JOURNAL OF ADOLESCENT HEALTH 2001;29:279–288

ORIGINAL ARTICLE

Relationship Between Perceived Life Satisfaction and

Adolescents’ Substance Abuse

KEITH J. ZULLIG, M.S.P.H., ROBERT F. VALOIS, Ph.D., M.P.H., E. SCOTT HUEBNER, Ph.D.,
JOHN E. OELTMANN, M.Ed., M.S.P.H., AND J. WANZER DRANE, Ph.D.

Purpose: To explore the relationship between per- Conclusion: Longitudinal studies are needed to deter-
ceived global life satisfaction and selected substance use mine whether dissatisfaction with life is a consequence
behaviors among 5032 public high school students. or determinant of substance abuse behavior for adoles-
Methods: The 1997 South Carolina Youth Risk Behav- cents. © Society for Adolescent Medicine, 2001
ior Survey substance abuse and life satisfaction variables
were used. An adjusted polychotomous logistic regres-
sion analysis utilizing SAS/SUDAAN, revealed a signif- KEY WORDS:
icant race/gender interaction. Subsequent multivariate Adolescents
models were constructed individually for four race/gen- Substance abuse
der groups. Adjusted odds ratios and 95% confidence Perceived life satisfaction
intervals were calculated to assess the magnitude of risk Quality of life
for selected substance abuse behaviors and their associ- Gender differences
ation with reduced global life satisfaction. Ethnic differences
Results: Cigarette smoking, chewing tobacco, mari-
juana, cocaine, regular alcohol use, binge drinking, in-
jection drug, and steroid use were significantly (p < .05)
associated with reduced life satisfaction for specific race/ Enhancing the quality of life (QOL) is the corner-
gender groups (white males; black males; white females; stone of health promotion efforts that seek to em-
and black females). In addition, age (<13 years) of first power people to improve their overall health [1,2].
alcohol drink, first marijuana use, first cocaine use, and Research bridging these two areas has been scarce,
first cigarette smoked were also significantly (p < .05) although some particularly noteworthy efforts have
associated with reduced life satisfaction. appeared recently [3]. This article considers the role
of one aspect of QOL, life satisfaction, and its rela-
tionship with adolescent substance abuse.
QOL research has been conceptualized from two
From the Department of Health Promotion & Education, School of perspectives: objective and subjective. The objective
Public Health, University of South Carolina, Columbia, South Carolina
(K.J.Z.); the Department of Health Promotion & Education, Family & perspective focuses on external conditions that con-
Preventive Medicine, Schools of Public Health and Medicine, University tribute to QOL, such as income levels, housing
of South Carolina, Columbia, South Carolina (R.F.V.); the Department quality, friendship networks, and access to health
of Psychology, University of South Carolina, Columbia, South Carolina
(E.S.H.); the Department of Epidemiology & Biostatistics, School of services. In contrast, subjective QOL refers to indi-
Public Health, University of South Carolina, Columbia, South Carolina viduals’ internal judgments of their quality of their
(J.E.O.); and the Department of Epidemiology & Biostatistics, Family & overall lives and/or specific life domains (e.g. satis-
Preventive Medicine, Schools of Public Health and Medicine, University
of South Carolina, Columbia, South Carolina (J.W.D.). faction with friends, family, school experiences) [4,5].
Address correspondence to: Robert F. Valois, Ph.D., M.P.H., Health In the development of their QOL model, Day and
Promotion & Education, School of Public Health, 800 Sumter Street, Jankey [6] concluded that although objective life
Room 216, University of South Carolina, Columbia, SC 29208. E-mail:
RValois@sph.sc.edu. circumstances are important in individuals determi-
Manuscript accepted March 29, 2001. nations of the quality of their lives, it is their subjec-
© Society for Adolescent Medicine, 2001 1054-139X/01/$–see front matter
Published by Elsevier Science Inc., 655 Avenue of the Americas, New York, NY 10010 PII S1054-139X(01)00269-5
280 ZULLIG ET AL. JOURNAL OF ADOLESCENT HEALTH Vol. 29, No. 4

