use of intermediate punishments across court
contexts.
Importantly, prior research has demonstrated that willingness to partic-
ipate in intermediate sanctions is directly tied to racial and gender dif-
ferences in their perceived severity. Men perceive sentencing alternatives
to be more severe than women and express less willingness to participate
in them (May et al., 2005; Wood and Grasmick, 1999) as do racial and
ethnic minorities relative to White offenders (May et al., 2005; Wood
and May, 2003). These differences in part reflect differing views about
the supervision level, probability of revocation, and burden of required
programming associated with intermediate sanctions (Gainey, Steen, and
Engen, 2005). Variation in willingness to participate, then, may well be
associated with the observed disparities reported in our study. The fact that
male and minority defendants tend to be less likely to receive alternatives
relative to both incarceration and probation is consistent with a stronger
role for offender agency and willingness to participate than for uniform
judicial bias in the use of these punishments. Future work is clearly needed,
though, that specifically begins to parcel out the independent effects of
potential judicial bias from individual offender motivation and agency. To
accomplish this goal, direct measures of offender willingness to participate
will need to be integrated into analyses of judicial decision making. As
such, the extent to which individual offenders have meaningful voices in
the judicial application of alternatives certainly deserves the attention of
future research.
As suggested by our third hypothesis, the use of intermediate sanctions
also varies by offense. Compared with violent offenders, defendants con-
victed of drug or alcohol crimes often were more likely to be diverted
from prison and jail to intermediate sanctions. This result may reflect the
fact that drug offenses often involve relatively severe punishments that
court actors may be inclined to “correct” through the use of intermediate
sanctions (Gainey, Steen, and Engen, 2005; Kramer and Ulmer, 2002). In
this way, sentencing alternatives may provide an important mechanism
for subverting punishments that are considered disproportionately severe
under the guidelines. The largest effects, though, occurred for DUI of-
fenses, which were very unlikely to receive any term of imprisonment and
were substantially more likely to receive intermediate sanctions compared
with probation. This finding reflects the fact that these offenses very rarely
result in imprisonment but very often involve short jail sentences in combi-
nation with time in community-based alcohol abuse programs.
We found support for our fourth hypothesis regarding the effects of mode
of conviction. Some prior work has suggested that court actors believe
sentencing alternatives may “serve the critical function of maintaining court
efficiency by keeping cases moving through the system” (Gainey, Steen, and
Engen, 2005: 515). Our results provide support for that assertion. Going to
trial had profound effects on the likelihood of both prison and jail diversion
but not on probation outcomes. In part, this finding may reflect the fact
that trials typically occur in more serious cases where prison or jail is the
presumptive sentence. In these cases, it is likely that guilty pleas signal
greater offender remorse, which may be a moral prerequisite for diversion
to sentencing alternatives that place greater emphasis on the rehabilitative
goals of sentencing.
Several of our research hypotheses also focused on the correlates of
intercourt variation in the use of intermediate punishments. Overall, the
individual likelihood of receiving an alternative sanction varied significantly
across both individual judges and county court contexts as suggested by
hypothesis 5. Judge-level control variables exerted relatively few signifi-
cant effects on the use of intermediate punishments, although a judge’s
race and gender were tied to the likelihood of prison diversion. Ideally,
more proximate measures of judge-level factors, such as judicial sentencing
philosophies and attitudes toward community punishments, will be needed
to improve investigation of judicial variation in intermediate sanctions.
Overall, though, the county-level variance estimates were larger than the
judge-level estimates, suggesting a greater role for court contexts than for
judicial factors in the use of intermediate sanctions.
Although prior work on court contexts emphasizes the importance of
caseload pressure (Ulmer and Johnson, 2004; Wooldredge, 1989), only
partial support emerged for its ability to explain intercourt variations in the
use of intermediate punishments in Pennsylvania. Higher court caseloads
were associated with greater use of probation sentences compared with
intermediate punishments. Intensive caseload pressure may place greater
emphasis on efficient case disposal, discouraging the use of more individu-
alized intermediate sentencing options in these contexts.
More uniform support was found for the expectation that courts with
high departure rates also would employ more liberal use of sentencing
alternatives. Prior theorizing has suggested that lower levels of guidelines
compliance may reflect courtroom cultural norms that emphasize increased
individuality in punishment and/or greater emphasis on rehabilitative ideals

