JOURNAL OF ADOLESCENT HEALTH 2002;30:448– 454
ORIGINAL ARTICLE
Social Skills and Attitudes Associated With Substance
Use Behaviors Among Young Adolescents
SHARI L. BARKIN, M.D., M.S.H.S., KAREN S. SMITH, M.P.H., R.H.Ed.,
AND ROBERT H. DuRANT, Ph.D.
Purpose: To examine how adolescents’ attitudes and
social skills affect current substance use and intentions
to use substances in the future.
Methods: An anonymous questionnaire was adminis-
tered to 2646 seventh graders in their classrooms. The
questionnaire was developed to measure the frequency
of tobacco, alcohol, and other substance use, anticipated
use, positive attitudes toward drug use, self-efficacy to
say “no,” decision-making skills, advertising-viewing
skills, anxiety-reducing skills, communication skills,
drug-resistance skills, perception of peer substance use,
and weapon-carrying behavior. Ethnicity classified re-
spondents as “white” or “students of color” and family
structure indicated one vs. two-parent families. Data
were analyzed with Spearman’s r, analysis of variance,
and multiple linear regression.
Results: Forty-one percent of students were minority,
50.6% female, over 90% were either 12- or 13-year-olds,
and 69.9% lived in two-parent families. A multiple linear
regression model demonstrated that self-efficacy to say
“no, positive attitudes toward drug use, perception of
peer substance use, male gender, weapon-carrying, and
fighting accounted for 51% of the variation in the current
use multiple substance scale. Anticipated substance use
during the subsequent year was significantly associated
with current substance use, positive attitudes toward
drug use, self-efficacy to say ”no, drug-resistance skills,
weapon-carrying, and fighting behavior. This model ac-
counted for 73.9% of the variance in anticipated sub-
stance use.
From the Brenner Center For Child and Adolescent Health, Depart-
ment of Pediatrics, Wake Forest University School Of Medicine, Win-
ston-Salem, North Carolina.
Address correspondence to: Shari Barkin, M.D., M.S.H.S., Wake
Forest University Baptist Medical Center, Medical Center Blvd., Win-
ston-Salem, NC 27157. E-mail: sbarkin@wfubmc.edu.
Manuscript accepted November 15, 2001.
1054-139X/02/$–see front matter
PII S1054-139X(01)00405-0 Published by Elsevier Science Inc., 655 Avenue of the Americas, New York, NY 10010
Conclusions: In today’s world, where drug use is com-
mon, building adolescents’ drug-resistance skills and
self-efficacy, while enhancing decision-making capacity,
may reduce their use of illegal substances. © Society for
Adolescent Medicine, 2002
KEY WORDS:
Young adolescents
Substance use
Self-efficacy skills
Engaging in health risk behaviors during early ado-
lescence not only threatens adolescents’ physical
health, but also negatively affects normal psychoso-
cial development [1– 4]. Among health risk behav-
iors, substance use is a significant threat to adoles-
cents [2]. Both early age of onset and frequency of
substance use clusters with violent and aggressive
behavior, weapon-carrying, early initiation of sexual
intercourse, suicide attempts, and poor academic
achievement [3,5– 8]. Experimentation with sub-
stances commonly begins in early adolescence. By
the eighth grade, almost 50% of students report
experimentation with cigarettes and almost one-
third report drinking alcohol to intoxication [2].
Alcohol and tobacco use has remained relatively
constant over the past two decades, whereas the use
of more illicit substances has declined [2]. Social
modeling in primary social groups, such as friends
and family, influences the development of health risk
behaviors [4]. In particular, some factors that are
associated with substance use include being raised in
a single-parent family with less parental supervision
and exposure to parents or friends who use sub-
stances [9 –12].
© Society for Adolescent Medicine, 2002
June 2002 YOUNG ADOLESCENTS AND SUBSTANCE USE 449