tive interpretations and evaluations of their experi- ronment. Furthermore, Newcomb et al. [18] found
ences that are most crucial. that alcohol use in grades seven through nine did not
Global life satisfaction has been defined as an directly affect later dissatisfaction with future oppor-
overall cognitive evaluation by the person of his or tunities; however, early dissatisfaction caused a di-
her life [7]. According to Pavot and Diener [8], rect, but small increase, in young adult alcohol use
satisfaction judgments are based on persons’ com- (aged 19 –22 years). For the same time period (grades
parisons between self-imposed criteria and their per- seven through nine into early adulthood, 4 years
ceived life circumstances. Such appraisals are ex- later), these adolescents reported better satisfaction
pected to influence the probability of subsequent regarding themselves and their relationships with
emotional and coping responses [9,10]. their parents, but lower satisfaction regarding their
Although demonstrating a significant degree of life pursuits and opportunities as young adults when
short-term and long-term stability [11], life satisfac- compared with their adolescent years.
tion reports are constructed as ongoing judgments The Clark and Kirisci study [17] yielded divergent
that reflect the influence of changing life circum- findings, however. This study was designed to ex-
stances [4]. As such, changes in global life satisfac- amine and compare the effects of Posttraumatic
tion may elicit changes in adolescents’ coping strat- Stress disorder (PTSD), major depression, and alco-
egies. For example, some individuals might respond hol use disorders on QOL during adolescence. Sub-
to declines in life satisfaction and associated emo- jects were 540 adolescents aged 12–18 years, re-
tions by engaging in various risk-taking behaviors cruited from clinical and community sources. In this
related to health (e.g. tobacco, alcohol, and other study, no relationship was observed between adoles-
drug abuse) in an attempt to improve their life cents’ life satisfaction and alcohol use disorders. It
satisfaction. Alternatively, engaging in a risk behav- should be noted that the Clark and Kirisci study
ior or behaviors may alter an individual’s life satis- employed a one-item interviewer rating of life satis-
faction. faction of unknown reliability; thus these results
Historically, life satisfaction research has been should be interpreted cautiously.
mostly limited to studies of adults [12]. During the There remains a paucity of research that attempts
past decade, however, increasing interest has been to associate adolescents’ perceived QOL with health
documented in the determinants, correlates, and risk behaviors leading to mortality and morbidity.
consequences of individual differences in life satis- Therefore, the purpose of this study was to investi-
faction among children and adolescents. Global life gate the relationship between adolescents’ life satis-
satisfaction reports of adolescents have been related faction and their self-reported substance abuse be-
to a variety of variables, including environmental haviors.
variables [13], demographic variables, such as socio-
economic status (SES) [14], and personality and tem-
perament variables [15,16]. Life satisfaction measures Methods
have been found to be related, but distinct from,
traditional measures of mental health [4,16]. Sample
A review of literature yielded only two studies The 1997 South Carolina Youth Risk Behavior Survey
that attempted to examine the relationship between (SCYRBS) used a sampling procedure designed to
life satisfaction and substance abuse behaviors obtain a representative sample (with the exception of
among adolescents [17,18]. Moreover, these two special education schools) of all South Carolina pub-
studies examined alcohol as the only substance of lic high school students in grades 9 –12 in the spring
abuse vis-à-vis life satisfaction. of 1997 [19]. The survey was previously determined
The longitudinal study by Newcomb et al. [18] to have adequate reliability [20]. The Youth Risk
suggests causal links between alcohol use and spe- Behavior Surveillance System (YRBSS) focuses on six
cific life satisfaction domains (self, peer relationships, major areas of adolescent behaviors: those that lead
future life opportunities, and globally-perceived en- to intentional and unintentional injuries, tobacco,
vironment), but reserved its conclusions strictly for alcohol and other drug use, sexual behaviors, dietary
alcohol. This study followed adolescents from grades behaviors, and physical inactivity [21].
seven through nine into early adulthood (4 years Schools were randomly selected using a system-
later, aged 19 –22 years), and found that alcohol use atic basis for each enrollment size category, by using
in grades seven through nine was associated with PCSAMPLE, a computerized sampling program,
dissatisfaction with peers and their perceived envi- which narrowed the 215 schools to 87 schools in 53
October 2001 LIFE SATISFACTION AND SUBSTANCE ABUSE 281

districts, statewide. The 215 schools were stratified options from the widely used Terrible–Delighted
by enrollment size into three enrollment categories: Scale [24] were used for each of these questions: (a)
schools with 74 to 874 students were classified as “Terrible,” (b) “Unhappy,” (c) “Mostly dissatisfied,”
small; schools with enrollments from 875 to 1278 (d) “Mixed” (equally satisfied and dissatisfied), (e)
were classified as medium; and schools of more than “Mostly satisfied,” (f) “Pleased,” and (g) “Delight-
1279 were classified as large. Our data include 63 of ed.”
the 87 eligible schools (72% response rate). School Cronbach alpha provided an estimate of internal
refusal was usually based on hectic schedules, prox- consistency for the Life Satisfaction Scale. The alpha
imity to standardized testing and accountability for coefficient for this total scale was .85. Removal of any
minutes in classroom instruction. Classes that met selected item resulted in alpha coefficients ranging
during second period were eligible for classroom from .80 to .85. Therefore, the total six-item scale was
level sampling selection to maximize student eligi- left intact.
bility. Using random starts, classes were selected
with a systematic equal probability sampling. When
the number of students selected by PCSAMPLE was Data Analysis
substantially less than expected, class samples were All six discrete variables from the Life Satisfaction
increased by the same systematic sampling of addi- Scale were pooled together to form a pseudo-contin-
tional classes. The desired number of participants uous dependent variable ranging in score from 6
(n ⫽ 125) was met by most schools, with the majority (1⫻6) to 42 (7⫻6). The satisfaction score was ex-
of schools sampling between 115 and 145 students. pressed as Mean Satisfaction Score (MSS) with lower
The survey was anonymous in that no personal scores indicative of reduced satisfaction with life.
identification was recorded. A school and class code The analyses were conducted separately for the
were used for CDC (Centers for Disease Control and four race/gender groups, white females (WF), black
Prevention) reporting purposes. Parent notification females (BF), white males (WM), and black males
forms were distributed at least five days in advance (BM). The use of indicator or dummy variables to
of survey administration; those parents who did not represent the four groups (WF, BF, WM, BM) would
want their children to participate were required to have assumed that the coefficients of all risk and
return the form (passive consent). Survey adminis- confounding variables were constant across all four
trations were conducted by trained data collectors, race/gender groups. Because this assumption was
emphasizing anonymity, privacy, and confidential- false, the four groups were analyzed separately.
ity. This research was approved by the host univer- To account for the three levels of outcome vari-
sity’s review board for the rights of human subjects ables (“satisfied,” “mid-range,” and “dissatisfied”)
in research. polychotomous logistic regression was conducted
using SAS callable Survey Data Analysis (SUDAAN)
[25] at ␣ ⫽ .05 for this analysis. SUDAAN takes into
Life Satisfaction Instrumentation and account the weighting as well as the stratification of
Instrument Validity schools and clustering within schools and classes
The Life Satisfaction Scale was based on the six nested within schools. The pooled dependent vari-
domains (family; friends; school; self; living environ- able was collapsed/categorized into three outcome
ment; and a global item, overall life) of the Multidi- levels: “dissatisfied,” “mid-range,” and “satisfied”
mensional Students’ Life Satisfaction Scale [22,23], adolescents. Because adolescents falling into the
which has been validated through exploratory and mid-range category were measured with one re-
confirmatory factor analytic procedures. Six life sat- sponse option, whereas the other two categories
isfaction items were developed for this study with comprised three response options collapsed into a
one question for each domain. The Instrument Reli- composite, a numeric collapse of the MSS deter-
ability six items were: “I would describe my satisfac- mined subject classification. MSS scores ranging
tion with my family life as;” “I would describe my from 6 to 27 were categorized as dissatisfied, MSS
satisfaction with my friendships as;” “I would de- scores ranging from 28 to 34 comprised the mid-
scribe my satisfaction with my school experience as;” range group, and MSS scores of 35 or greater defined
“I would describe my satisfaction with myself as;” “I the satisfied group. Adjusted odds ratios (OR) and
would describe my satisfaction with where I live as;” confidence intervals (CI) were calculated to deter-
and a global question; “I would describe my satis- mine which risk behaviors were associated with
faction with my overall life as.” Seven response reduced life satisfaction for each race/gender group
282 ZULLIG ET AL. JOURNAL OF ADOLESCENT HEALTH Vol. 29, No. 4