(Johnson, 2006; Kautt, 2002). As such, low guidelines compliance is re-
flective of cultural norms favoring individualization in sentencing, which
is consistent with an increased propensity to divert offenders to interme-
diate punishments. A proper test of these theoretical arguments, how-
ever, requires additional data that include more direct measures of the
normative attitudes of court actors and cultural norms toward community
punishments. Additional research also is needed that contrasts the use of
intermediate punishments with other diversionary sentencing options like
guidelines departures, which may represent complementary or competitive
approaches to the individualization of punishment under sentencing guide-
lines (Engen et al., 2003; Johnson, 2005).
The results for court size were unexpected. Some prior work has found
that larger courts are less punitive (Ulmer and Johnson, 2004), leading us to
predict a positive relationship between court size and the diversionary use
of intermediate sanctions. Not only was this not the case, but our findings
suggest large courts are in fact less likely to use intermediate punishments.
Notably, this result corroborates survey research by Wooldredge and Gor-
don (1997) that found larger, more bureaucratic court organizations offer
less opportunity to individualize punishments via alternative sanctions. One
limitation of the current study is that court size tends to be highly related
to other court community characteristics, including the proportion of the
population that is African American and the county-level crime rate. These
additional factors were too highly related to be simultaneously examined.
Supplemental models using a combined factor for court size, percent Black,
and the violent crime rate produced similar findings, but it is important
to note that the findings for court size may partly reflect the latent influ-
ence of these and related omitted variables. Future research is therefore
needed in additional contexts, incorporating alternative measures of court
structure, such as its level of bureaucratization, in addition to other omitted
community-level influences to parcel out their unique contributions better.
Finally, court community and organizational efficiency perspectives sug-
gest that availability of local resources is likely to be tied to jurisdictional
variation in punishment (Dixon, 1995; Eisenstein and Jacob, 1977). This
study offers the first examination of the impact of local funding measures
on the use of sentencing alternatives. Overall, funding variables were a
strong predictor of the use of intermediate sanctions in individual cases.
Offenders sentenced in counties with more funding for alternative sen-
tences were more likely to receive these outcomes in lieu of prison and
jail time. Levels of state funding, however, also were tied to the degree
of net-widening that occurs in a jurisdiction—courts with greater funding
were more likely to sentence offenders to intermediate sanction programs
instead of probation. The exact theoretical mechanisms responsible for
this relationship remain somewhat ambiguous. It is likely that more fund-