Social Cognitive Theory has been applied to min- Table 1. Demographic Characteristics of Seventh Grade
imize adolescent health risk behaviors, such as sub- Student Participants (n ⫽ 2,646)
stance use [13–16]. Multiple components are used: (a) n %
knowledge about the consequences of substance use Age (years)
must be acquired, (b) social and self-regulative skills ⬍12 78 2.9
for resisting substance use must be developed, (c) 12 1685 63.7
13 724 27.4
opportunities for guided practice and corrective ⬎13 105 4.0
feedback to build self-efficacy must be provided; and Missing 54 2.0
(d) changes in social support and norms concerning Gender
substance use must be addressed [13–16]. Based on Male 1286 48.6
Female 1340 50.6
Social Cognitive Theory, we hypothesize that the Missing 20 0.8
frequency with which young adolescents currently Family Structure
engage in substance use and anticipate engaging in Single-parent home 668 25.2
substance use in the coming year will be inversely Two-parent home 1850 69.9
Missing 128 4.8
associated with their self-efficacy to say “no” to
Ethnicity
anyone who offers them substances (tobacco, alco- White 1544 58.4
hol, marijuana, or other illicit substances), their abil- Students of color 1095 41.4
ity to critically review advertisements indicating a Missing 7 0.3
better knowledge base, their critical decision-making
skills, their anxiety-reducing skills, their communi-
cation skills, and their drug-resistance skills indicat- The Instrument
ing social and self-regulative skills. We expect that The questionnaire and all the constructs were
their own social norms, influenced by their culture/ adapted from the instrument used by Botvin et al. to
ethnicity and parental resources (one- vs. two-parent evaluate the Life Skills Training curriculum [13–15].
homes) could influence their decisions [3,4,17]. The primary modifications included evaluating each
substance use variable over the period of “30-day use
vs. ”lifetime use“ and decreasing the number of total
items asked about each substance. The frequency of
Methods substance use during the previous 30 days (current
Study Design use) was measured on seven-point ordinal scales,
The study protocol was approved by the Internal ranging from ”never“ to ”more than once a day.“
Review Board at Wake Forest University School of Anticipated use of each substance during the upcom-
Medicine. Verbal consent was obtained from each ing year was measured on five-point ordinal scales
ranging from ”definitely not“ to ”definitely will.“ A
student. Written informed consent was not obtained
multiple substance use scale was constructed by
from a parent or guardian because this study was
adding the number of substances used and the
part of an evaluation of a new curriculum in the
frequency that the drug was used during the previ-
middle schools. An anonymous questionnaire was
ous 30 days. Likewise, an anticipated substance use
administered to 2646 seventh grade students in class-
scale was constructed with similar questions that
rooms by investigators before implementing the Life projected students’ thoughts on whether they would
Skills Training curriculum in 1999. All enrolled sev- be using any substances within the next year. Com-
enth grade students in Winston-Salem Forsyth mon substances included were tobacco, alcohol, mar-
County public schools participated (Table 1). The ijuana, cocaine/crack, and inhalants. The internal
curriculum was designed to address the use of consistency of the scales were measured with Cron-
substances such as tobacco, alcohol, and other drugs bach alpha. These questions were used to construct a
by educating students about general health, building scale to assess frequency of current substance use
essential life skills, and developing drug-resistance (␣ ⫽ .79) and anticipated use (␣ ⫽ .73).
skills. Those students who attended sixth grade in The survey developed by Botvin et al. [13–15]
Forsyth County participated in a 15-session curricu- contained questions that were used to create scales
lum. This paper reports on the cross-sectional data measuring positive attitudes toward substance use
collected before seventh grade students participated (␣ ⫽ .95), self-efficacy to say ”no“ if offered tobacco,
in a 10-session booster curriculum. alcohol, or other drugs (␣ ⫽ .85), drug-resistance
450 BARKIN ET AL.
skills (␣ ⫽ .82), skills to critically view advertising
(␣ ⫽ .85), critical decision-making skills (␣ ⫽ .92),
anxiety-reducing skills (␣ ⫽ .87), and basic commu-
nication skills (␣ ⫽ .76). Additionally, questions were
asked about students’ perceptions of peer substance
use (␣ ⫽ .83), recent weapon-carrying behavior, and
recent involvement in physical fights. Data gathered
on weapon-carrying behavior and physical fights