for both the mid-range and dissatisfied levels. Stu- (Table 3). An overall Chi-square value of 32.8 with 6
dents at the satisfied level served as the referent degrees of freedom with a p-value of .00001 was
group. Variables that did not meet the .05 signifi- determined. To focus on isolated differences either
cance level were not retained in the model. All above or below expectations, cell Chi-squares were
students who did not answer questions used in the calculated and tested with an alpha of 0.10 noting
regression procedure were excluded from the analy- observed (O) and expected (E) cases for the cell
sis by the numerical procedures of SAS/SUDAAN. Chi-squares (CCS).
The independent variables explored in this study Deviations from expectations were found in white
were use of tobacco, marijuana, alcohol, cocaine, females for Mid-Range LS (O ⫽ 557, E ⫽ 519, CCS ⫽
inhalants, steroids, illegal injected drugs, along with 2.80, p ⫽ .094), black females for High LS (O ⫽ 389,
age of first use for these substances (Table 1). Ac- E ⫽ 430, CCS ⫽ 3.91, p ⫽ .048), black males for
cording to the literature, the age of 13 years is a Mid-Range LS (O ⫽ 349, E ⫽ 415, CCS ⫽ 10.5, p ⫽
pivotal age for adolescents, with those experiment- .001), and black males for High LS (O ⫽ 395, E ⫽ 352,
ing and using before age 13 years being at greater CCS ⫽ 513, p ⫽ .024).
risk for continued use (and other consequences) Among the four race/gender groups there were
compared with those who delay substance use [26]. no statistical departures from expectations for Lower
As a result, the age variable was divided into initi- Satisfaction. At Mid-Range, white females were
ating the behavior at or before the age of 13 years overly represented (O-E ⫽ 38, p ⫽ .09), whereas
versus after the age of 13 years, with the referent black males are underrepresented (O-E ⫽ ⫺66,
group being the adolescents who did not drink. p ⫽.00001). Finally, for Higher Satisfaction, black
Adolescents were categorized as having participated females are under-represented (O-E ⫽ ⫺41, p ⫽ .05)
in the behavior or not, with the referent group and black males are overly represented (O-E ⫽ 30,
designated as those who have never participated in p ⫽ .02).
any of the drug behaviors. SAS/SUDAAN automat- None of the above departures or odds ratios that
ically deletes any incomplete record. Frequencies could be constructed from them have been adjusted
and percentages for the selected substance use be- for the risk factors located in Table 4.
haviors are presented in Table 1, by race and gender.

Relationship Between Mid-Range Level of Life

Results Satisfaction and Substance Use Behaviors
Description of Subjects Results here identified only three significant associ-
The total number of subjects who participated in the ations for one of the four race/gender groups. A
1997 SCYRBS was 5544 (Table 2). However, only significant association was established between self-
5032 valid observations were available owing to: (a) reported alcohol use in the last 30 days and the
nonresponse by subjects to the variables of interest, mid-range level of LS for white females (OR ⫽ 1.56,
(b) responses that could not be read, (c) out-of-range CI ⫽ 1.02–2.38). Also for white females, a significant
responses, and (d) exclusion of subjects reporting association was established among cigarette smoking
their race as “Other” than black or white (n ⫽ 388, (past 30 days) (OR ⫽ 1.28, CI ⫽ 1.01–1.63) and first
7.71%). Failure to answer any variable of interest cigarette at ⱕ age 13 years (OR ⫽ 1.33, CI ⫽
question also resulted in deletion (n ⫽ 125, 2.48%). Of 1.04 –1.68) and the mid-range level of LS. Owing to
the survey participants, 2650 (52.7%) were female the lack of association between substance use vari-
and 2382 (47.3%) were male. The study consisted of ables and the mid-range level of LS, these data are
2382 (47.3%) black students and 2650 (52.7%) white neither tabled nor discussed.
students. White females accounted for the largest
population (26.6%) followed by both black females
and white males (26.1%), and black males (21.2%). Relationship Between Life Dissatisfaction and
Selected Substance Use Behaviors
White females. Significant relationships were es-
Perceived Life Satisfaction tablished between dissatisfaction with life and all
A Chi-square analysis was performed testing the twenty-one substance use behaviors for white fe-
hypothesis of equal proportions (percentages) across males. Cigarette smoking (ⱕ age 13 years), cigarette
race/gender groups for levels of life satisfaction (LS) smoking (⬎ age 13 years), cigarette smoking (past 30
October 2001 LIFE SATISFACTION AND SUBSTANCE ABUSE 283