ing increases program availability and encourages increased judicial use
of intermediate sanctions—from an organizational efficiency perspective,
increased resources should translate into greater use of sentencing alter-
natives. However, it also is possible that this relationship operates recip-
rocally. That is, courts with latent propensities to sentence offenders to
community sanctions may be more aggressive in obtaining funding for these
programs. Qualitative research with court actors that examines the process
of how and why some county courts obtain greater funding would be needed
to sort out this causal ordering.
One unique finding of this work is that substantial net-widening is associ-
ated with intermediate sanctions in Pennsylvania. Relatively little empirical
research specifically has examined this issue in the context of sentencing
guidelines. Tonry and Lynch (1996: 101) suggested that “the availability
of new sanctions presents almost irresistible temptations to judges and
corrections officials to use them for offenders other than those for whom
the program was created.” In the current study, the largest proportion
of intermediate punishments occurred in level 1 of the guidelines, where
probation and restorative sanctions (i.e., fines or restitution) are the only
recommended sentences. Nearly 12 percent of these offenders received
intermediate sanctions, compared with 7 percent in levels 3 and 4 of the
guidelines where alternatives can be exchanged for jail and prison terms.
This finding suggests that judges use intermediate punishments for both
presumptive probationers and presumptive incarcerates.
Despite the negative rhetoric surrounding net-widening, though, we stop
short of deriding the process as necessarily undesirable. One benefit of
intermediate punishments is that they expand the continuum of punish-
ments to better serve goals like proportionality, community protection, and
individual just deserts. Some offenders who would have otherwise received
probation may be more effectively punished with community sanctions,
just as some incarcerated offenders will be more appropriately punished
through diversion. From the perspective of the judge, the decision to divert
some offenders from probation is understandable, even though it is likely to
conflict with larger programmatic goals like correctional cost effectiveness
and conservation of correctional resources (Tonry, 1997; Tonry and Lynch,
1996). The question of where to draw this line is a moral and philosophical
one that will require thoughtful debate and careful future scrutiny among
academics and justice practitioners alike.
CONCLUSION
The process of criminal sentencing has always involved “conflicts be-
tween the goals of equal justice under the law and individualized justice
with punishment tailored to the offender” (Blumstein et al., 1983: 39). As

Richard Frase (2000: 429) pointed out, the use of intermediate sanctions
has rapidly expanded, in part because of “rising prison populations, and
a desire to encourage judges to consider effective non-custodial sentenc-
ing options,” although few guidelines systems adequately “regulate inter-
mediate sanctions.” Sentencing alternatives represent a unique locus of
court actor discretion—they are tailored specifically to individual offender
considerations but lack systematic guidance in application, raising a new
breed of concerns over social inequities in punishment. Although a growing
literature has evolved documenting the successes and failures of specific
intermediate sanctions (e.g., Byrne, Lurigo, and Petersilia, 1996; Gendreau
et al., 2000; Harland, 1993; Langan, 1994; Morris and Tonry, 1990; Padgett,
Bales, and Blomburg, 2006; Petersilia, 1998; Petersilia and Turner, 1993;
Tonry and Lynch, 1996), very little research attention has been devoted to
understanding how and why they are employed as a sentencing option in
criminal courts. This lack of attention is unfortunate given their prolifera-
tion and their potential promise to redress jail and prison crowding in the
face of rising costs of incarceration (Wooldredge and Gordon, 1997).
The current findings highlight the fact that certain classes of offender,
namely, male and minority defendants, have systematically lower odds of
receiving alternative sanctions in place of jail and prison sentences. These
same offenders, however, also tend to be less likely to receive alternatives
in lieu of probation. In part, these differences likely reflect legitimate
sentencing concerns as well as real-world organizational constraints. Indi-
viduals vary in their amenability to intermediate punishments and in their
likelihood of successfully completing alternative sanction programs, both
of which are associated with offender demographic and background char-
acteristics (Wood and Grasmick, 1999; Wood and May, 2003). A complete
test of these theoretical arguments, however, will require future work that
collects additional data on both judicial sentencing rationales and offender
participation in sentencing hearings. Such research should be a top priority
of future work on community-based punishments.
The judicial use of intermediate punishments also varies notably across
individual judges and across county courts. Jurisdictional variations are tied
to differences in organizational structure and variation in available funding
for alternative sanction programs. Prior work has suggested criminal courts
develop localized cultural standards that guide criminal case processing
(Dixon, 1995; Eisenstein and Jacob, 1979; Ulmer, 1997), which likely trans-
late into court-specific policies and standards for how and when to divert
offenders to intermediate sanctions. Whether contextual disparities are
warranted is debatable, but geographic variation in their application in part
reflects the fact that intermediate punishments serve to “promote account-
ability of offenders to their local community” and to “address local needs
through the expansion of punishment” (Sims and Shi, 1999: 54). Given
the diversity in the funding and availability of intermediate punishment
programs across local court environments, some jurisdictional variation is
not surprising and may even be desirable.
One theme that emerges is that there remains a lack of clear structure
and guidance about how to employ intermediate sanctions as substitutes
for other guidelines punishments. Sentencing alternatives present judges
with added discretion that warrants the consideration of individual offender
characteristics. Inevitably, this issue raises legitimate concerns about their
fair and equal application. As Engen et al. (2003: 109) noted, “Sentenc-
ing alternatives are important . . . because they encourage consideration of
sentencing criteria that may make disparities by race-ethnicity or gender
even more likely.” Ultimately, selection and interchangeability issues may
require the development of guidelines policies to define better the relative
weight and priority of different sentencing considerations (Harland, 1993),
although simplicity of application remains an important counterbalance in
this debate (Frase, 2000).
Finally, intermediate sanctions also are unique in that they often allow
offenders to exercise some degree of autonomy in the sentencing process.
Observed disparities, like those in this study, may reflect court actor predis-
positions, stereotypes, or bias; they may be associated with the recommen-
dations and arguments levied by prosecutors and defense counsel; or they
may be intimately tied to individual preferences, willingness to participate,
or ability to pay for available programs. In all likelihood, different individ-
ual cases involve some complex melding of these factors. Future research
is therefore needed that continues to theorize and begins to examine em-
pirically the important and unique role that offenders play in the receipt
of sentencing alternatives under guidelines; through the use of surveys,
interviews, and other innovative data collection techniques like courtroom
observation, researchers can begin to collect and integrate more proximate
measures of important theoretical constructs. Data on judicial attitudes and
sentencing philosophies and assessments of offender blameworthiness and
dangerousness would substantially improve the theoretical specification of
sentencing models as would offender-based measures of attitudes toward
alternative sanction programs and willingness to participate in them.
The current research has some important limitations that future research
should work to overcome. Intermediate sanctions encompass numerous
types of community-based punishments, and the factors that are most
salient may be different for different options (Engen et al., 2003). In-
termediate punishments range from treatment-oriented sanctions, like in-
patient clinics, to more control-oriented punishments, like house arrest, and
they may be combined to meet varying degrees of offender risk and need
(Ostrom, Ostrom, and Kleiman, 2004). The current analysis examines the
judicial use of these sentencing alternatives as a whole, but future research