used questions from the Youth Risk Behavior Survey
(YRBS) [2]. Data on gender, age, ethnicity, and family
structure were also gathered.
Most of the constructed scales reflected students’
attitudes and behaviors. For example, the positive
attitudes toward substance use scale contained 15
items, including statements such as, ”kids who drink
alcohol are more grown-up“ and ”kids who smoke
have more friends.“ Response categories ranged
from ”strongly disagree“ to ”strongly agree.“ The
self-efficacy to say ”no“ items included five ques-
tions like, ”[would you] say ’no’ when someone tries
to get you to smoke marijuana?“ Response categories
ranged from ”definitely would say no“ to ”definitely
would not say no.“ Five items addressed drug-
resistance skills such as telling peers you don’t want
to do it, changing the subject, making an excuse, and
leaving the situation. Response categories included
”definitely would“ to ”definitely would not.“ Critical
advertising (ad) viewing skills (5 questions) investi-
gated if students think about whether what the ad
says is true and if they tell themselves that smoking
cigarettes will not make their life better when they
hear an ad. Critical decision-making skills (5 ques-
tions) queried if students ”get the information
needed to make the best choice“ and if they ”think of
as many possible choices or ways of solving the
problem as [they] can.“ Anxiety-reducing skills in-
cluded five items such as: ”relaxing all the muscles in
my body“ and ”practicing doing whatever makes me
feel nervous until I feel more confident.“ Communi-
cation skills emphasized 4 behaviors such as ”mak-
(ing) sure that what I say matches my tone of voice,
how I stand, and the expression on my face“ and
”asking questions if (other people) say something
that isn’t clear.“ For critical decision-making, anxi-
ety-reduction, and communication skills, response
categories ranged from ”never“ to ”always.“
The peer perception scale (5 items) gathered infor-
mation on students’ perceptions of how many peers
are using substances such as tobacco, alcohol, mari-
juana, and other illicit drugs. Students rated how
many people their age use substances as ”none, “less
than half, ”about half, “more than half, and ”all or
almost all.“
JOURNAL OF ADOLESCENT HEALTH Vol. 30, No. 6
Weapon-carrying behavior was assessed with the
question, ”During the past 30 days, on how many
days did you carry a weapon such as a knife or a
wooden bat?“ Additional weapon-carrying behav-
iors were assessed, including carrying a gun on
school property and not going to school because they
felt unsafe. However, these questions had limited
variation and, therefore, were not included in the
model of current and anticipated substance use.
Likewise, physical-fighting behavior had the greatest
variation with the question, ”During the past 12
months, how many times were you in a physical
fight?“ Other questions included, ”During the past 12
months, how many times did you start a physical fight
with a boyfriend, girlfriend, or date?“ and ”Did a
boyfriend/girlfriend/date start a fight with you?“
However, these questions also yielded little variation.
Statistical Analysis
Data were collected to examine two dependent vari-
ables: (a) substance use in the past 30 days (current
use), and (b) anticipated substance use in the next
year. Independent variables collected included: gen-
der, ethnicity, family structure, positive attitudes
about substance use, self-efficacy to say ”no“ skills,
drug-resistance skills, critical advertising viewing
skills, critical decision-making skills, anxiety-reduc-
ing skills, and communication skills. The ethnicity
variable was recoded to classify respondents as
”white“ or ”students of color“ and family structure
included one-parent or two-parent families. Bivariate
relationships were examined between the dependent
variables and independent variables including weap-
on-carrying and physical fighting, and students’ per-
ceptions of peer substance use, as well as descriptive
univariate analyses. The data were analyzed with anal-
ysis of variance and multiple linear regression tests.
Our models were constructed by putting all variables
in at once, reflecting our specific aims. Therefore, for
the multiple linear regression analysis a partial r was
computed to determine the strength of the relationship
between each independent variable and the dependent
variables after adjusting for all the other variables in the
model. The R2 was used to describe the percent varia-
tion accounted for in the dependent variable by all the
independent variables in the model.
Results
Descriptive Statistics
Forty-one percent were students of color and 50.6%,
female. The majority ranged in age from 12 to 13
June 2002 YOUNG ADOLESCENTS AND SUBSTANCE USE 451