Table 1. Potential Substance Abuse Correlates by Race/Gender Group: Weighted Frequency (n) and Percent (%)
White Females Black Females White Males Black Males
Risk Variable n (%) n (%) n (%) n (%)
Ever try cigarette smoking (1 or 2 puffs)
Yes 1036 (21.0) 902 (18.3) 983 (19.9) 778 (15.8)
No 298 (6.0) 384 (7.8) 300 (6.1) 253 (5.1)
Age smoked first whole cigarette (y)
Never smoked 404 (8.2) 670 (13.6) 426 (8.7) 403 (8.2)
ⱕ13 410 (8.3) 160 (3.3) 487 (9.3) 221 (4.5)
⬎13 514 (10.4) 445 (9.1) 372 (7.5) 409 (8.3)
Cigarettes (past 30 days)
None 660 (13.9) 962 (20.2) 696 (14.4) 628 (13.2)
At least one cigarette 647 (13.6) 274 (5.8) 560 (11.6) 347 (7.3)
Spit tobacco (past 30 days)
None 1310 (26.4) 1291 (26.0) 1051 (21.0) 1014 (20.4)
At least one use 24 (0.5) 10 (0.3) 240 (4.8) 31 (0.6)
Age of first drink (y)
Never drank 256 (5.6) 334 (7.3) 275 (6.0) 254 (5.6)
ⱕ13 385 (8.5) 353 (7.7) 482 (10.6) 409 (9.1)
⬎13 612 (13.5) 481 (10.6) 448 (9.5) 269 (6.0)
Lifetime alcohol use
Never drank 256 (5.7) 333 (7.4) 274 (6.2) 256 (5.7)
At least one day 987 (22.0) 805 (18.0) 920 (20.5) 651 (14.5)
Alcohol use (past 30 days)
None 616 (13.1) 813 (17.3) 589 (12.6) 571 (12.2)
At least one drink 676 (14.4) 419 (8.9) 627 (13.4) 379 (8.1)
Binge drinking (past 30 days)
None 932 (19.0) 1159 (23.6) 793 (16.2) 840 (17.1)
Binge drank at least once 392 (8.0) 135 (2.8) 466 (9.5) 187 (3.8)
Age of 1st marijuana use (y)
Never used marijuana 750 (15.2) 845 (17.1) 669 (13.5) 492 (9.9)
ⱕ13 110 (2.2) 43 (0.8) 178 (3.6) 135 (2.7)
⬎13 474 (9.5) 405 (8.2) 442 (8.9) 411 (8.4)
Lifetime marijuana use
Did not use marijuana 750 (15.3) 844 (17.2) 669 (13.6) 492 (10.0)
Used marijuana at least once 583 (11.8) 442 (9.0) 610 (12.4) 525 (10.7)
Marijuana use (past 30 days)
No use 1007 (20.5) 1060 (21.6) 894 (18.2) 682 (13.9)
Used marijuana at least once 321 (6.5) 219 (4.5) 377 (7.7) 345 (7.1)
Age of first cocaine use (y)
Never used cocaine 1226 (24.7) 1290 (26.0) 1165 (23.4) 1015 (20.4)
ⱕ13 11 (0.3) 9 (0.2) 27 (0.5) 9 (0.3)
⬎13 97 (1.9) 4 (0.1) 97 (1.8) 15 (0.4)
Lifetime cocaine use
No use 1223 (24.8) 1279 (26.0) 1164 (23.6) 1007 (20.4)
Used cocaine at least once 106 (2.2) 9 (0.2) 113 (2.3) 24 (0.5)
Cocaine use (past 30 days)
No use 1301 (26.1) 1300 (26.0) 1255 (25.1) 1039 (20.8)
Used cocaine at least once 39 (0.8) 5 (0.2) 42 (0.8) 11 (0.2)
Lifetime crack cocaine use
Never used crack cocaine 1274 (25.4) 1294 (26.0) 1218 (24.4) 1046 (21.0)
Used crack at least once 60 (1.2) 8 (0.2) 73 (1.5) 13 (0.3)
Lifetime inhalant use
No use 1015 (20.3) 1187 (23.6) 965 (19.3) 980 (19.6)
Used inhalants at least once 324 (6.5) 118 (2.4) 333 (6.7) 81 (1.6)
Lifetime injection drug use
No use 1308 (26.3) 1301 (26.1) 1255 (25.2) 1036 (20.8)
Used illegal injection drugs at least once 25 (0.5) 1 (0.0) 36 (0.7) 19 (0.4)
Lifetime steroid use
No use 1279 (25.5) 1296 (25.8) 1224 (24.3) 1032 (20.5)
Used steroids at least once 62 (1.2) 12 (0.2) 84 (1.7) 39 (0.8)
284 ZULLIG ET AL. JOURNAL OF ADOLESCENT HEALTH Vol. 29, No. 4

Table 2. Demographic Characteristics of the Sample black females. Cigarette smoking (ⱕ age 13 years),
Variable n Weighted Percentage chewing tobacco use (past 30 days), alcohol use (ⱕ
Females 2650 52.7
age 13 years), alcohol use (⬎ age 13 years), lifetime
Males 2382 47.3 alcohol use, alcohol use (past 30 days), binge drink-
Whites 2650 52.7 ing (past 30 days), marijuana use (ⱕ age 13 years),
Blacks 2382 47.3 lifetime marijuana use, marijuana use (past 30 days),
White females 1341 26.7 lifetime cocaine use, lifetime crack or freebase co-
Black females 1309 26.0
White males 1309 26.0
caine use, lifetime inhalant use, and lifetime steroid
Black males 1073 21.3 use were all associated with reported decreased LS
Grade 9 1774 35.3 (Table 4). Substance-using black adolescent females
Grade 10 1233 24.5 increased the odds of reporting dissatisfaction with
Grade 11 1102 21.9 their lives ranging from 1.28 times (lifetime mari-
Grade 12 920 18.3*
juana use) to 6.21 times (lifetime crack or freebase use
This study analyzed black and white respondents only; N ⫽ of cocaine) in comparison to those black females who
5032.
* Grade was missing for 3 subjects; N ⫽ 5029. did not report substance use (Table 4).