is needed that begins to disentangle variations in the use of different types
of intermediate punishments for different types of offenders. Because in-
termediate sanctions are likely to play an expanded role in the future of
criminal sentencing, the current findings should be replicated with more
recent data and in additional research contexts. The current data cannot
speak to issues involving the global effects of sentencing alternatives on
such issues as correctional populations, cost effectiveness, prison crowd-
ing, and offender recidivism. A related literature addresses these issues
(e.g., Morris and Tonry, 1990; Petersilia, 1998; Tonry and Lynch, 1996),
but ideally research on the judicial use of sentencing alternatives must be
combined with research on their correctional effectiveness. In the end, we
agree with Frase (2000: 443) that “guidelines systems can and should seek
to develop standards—and resources—that promote increased and more
effective use of intermediate sanctions,” as long as they carefully balance
issues of unwarranted disparity and individualized justice, with issues of
effective prison diversion and concerns over potential net-widening in the
criminal justice system.
REFERENCES
Albonetti, Celesta. 1991. An integration of theories to explain judicial
discretion. Social Problems 38:247–66.
Austin, James, and John Irwin. 2001. It’s About Time: America’s Imprison-
ment Binge. Belmont, CA: Wadsworth.
Blumstein, Alfred, and Allen J. Beck. 1999. Population growth in U.S.

prisons, 1980–1996. Crime and Justice 6:17–61.
Blumstein, Alfred, Jacqueline Cohen, Susan E. Martin, and Michael

H. Tonry. 1983. Research on Sentencing: The Search for Reform.
Washington, DC: National Academy Press.
Britt, Chester L. 2000. Social context and racial disparities in punishment

decisions. Justice Quarterly 17:707–32.
Byrne, Jim M., Arthur J. Lurigo, and Joan Petersilia. 1996. Smart Sen-
tencing: The Emergence of Intermediate Sanctions. Newbury Park, CA:
Sage.
Crimes Code of Pennsylvania. 2006. Lexis Nexus Publishing.
Dixon, Jo. 1995. The organizational context of criminal court sentencing.