White males. Significant relationships were estab-

days), chewing tobacco use (past 30 days), alcohol lished between dissatisfaction with life and sixteen of
use (ⱕ age 13 years), alcohol use (⬎ age 13 years), the twenty-one substance use behaviors for white
lifetime alcohol use, alcohol use (past 30 days), binge males. Cigarette smoking (ⱕ age 13 years), cigarette
drinking (past 30 days), marijuana use (ⱕ age 13 smoking (past 30 days), chewing tobacco use (past 30
years), marijuana use (⬎ age 13 years), lifetime days), alcohol use(ⱕ age 13 years), lifetime alcohol
marijuana use, marijuana use (past 30 days), cocaine use, alcohol use (past 30 days), binge drinking (past
use (ⱕ age 13 years), cocaine use (⬎ age 13 years), 30 days), marijuana use (ⱕ age 13 years), marijuana
lifetime cocaine use, lifetime crack or freebase co- use (⬎ age 13 years), lifetime marijuana use, mari-
caine use, cocaine use (past 30 days) lifetime inhalant juana use (past 30 days), cocaine use (ⱕ age 13
use, lifetime steroid use, and illegal injection drug years), cocaine use (⬎ age 13 years), lifetime cocaine
use were all associated with reported decreased LS use, lifetime crack or freebase cocaine use, cocaine
(Table 4). Substance-using white adolescent females use (past 30 days), lifetime inhalant use, lifetime
increased the odds of reporting dissatisfaction with steroid use, and illegal injection drug use were
their lives ranging from 1.43 times (binge drinking) associated with reported decreased LS (Table 4).
to 11.89 times (first cocaine use ⱕ age 13 years) in Substance-using white adolescent males increased
comparison to those white females who did not the odds of reporting dissatisfaction with their lives
report substance use (Table 4). ranging from 1.54 times (chewing tobacco [past 30
days] and first marijuana use age ⬎ 13 years) to 11.96
Black females. Significant relationships were deter- times (first cocaine use ⱕ age 13 years) in compari-
mined between reported dissatisfaction with life and son to those white males who did not report sub-
sixteen of the twenty-one substance use behaviors for stance use (Table 4).

Black males. Significant relationships were estab-

Table 3. Weighted Frequency and Percentage of lished between dissatisfaction with life and sixteen of
Perceived Life Satisfaction Level by Race and Gender the twenty-one substance use behaviors for black
(N ⫽ 5032) males. Cigarette smoking (ⱕ age 13 years), cigarette
Lower Higher smoking (⬎ age 13 years), cigarette smoking (past 30
Race and Gender Satisfaction Mid-Range Satisfaction days), chewing tobacco use (past 30 days), alcohol (ⱕ
White females 375 (28.0%) 557 (41.5%)⫹ 409 (30.5%) age 13 years), lifetime alcohol use, alcohol use (past
Black females 382 (29.2%) 538 (41.1%) 389 (29.7%)⫺ 30 days), marijuana use (ⱕ age 13 years), marijuana
White males 346 (26.4%) 503 (38.4%) 460 (35.1%) use (⬎ age 13 years), lifetime marijuana use, cocaine
Black males 329 (30.7%) 349 (32.5%)⫺ 395 (36.8%)⫹ use (ⱕ age 13 years), cocaine use (⬎ age 13 years),
⫺ ⫽ Significantly low. lifetime cocaine use, lifetime crack or freebase co-
⫹ ⫽ Significantly high (p ⫽ .10). caine use, cocaine use (past 30 days), lifetime inhal-
Chi-square 1 df, .90 ⫽ 2706 (If cell Chi-square ⬎ 2.706, declare
cell aberrant.). ant use, lifetime steroid use, and illegal injection
Chi-square 6df ⫽ 32.8, p ⫽ .00011. drug use were associated with reported decreased LS
October 2001 LIFE SATISFACTION AND SUBSTANCE ABUSE 285

Table 4. Associated Substance Use Behaviors for Race/Gender Groups at the Dissatisfied Level of Life Satisfaction
White Female Black Female White Male Black Male
Risk Factor OR (95% CI) OR (95% CI) OR (95% CI) OR (95% CI)
First cigarette smoke ⱕ age 13 2.63**** (2.01–3.46) 1.55* (1.02–2.36) 2.22**** (1.64 –3.00) 2.14*** (1.42–3.22)
First cigarette smoked ⬎ age 13 1.59** (1.16 –2.18) 1.29 (0.98 –1.69) 1.26 (0.89 –1.78) 1.62* (1.11–2.36)
Smoked cigarettes, past 30 days 2.42**** (1.93–3.04) 1.42 (1.03–1.24) 1.95*** (1.39 –2.73) 1.90**** (1.41–2.55)
Chewing tobacco use, past 30 days 3.33** (1.58 –7.03) 3.78* (1.00 –14.29) 1.54* (1.10 –2.14) 2.73* (1.09 – 6.82)
First alcohol drink ⱕ age 13 3.24**** (2.39 – 4.40) 2.04** (1.36 –3.40) 2.40**** (1.59 –3.63) 1.72** (1.99 –2.48)
First alcohol drink ⬎ age 13 2.09** (1.29 –3.38) 1.46* (1.04 –2.05) 1.34 (0.87–2.07) 1.59 (0.95–2.68)
Lifetime alcohol use 2.49**** (1.69 –3.65) 1.71** (1.21–2.41) 1.86** (1.26 –2.73) 1.58* (1.11–2.24)
Alcohol use, past 30 days 1.53**** (1.25–1.88) 1.63** (1.17–2.25) 1.59* (1.11–2.28) 1.53* (1.13–2.08)
Binge drank, past 30 days 1.43** (1.13–1.81) 2.05*** (1.43–2.29) 1.43* (1.07–1.92) 1.71** (1.28 –2.29)
First marijuana use ⱕ age 13 4.23**** (2.78 – 6.42) 1.65* (1.02–16.59) 3.15**** (2.29 – 4.35) 1.74 (0.98 –3.10)
First marijuana use ⬎ age 13 1.79**** (1.47–2.64) 1.26 (0.99 –1.60) 1.54* (1.04 –2.29) 1.31* (1.02–1.68)
Marijuana use lifetime 2.30**** (1.75–3.20) 1.28* (1.00 –1.63) 1.91**** (1.40 –2.61) 1.39* (1.05–1.84)
Marijuana use, last 30 days 2.53**** (1.90 –3.37) 1.61** (1.22–2.13) 2.16**** (1.60 –2.92) 1.33 (0.97–1.48)
First cocaine use ⱕ age 13 11.89**** (5.13–27.57) 4.40 (0.83–23.42) 11.96**** (4.43–32.29) 16.80* (2.00 –140.95)
First cocaine use ⬎ age 13 3.66**** (2.60 –5.15) 2.20 (0.39 –12.50) 2.35** (1.51–3.65) 3.69* (1.46 –9.28)
Lifetime cocaine use 4.05**** (2.92–5.62) 4.22* (1.07–16.59) 2.82**** (1.87– 4.26) 5.73** (2.14 –15.34)
Lifetime crack or freebase use 4.31**** (2.45–7.59) 6.21* (1.24 –31.02) 3.13**** (1.98 – 4.93) 11.71** (2.37–57.93)
Cocaine use, past 30 days 3.50*** (1.80 – 6.81) 2.93 (0.45–19.17) 3.24*** (1.79 –5.83) 11.09* (1.51– 81.25)
Lifetime inhalant use 2.05** (1.59 –2.64) 2.05** (1.28 –3.26) 2.28* (1.70 –3.07) 2.58**** (1.67–3.89)
Lifetime steroid pills or shots 2.11** (1.36 –3.29) 4.63* (1.35–15.87) 2.11** (1.28 –3.48) 7.90**** (3.61–17.29)
without prescription
Inject illegal drugs lifetime 5.79**** (2.59 –12.90) Inadequate cell 4.96**** (2.68 –9.17) 11.78**** (3.84 –36.20)
size to estimate risk
* p ⬍ .05; ** p ⬍ .005; *** p ⬍ .0005; **** p ⬍ .0001.
OR ⫽ odds ratio; CI ⫽ confidence interval.

(Table 4). Substance-using black adolescent males life satisfaction and inhalant use has not been previ-
increased the odds of reporting dissatisfaction with ously described, it is congruent with previous find-
their lives ranging from 1.31 times (first marijuana ings concerning various negative characteristics of
use ⬎ age 13 years) to 16.8 times (first cocaine use ⱕ youth involved in inhalant use [27–30]. The direction
age 13 years) in comparison to those black males who of this relationship cannot be discerned from this
did not report substance use (Table 4). study. Young adolescents dissatisfied with life might
abuse inhalants as an ill-advised coping mechanism
for dealing with life dissatisfaction, or experiment
Discussion with inhalants out of curiosity. As inhalant use
Results suggest that a substantial number of public becomes less satisfying, a young adolescent may
high school students are reporting dissatisfaction pursue other substances [31], which may, in turn,
with their lives. This study also suggests that a increase dissatisfaction with life.
substantial number of public high school adolescents All four tobacco use behaviors were significantly
are also engaging in substance abuse behaviors. Most related to reported life dissatisfaction for at least two
importantly, this study identified significant rela- and sometimes all four of the race/gender groups.
tionships between self-reported LS and substance Smoking the first cigarette at or less than age 13
use behaviors. Furthermore, this study demonstrated years, and chewing tobacco use was significantly
the importance of two contextual factors as modera- associated with life dissatisfaction for adolescents.
tors of the relationships between life satisfaction and For white females and back males, smoking the first
specific types of substance use. That is, the nature cigarette after age 13 years and for these two race/
and magnitude of the substance use: life satisfaction gender groups in addition to white males, smoking
associations were significantly influenced by both in the past 30 days was associated with reported life
student gender and ethnicity. dissatisfaction. Tobacco use has been associated with
Lifetime inhalant use appears to be strongly asso- truancy [26], increased sexual risk behavior [32]
ciated with life dissatisfaction for all four race/ disconnectedness with parents [33], school perfor-
gender groups. Whereas this relationship between mance, and lower household income [34]. No exist-
286 ZULLIG ET AL. JOURNAL OF ADOLESCENT HEALTH Vol. 29, No. 4

ing literature connects life satisfaction with tobacco have adjustment difficulties and problems by age 16
use; however, reduced perceived life satisfaction [40].
could be associated with any of these aforemen- For all four race/gender groups, using steroids
tioned behaviors. If an adolescent is dissatisfied with was significantly associated with life dissatisfaction.
life, he/she could be more apt to engage in risk- The same is true for using illegal injection drugs
taking behaviors and perform poorly in school. Con- (lifetime) with the exception of black females. Ana-
versely, if an adolescent is performing poorly in bolic steroid use is associated with being physically
school and taking health-related risks, they may be aggressive and powerful, psychologically more ag-
dissatisfied with life. Tobacco use can be the first gressive, socially assertive, sexually aroused, and
indication of adolescent problem behavior [26]. socially accepted [41]. Whether steroid use is chosen
With the exception of black females, age of cocaine to improve physical performance, appearance or
use (ⱕ13 and ⬎13), and 30 day prevalence of cocaine mental confidence, users who reported steroid use
use, significant relationships between all the cocaine may be dissatisfied with at least one aspect of their
use behaviors and dissatisfaction with life were adolescent development.
found for all four race/gender groups. Documented Adolescent risk behaviors tend to cluster
personality disorders connected with cocaine use [26,32,42] and perhaps, dissatisfaction with life
include: antisocial, borderline, self-defeating and de- among adolescents could be associated with other
pendency behaviors, narcissism, low self-esteem, health risk behaviors such as sexual risk-taking,
passivity, and “inner emptiness” [35]. Adolescents violence, aggression, suicide ideation, dieting and
who reported a dissatisfied life in this study and also physical activity behaviors, among others. Future
reported crack use could be at the “final stage” of research should consider the association with re-
substance use [36]. When comparing cocaine users to ported life satisfaction and other risk-taking behav-
crack users, crack users are more likely to be iors for adolescents.
younger, make poorer grades, be more depressed, Behavioral health specialists have recently begun
and more alienated from family and friends than to focus greater attention on the assessment and
cocaine users [37]. Cocaine users have also been promotion of youth developmental assets, and the
shown to be more prone to mental problems than positive aspects of psychological well-being and ad-
nonusers [35]. aptation [43– 46]. This focus broadens the matrix of
With the exception of males, for having their first youth assessment beyond those of risk behaviors and
alcohol drink after age 13 years and binge drinking psychiatric symptoms [22,26,32,42], and thus, may be
for black males, every alcohol behavior was signifi- more sensitive to subtle, but clinically meaningful,
cantly associated with self-reported life dissatisfac- changes in adolescent emotional well-being. Subjec-
tion. The significant association between binge tive life satisfaction measures can also extend the
drinking and reduced life satisfaction is an important scope of well-being indicators beyond the objective
finding vis-à-vis characteristics significant in previ- indicators often used exclusively in federally spon-
ous studies including: self-derogation, dissatisfaction sored large-scale surveys of American children’s
with peer relationships, dissatisfaction with future well-being. For example, one recent federal survey of
opportunity, dissatisfaction with perceived environ- children entitled “America’s Children: Key National
ment, and depression [18,38]. Indicators of Well-Being” [47] was comprised en-
With the exception of black females using mari- tirely of objective indicators such as teenage preg-
juana for the first time after the age of 13 years, and nancies, drug and alcohol usage, and family income.
black males past 30 day use of marijuana, all four The many potential uses of measures of adolescents’
marijuana behaviors were significantly associated perceptions of their QOL, such as general life satis-
with reports of life dissatisfaction. Literature in this faction measures, have been delineated in a number
regard suggests that early use of marijuana has of sources [48 –50], including uses in social and
consequences for later psychological development health policy formulation, needs assessment, and
[39]. Young people who have used marijuana on ten program evaluation.
or more occasions have been reported to be signifi- Study limitations should be noted here. These
cantly more likely to have come from socially disad- include: the use of a cross-sectional study that pre-
vantaged backgrounds, to have been exposed to cludes determination of the temporal sequence of
family adversities during childhood, to have formed reduced life satisfaction and substance use. This
affiliations with delinquent or substance-using peers, study may reflect only South Carolina youth, and
to have reported poorer parental attachment, and to may not be nationally representative. The logistic
October 2001 LIFE SATISFACTION AND SUBSTANCE ABUSE 287