American Journal of Sociology 100:1157–98.
Eisenstein, James, Roy B. Flemming, and Peter F. Nardulli. 1988. The

Contours of Justice: Communities and Their Courts. Boston, MA: Little,
Brown.
Eisenstein, James, and Herbert Jacob. 1977. Felony Justice: An Organiza-

tional Analysis of Criminal Courts. Boston, MA: Little, Brown.
Enders, Craig K., and Davood Tofighi. 2007. Centering predictor variables
in cross-sectional multilevel models: A new look at an old issue. Psy-
chological Methods 12:121–38.
Engen, Rodney L., and Randy R. Gainey. 2000. Modeling the effects of
legally relevant and extralegal factors under sentencing guidelines: The
rules have changed. Criminology 38:1207–30.
Engen, Rodney L., Randy R. Gainey, Robert D. Crutchfield, and Joseph G.

Weis. 2003. Discretion and disparity under sentencing guidelines: The
role of departures and structured sentencing alternatives. Criminology
41:99–130.
Frase, Richard. 1991. Defendant amenability to treatment or probation

as a basis for departure under the Minnesota and federal sentencing
guidelines. Federal Sentencing Reporter 3:328–33.
Frase, Richard. 2000. Is guided discretion sufficient? Overview of state

sentencing guidelines. St. Louis Law Review 44:425–49.
Gainey, Randy R., Sara Steen, and Rodney L. Engen. 2005. Exercising
options: An assessment of the use of alternative sanctions for drug
offenders. Justice Quarterly 22:488–520.
Gendreau, Paul, Claire Goggin, Francis T. Cullen, and Don A. Andrews.

2000. Effects of community sanctions and incarceration on recidivism.
Forum on Corrections Research 12:10–3.
Gibson, James, L. 1978. Judges role orientations, attitudes, and deci-

sions: An interactive model. American Political Science Review 72:
911–24.
Gruhl, John, Cassia Spohn, and Susan Welch. 1981. Women as policy-
makers: The case of trial judges. American Journal of Political Science
25:308–22.
Harland, Allen T. 1993. Defining a continuum of sanctions: Some research
and policy implications. Perspectives 17:6–15.
Helms, Ronald, and David Jacobs. 2002. The political context of sentencing:
An analysis of community and individual determinants. Social Forces
81:577–604.
Hogarth, John. 1971. Sentencing as a Human Process. Toronto, Ontario,

Canada: University of Toronto Press.
Holleran, David, and Cassia Spohn. 2004. On the use of the total incarcera-
tion variable in sentencing research. Criminology 42:211–40.
Johnson, Brian D. 2005. Contextual disparities in guidelines departures.

Criminology 43:761–97.
Johnson, Brian D. 2006. The multilevel context of criminal sentenc-

ing: Integrating judge-and county-level influences. Criminology 44:
259–98.
Johnson, Brian D. 2010. Multilevel analysis in the study of crime and

punishment. In the Handbook of Quantitative Criminology, eds. David
Weisburd and Alex Piquero. New York: Springer-Verlag.
Kautt, Paula. 2002. Location, location, location: Interdistrict and intercir-

cuit variation in sentencing outcomes for federal drug-trafficking of-
fenses. Justice Quarterly 19:633–71.
Kramer, John K., and Jeffrey T. Ulmer. 2002. Downward departures for
serious violent offenders: Local court “corrections” to Pennsylvania’s
sentencing guidelines. Criminology 40:897–932.
LaFree, Gary. 1985. Official reactions to Hispanic defendants in the South-