regression (SAS/SUDAAN) called for the elimina- 7. Pavot W, Diener E, Colvin CR, et al. Further validation of the
Satisfaction with Life Scale: Evidence for the cross-method
tion of subjects with missing data on the variables of convergence of well being measures. Pers Assessment 1991;
interest. Therefore, observed associations are likely 57:149 – 61.
to be conservative (i.e., an underestimate of the 8. Pavot W, Diener E. Review of the Satisfaction with Life Scale.
magnitude of the reported associations). Psychol Assessment 1993;5:164 –73.
The study strengths include: the large sample size; 9. Diener E. Assessing subjective well-being: Progress and op-
portunities. Soc Indic Res 1994;31:103–57.
the random sampling technique; the stratification of
10. Lazarus R. Progress on a cognitive-motivational-relational
schools by size; good school and class response rates theory of emotion. Am Psychol 1991;46:819 –34.
(72% and 88%, respectively); reliability and construct 11. Gilman R, Huebner ES. Children’s reports of their life satis-
validity of the life satisfaction measure; the number faction: Convergence across raters, time and response formats.
and types of drug abuse variables used; and the Sch Psychol Int 1997;18:133– 47.
analysis of results by race and gender. In addition, 12. Veenhoven R. Bibliography of Happiness: 2472 Contemporary
Studies of Subjective Appreciation in Life. Rotterdam, Neth-
the dichotomized variables and associations estab- erlands: RISBO, 1988.
lished between selected substances and life satisfac- 13. McCullough G, Huebner ES, Laughlin JE. Life events, self-
tion could be even stronger and more conservative concept and adolescents’ positive subjective well-being. Psy-
(i.e., an underestimation of the magnitude of risk and chol in Schs 2000;37:281–90.
14. Dew T, Huebner ES. Adolescent’s perceived quality of life: An
more toward the null hypothesis of no association) exploratory investigation. J Sch Psychol 1994;32:185–99.
than reported. 15. Huebner ES. Correlates of life satisfaction in children. Sch
Psychol Q 1991;6:103–111.
16. Huebner ES, Alderman GL. Convergent and discriminant
validation of a children’s life satisfaction scale: Its relationship
Conclusions to self- and teacher-reported psychological problems and
It is difficult to explain whether these adolescents school functioning. Soc Indic Res 1993:30;71– 82.
participate in substance abuse behaviors because of 17. Clark DB, Kirisci L. Posttraumatic Stress disorder, depression,
alcohol use disorders and quality of life in adolescents. Anx-
dissatisfaction with life, or whether they are dissat- iety 1996;2:226 –33.
isfied with life owing to substance abuse. Longitudi- 18. Newcomb MD, Bentler PM, Collins C. Alcohol use and
nal studies are needed to determine whether dissat- dissatisfaction with self and life: A longitudinal analysis of
isfaction with life is a consequence or determinant of young adults. J Drug Issues 1986;63:479 –94.
19. Valois RF. South Carolina Youth Risk Behavior Survey Re-
substance abuse behavior. port—1997. Columbia, SC: Department of Health Promotion &
Education, School of Public Health, University of South Caro-
This research was funded by Cooperative Agreement #U63/CCU lina, 1998.
802750-04, US Centers for Disease Control & Prevention, National
Center for Disease Prevention and Health Promotion, Division of 20. Brener ND, Collins JL, Kann, L, et al. Reliability of the Youth
Adolescent and School Health, Atlanta, GA and Cooperative Risk Behavior Survey questionnaire. Am J Epidemiol 1995;
Agreements with the South Carolina Department of Education. 41:575–580.
21. Kolbe LJ. An epidemiological surveillance system to monitor
the prevalence of youth behaviors that most affect health.
J Health Educ1990; 21(6):44 – 8.
22. Huebner ES, Laughlin JE, Ash C, Gilman R. Further validation
References of the multidimensional students’ satisfaction scale. J Psycho-
1. Epp J. Achieving Health for All: A Framework for Health educational Assess 1998:16:118 –34.
Promotion. Ottawa, Canada: Health and Welfare Canada, 23. Gilman R, Huebner ES, Laughlin JE. A first study of the
1986. multidimensional students’ life satisfaction scale with adoles-
2. Raphael D. The determinants of adolescent health: Evolving cents. Soc Indic Res 2000;52:135–160.
definitions, recent findings, and proposed research agenda. J 24. Andrews FM, Withey SB. Social Indicators of Well-Being:
Adolesc Health 1996;19:6 –16. Americans’ Perceptions of Life Quality. New York: Plenum,
3. Renwick R, Brown I, Nagler N (eds). Quality of Life in Health 1976.
Promotion and Rehabilitation: Conceptual Approaches, Issues 25. Shah BV, Barnwell BG, Bieler GS. SUDAAN User’s Manual,
and Applications. Thousand Oaks, CA: Sage, 1996. Release 7.0. Research Triangle Park, NC: Research Triangle
Institute, 1996.
4. Diener E, Suh E, Oishi S, et al. Subjective well-being: Three
decades of progress. Psychol Bull 1999;125:276 –302. 26. Dryfoos JG. Adolescence at Risk: Prevalence and Prevention.
New York: Oxford University Press, 1990.
5. Huebner ES, Dew T. The interrelationships of positive affect,
negative affect and life satisfaction in an adolescent sample. 27. Edwards RW, Oetting ER. Inhalant Use in the United States.
Soc Indic Res 1996;38:129 –37. NIDA Research Monograph, 148, 8 –28, Washington DC, 1995.
6. Day H, Jankey SG. Lessons from the literature: Toward a 28. Heath DB. Inhalant abuse moving into the mainstream. Behav
holistic model of quality of life. In: Renwick R, Brown I, Nagler Health Manage 1994;14(3):47– 8.
M (eds). Quality of Life in Health Promotion and Rehabilita- 29. Johnson CC, Myers L, Webber LS, et al. Alcohol consumption
tion: Conceptual Approaches, Issues and Applications. Thou- among adolescents and young adults: The Bogalusa heart
sand Oaks, CA: Sage, 1996:39 –50. study, 1981 to 1991. Am J Public Health 1995;85(7):979 – 85.
288 ZULLIG ET AL. JOURNAL OF ADOLESCENT HEALTH Vol. 29, No. 4

30. McGarvey EL, Canterbury RJ, Waite D. Delinquency and 40. Fergusson DM, Horwood J. Early onset of cannabis use and
family problems in incarcerated adolescents with and without psychosocial adjustment in young adults. Addiction 1997;
a history of inhalant use. Addict Behav 1996;21(4):537– 42. 92(3):279 –96.
31. Martin C, Arria AM, Mezzich AC, Bukstein OG. Patterns of 41. Schwerin MJ, Cocoran KJ. Beliefs about steroids: User vs.
polydrug use in adolescent alcohol abusers. Am J Drug non-user comparisons. Drug Alcohol Dependence 1996;40:
Alcohol Abuse 1993;19(4):511–21. 221–5.
42. Valois RF, McKeown R, Garrison C, Vincent ML. Correlates of
32. Valois RF, Oeltmann JE, Waller J, Hussey JR. Relationship
aggressive and violent behaviors among public high school
between number of sexual intercourse partners and selected
adolescents. J Adol Health 1995:16(1):26 –34.
health risk behaviors among public high school adolescents. J
Adol Health 1999;25(5):328 –35. 43. Compas B. Promoting positive mental health in adolescence.
In: Millstein SG, Peterson AC, Nightingale EO (eds). Promot-
33. Diem EL, McKay LC, Jamieson JL. Female adolescent alcohol, ing the Health of Adolescents. New York: Oxford University
cigarette, and marijuana use: Similarities and differences in Press, 1993:159 –79.
patterns of use. Int J Addict 1994;29(8):987–97. 44. Huebner ES. Initial development of the Students’ Life Satis-
34. Teh-wei H, Lin Z, Keeler T. Teenage smoking, attempts to faction Scale. School Psych Int 1991;12:231– 40.
quit, and school performance. Am J Public Health 1998;88(6): 45. Seligman M. What is the good life? Am Psych Assn Monitor
940 –3. 1998;29(10):2.
35. Fieldman NP, Woolfolk RL, Allen LA. Dimensions of self- 46. Scales PC. Reducing risks and building develomental assets:
concept: A comparison of heroin and cocaine addicts. Am J Essential actions for promoting adolescent health. J School
Drug Alchohol Abuse 1995;21(3):315–26. Health 1999;69(3):113–9.
36. Kandel D, Yamaguchi K. From beer to crack: Developmental 47. Federal Interagency Forum on Child and Family Statistics.
patterns of drug involvement. Am J Public Health 1993;83(6): America’s Children: Key National Indicators of Well-being.
851–5. Washington, DC: Federal Interagency Forum on Child and
Family Statistics, 1997.
37. Ringwalt CL, Palmer JH. Cocaine and crack users compared.
48. Dew T, Huebner ES. Adolescents perceived quality of life: An
Adolescence 1989;24(96):851– 8.
exploratory investigation. J School Psych 1994;32:185–99.
38. DiSimone A, Murray P, Lester D. Alcohol use, self-esteem, 49. Frisch MB, Cornell J, Villanueva M, Retzlaff B. Clinical vali-
depression, and sucidality in high school students. Adoles- dation of the Quality of Life Inventory. Psych Assessment
cence 1994;24(116):939 – 42. 1992;4:92–101.
39. Kandel DB, Davies M, Karus D, Yamaguchi K. The conse- 50. Raphael D, Rukholm E, Brown I, et al. The Quality of Life
quences in young adulthood of adolescent drug involvement: Profile—Adolescent version: Background, description, and
An overview. Arch Gen Psychiatry 1986;43:746 –54. initial validation. J Adol Health 1996;19(5):366 –75.