west. Journal of Research in Crime and Delinquency 22:213–37.
Langan, Patrick A. 1994. Between prison and probation: Intermediate sanc-

tions. Science 264:791–3.
MacKenzie, Doris L. 2001. Correction and sentencing in the 21st century:
Evidence-based corrections and sentencing. The Prison Journal 81:299–
312.
Martinson, Robert. 1974. What works? Questions and answers about prison
reform. The Public Interest 35:22–54.
May, Donald, Peter B. Wood, Jennifer L. Mooney, and Kevin I. Mi-

nor. 2005. Predicting offender-generated exchange rates: Implica-
tions for a theory of sentence severity. Crime & Delinquency 51:
373–99.
Mears, Daniel P. 1998. The sociology of sentencing: Reconceptualiz-

ing decision-making processes and outcomes. Law & Society Review
32:667–724.
Mitchell, Ojmarrh. 2005. A meta-analysis of race and sentencing research:

Explaining the inconsistencies. Journal of Quantitative Criminology
21:439–66.
Morris, Norval, and Michael H. Tonry. 1990. Between Prison and Probation:
Intermediate Punishments in a Rational Sentencing System. New York:
Oxford University Press.
Ostrom, Charles W., Brian J. Ostrom, and Matthew Kleiman. 2004. Judges
and discrimination: Assessing the theory and practice of criminal sen-
tencing. Final Grant Report to the National Institute of Justice (Grant
98-CE-VX-0008).
Padgett, Kathy G., William D. Bales, and Thomas G. Blomberg. 2006.

Under surveillance: An empirical test of the effectiveness and con-
sequences of electronic monitoring. Criminology & Public Policy 5:
61–92.
Petersen, Rebecca, and Dennis J. Palumbo. 1997. The social construction

of intermediate punishments. Prison Journal 77:77–91.
Petersilia, Joan. 1998. Community Corrections. New York: Oxford Univer-

sity Press.
Petersilia, Joan, and Susan Turner. 1993. Intensive probation and parole. In
Crime and Justice: A Review of Research, vol. 17., ed. Michael H. Tonry.
Chicago, IL: University of Chicago Press.
Portillos, Edwardo L. 1998. Latinos and Latinas: Latinos, gangs, and drugs.
In In Images of Color, Images of Crime, eds. Coramae. Richey Mann
and Marjorie Zatz. Oxford, U.K.: Oxford University Press.
Raudenbush, Stephen W., and Anthony S. Bryk. 2002. Hierarchical Linear

Models: Applications and Data Analysis Methods, 2nd ed. Thousand
Oaks, CA: Sage.
Ruth, Henry, and Kevin Reitz. 2003. The Challenge of Crime. Cambridge,
MA: Harvard University Press.
Savelsberg, Joachim. 1992. Law that does not fit society: Sentencing guide-
lines as a neoclassical reaction to the dilemmas of substantivized law.
American Journal of Sociology 97:1346–81.
Sims, Barbara, and Wenbo Shi. 1999. Assessment of Pennsylvania interme-

diate punishment programs. Corrections Management Quarterly 3:53–
62.
Spohn, Cassia. 1990. The sentencing decision of Black and White judges:
Some expected and unexpected similarities. Law & Society Review
24:1197–216.
Spohn, Cassia. 2000. Thirty years of sentencing reform: The quest for a
racially neutral sentencing process. Criminal Justice: The National In-
stitute of Justice Journal 3:427–501.
Steffensmeier, Darrell, and Chris Hebert. 1999. Women and men poli-
cymakers: Does the judge’s gender affect the sentencing of criminal
defendants? Social Forces 77:1163–96.
Steffensmeier, Darrell, Jeffery T. Ulmer, and John K. Kramer. 1998. The
interaction of race, gender, and age in criminal sentencing: The punish-
ment cost of being young, black, and male. Criminology 36:763–98.
Tonry, Michael H. 1996. Sentencing Matters. New York: Oxford University

Press.
Tonry, Michael H. 1997. Intermediate Sanctions in Sentencing Guidelines.

Washington, DC: U.S. National Institute of Justice.
Tonry, Michael H. 2004. The Future of Imprisonment. New York: Oxford

University Press.
Tonry, Michael H., and Mary Lynch. 1996. Intermediate sanctions. In Crime
and Justice: A Review of Research, vol. 20, ed. Michael H. Tonry.
Chicago, IL: University of Chicago Press.
Ulmer, Jeffery T. 1997. Social Worlds of Sentencing: Court Communities

under Sentencing Guidelines. Albany: State University of New York
Press.
Ulmer, Jeffery T. 2000. The rules have changed—so proceed with caution:
A comment on Engen and Gainey’s method for modeling sentencing
outcomes under guidelines. Criminology 38:1231–44.
Ulmer, Jeffery T., and Brian D. Johnson. 2004. Sentencing in context: A

multilevel analysis. Criminology 42:137–77.
von Hirsch, Andrew. 1976. Doing Justice: The Choice of Punishments. New
York: Hill and Wang.
Weber, Max. 1968. Economy and Society, eds. Guenther Roth and Claus
Wittich. New York: Bedminster Press.
Welch, Susan, Michael Combs, and John Gruhl. 1988. Do Black judges
make a difference? American Journal of Political Science 32:126–36.
Wellford, Charles F. 2007. Sentencing research for sentencing reform.
Criminology & Public Policy 6:399–402.
Wood, Peter B., and Harold G. Grasmick. 1999. Toward the development
of punishment equivalencies: Male and female inmates rate the severity
of alternative sanctions compared to prison. Justice Quarterly 16:19–50.
Wood, Peter B., and David C. May. 2003. Racial differences in perceptions
of the severity of sanctions: A comparison of prison with alternatives.
Justice Quarterly 20:605–31.
Wooldredge, John. 1989. An aggregate-level examination of the caseload

pressure hypothesis. Journal of Quantitative Criminology 5:259.
Wooldredge, John. 2007. Neighborhood effects on felony sentencing. Jour-

nal of Research in Crime and Delinquency 44:238–63.
Wooldredge, John, and Jill Gordon. 1997. Predicting the estimated use
of alternatives to incarceration. Journal of Quantitative Criminology
13:121–42.
Zatz, Marjorie. 2000. The convergence of race, ethnicity, gender, and class
on Court decision-making: Looking toward the 21st century. Criminal
Justice: The National Institute of Justice Journal 3:503–52.
Brian D. Johnson is an associate professor in the Department of Crim-

inology and Criminal Justice at the University at Maryland. He received
his PhD in crime, law, and justice from Pennsylvania State University in
2003. His research examines various issues in juvenile and criminal justice,
with a particular focus on social inequality and contextual variations in
punishment. Prof. Johnson received the 2008 ASC Ruth Shonle Cavan
Award for distinguished young scholar, and he is a faculty associate of the
Maryland Population Research Center.
Stephanie M. DiPietro is an assistant professor in the Department of

Criminology and Criminal Justice at the University of Missouri–St. Louis.
Her research interests involve immigration and crime, with a focus on immi-
grant youth. She is the co-recipient of the 2011 W.E.B. Du Bois Fellowship
from the National Institute of Justice.
SUPPORTING INFORMATION
The following supporting information is available for this article:
Table S1. 1997 Pennsylvania Sentencing Guidelines

Details regarding the formal specification of Level 3 model
Table S2. Multinomial HLM Mixed Models with Alternative Level 3
Specification
Supporting Information may be found in the online version of this article.
Please note: Wiley-Blackwell is not responsible for the content or func-

tionality of any supporting information supplied by the authors. Any
queries (other than missing material) should be directed to the correspond-
ing author for the article.
View publication stats

Kavita Mam

Uploaded by

Copyright:

Available Formats

Kavita Mam

Uploaded by

Document Information

Original Description:

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

Kavita Mam

Uploaded by

Copyright:

Available Formats

See discussions, stats, and author profiles for this publication at: https://www.researchgate.

The Power of Diversion: Intermediate Sanctions and

Article in Criminology · August 2012

Brian Johnson Stephanie M. Dipietro

SEE PROFILE SEE PROFILE

The user has requested enhancement of the downloaded file.

1. Intermediate sanctions are variously referred to as intermediate punishments,

invokes a less structured and more substantively rational decision-making

punishment. Second, the emergence of “just deserts” emphasized the need

PRIOR RESEARCH ON INTERMEDIATE

Spohn, 2000; Zatz, 2000). Examinations of the judicial use of intermediate

INTERMEDIATE PUNISHMENTS UNDER THE

for jail or prison at the discretion of the sentencing judge.3 Level 5 of

THEORETICAL IMPORT OF INTERMEDIATE

subsequent discussion, examination of these net-widening processes can

SUBSTANTIVE RATIONALITY, ATTRIBUTION, AND FOCAL

Hypothesis 1: Male and minority offenders will be less likely to re-

Similar processes may characterize judicial decisions regarding net-

SUITABILITY OF PUNISHMENT AND INDIVIDUAL OFFENDER

Hypothesis 2: Male and minority offenders will be less likely to re-

The perceived suitability of punishment also is likely to be affected by

to Gainey, Steen, and Engen (2005: 491–3), intermediate sanctions involve

Hypothesis 3: Compared with violent crimes, offenders convicted of

According to organizational efficiency theory, efficient case disposition

Hypothesis 4: Offenders convicted through trials will be less likely to

COURT COMMUNITIES, LOCAL RESOURCES,

Hypothesis 5: The use of intermediate punishments will vary signifi-

Finally, intermediate sanctions are of additional theoretical import be-

Hypothesis 7: Offenders sentenced in courts with greater total fund-

DATA AND METHODOLOGY

data on individual judges and local court sentencing environments, making

option, although it is important for future research also to examine indi-

capture case-processing effects, the mode of conviction also is included,

Table 1. Summary of Dependent and Independent Variables, PCS Data 1998–2000

The outcome (ηmijk ) represents the log odds of membership, m, in one of

9. Hierarchical models allow the researcher to specify individual-level predictors

this problem by estimating a unique random effect for each organizational

Table 2. Descriptive Statistics for Pennsylvania Commission

Table 3. Correlation Matrices for Judge- and County-Level

MULTILEVEL STATISTICAL MODELS—INDIVIDUAL EFFECTS

b SE Exp(b ) b SE Exp(b ) b SE Exp(b )

County-Level Variables (N = 60)

In addition to offender characteristics, several legal considerations also

MULTILEVEL STATISTICAL MODELS—CONTEXTUAL EFFECTS

estimated supplemental models including cross-level interaction terms between

while the underclass, particularly black and Hispanic underclass, [go] to

individuality in punishment and/or greater emphasis on rehabilitative ideals

this relationship remain somewhat ambiguous. It is likely that more fund-

with punishment tailored to the offender” (Blumstein et al., 1983: 39). As

judicial use of these sentencing alternatives as a whole, but future research

Blumstein, Alfred, and Allen J. Beck. 1999. Population growth in U.S.

Blumstein, Alfred, Jacqueline Cohen, Susan E. Martin, and Michael

Britt, Chester L. 2000. Social context and racial disparities in punishment

Crimes Code of Pennsylvania. 2006. Lexis Nexus Publishing.

Dixon, Jo. 1995. The organizational context of criminal court sentencing.

Eisenstein, James, Roy B. Flemming, and Peter F. Nardulli. 1988. The

Eisenstein, James, and Herbert Jacob. 1977. Felony Justice: An Organiza-

Engen, Rodney L., Randy R. Gainey, Robert D. Crutchfield, and Joseph G.