Table 2. Multiple Linear Regression Model of “Current” Drug Use

Variables Beta SE Partial r p value
(Constant) 3.628 .375 .0001
Gender ⴚ.267 .091 ⴚ.069 .003
Family Structure ⫺.168 .111 ⫺.036 .130
Ethnicity .030 .100 .007 .767
Self-efficacy to say “no” .545 .024 .473 .0001
Critical ad viewing skills .015 .009 .038 .106
Positive attitude toward drugs .076 .008 .228 .0001
Critical decision-making skills ⫺.009 .013 .017 .475
Anxiety-reducing skills .014 .011 .030 .204
Communication skills ⫺.009 .018 ⫺.012 .622
Drug-resistance skills ⫺.014 .012 ⫺.028 .245
Perception of peer substance use .072 .011 .069 .0001
Weapon-carrying behavior .129 .044 .166 .0001
Fighting behavior .064 .042 .079 .001
Model Adjusted R2: 51.0%
SE ⫽ standard error.

years, and 69.9% lived in two-parent families. The
frequency of substance use in the previous 30 days
(ranging from 1 ⫽ ”never“ to 7 ⫽ ”more than once a
day“) indicated that 14.1% used cigarettes at least
once, 18.5% used alcohol at least once, 5% smoked
marijuana at least once, 6.5% used inhalants at least
once, and 0.6% used cocaine or other ”hard drugs“ at
least once. For anticipated use in the next year,
almost one-quarter of students reported that they
might use either cigarettes or alcohol, about 10%
thought they might use marijuana or inhalants, and
fewer than 3% thought they might use ”hard drugs.“
Contrary to their own reported use, students
perceived that a significant number of peers used
substances. About one-third of students thought 50%
of their peers used cigarettes or alcohol and approx-
imately one-half of students thought less than half of
their peers used marijuana, inhalants, or hard drugs.
About 10% of students thought that smoking ciga-
rettes or drinking alcohol makes kids look more
grown-up, look ”cool, or lets you have “more fun.”
About 15% of students thought using substances
makes you have more friends.
Saying, “no” to smoking a cigarette or drinking
alcohol was not a clear course of action for more than
a quarter of students. About one-fifth of students
were unsure about saying “no” to someone getting
them to try inhalants or marijuana. Students re-
ported considering using other drug-resistance skills
such as changing the subject or leaving the situation
about 75% of the time. Twenty percent of students
reported that they would not be able to say that they
didn’t want to do it. At least one-third of students
reported minimal critical advertising viewing skills
and limited critical decision-making skills. Half of
students did not report using anxiety-reducing skills
when they feel anxious. Likewise, only half of stu-
dents reported using clear communication skills with
peers.
More than one-third of students reported being in
a physical fight over the past year, with 15% report-
ing that the physical fight occurred on school prop-
erty. Moreover, 15% of students reported carrying a
weapon such as a knife or a bat during the past 30
days. Six percent of respondents reported carrying a
gun other than for hunting or target practice.
Bivariate analysis revealed that weapon-carrying
and fighting behavior was significantly associated
with substance use, both current and anticipated
(p ⬍ .0001). Also, perceptions of peer substance use
were significantly associated with both current and
anticipated substance use (p ⬍ .0001).
Multiple Linear Regression
A regression model containing self-efficacy to say
“no, critical advertising viewing skills, positive atti-
tudes toward drug use, critical decision-making
skills, anxiety-reducing skills, communication skills,
drug-resistance skills, perceptions of peer substance
use, weapon-carrying and fighting behavior (control-
ling for gender, family structure, and ethnicity) ac-
counted for 51.0% of the variation in the current use
multiple substance scale (Table 2). Family structure,
ethnicity, critical advertising viewing skills, decision-
making skills, anxiety-reducing skills, communica-
tion skills, and drug-resistance skills were not signif-
icant predictors of current use. Instead the variables
explaining most of the variation included self-effi-
cacy to say ”no,“ positive attitudes toward drug use,
452 BARKIN ET AL. JOURNAL OF ADOLESCENT HEALTH Vol. 30, No. 6

Table 3. Multiple Linear Regression Model of “Anticipated” Drug Use

Variables Beta SE Partial r p value
(Constant) .193 .232 .405
Gender .068 .055 .029 .218
Family Structure ⫺.071 .067 ⫺.025 .287
Ethnicity ⫺.004 .084 ⫺.006 .070
Current drug use .401 .014 .554 .0001
Self-efficacy to say “no” .312 .017 .409 .0001
Critical ad viewing skills ⫺.001 .006 ⫺.004 .881
Positive attitude toward drugs .021 .005 .104 .0001
Critical decision-making ⴚ.016 .008 ⴚ.051 .033
Anxiety-reducing skills .0001 .006 .001 .970
Communication skills .016 .011 .036 .133
Drug-resistance skills .036 .007 .118 .0001
Perception of peer substance use .015 .007 .028 .245
Weapon-carrying behavior .030 .028 .050 .034
Fighting behavior .038 .025 .064 .008
Model Adjusted R2: 73.9%
SE ⫽ standard error.

females using multiple substances less often than
males, perception that peers were using substances,
carrying a weapon, and fighting sometime in the past
12 months. When a reduced multiple linear regres-
sion model was tested including these variables and
controlling for family structure and ethnicity, the
model continued to account for 40.9% of the varia-
tion in current substance use. Although gender was
statistically significant, it explained little of the vari-
ation in the model.
A regression model containing current drug use,
self-efficacy to say ”no“ to drugs, critical advertising
viewing skills, positive attitudes toward drug use,
critical decision-making skills, anxiety-reducing
skills, communication skills, drug-resistance skills,
perception of peer substance use, weapon-carrying,
and fighting behavior (controlling for gender, ethnic-
ity, and family structure) accounted for 73.9% of the
variance in anticipated substance use (Table 3). Family
structure, ethnicity, and gender were not significant
predictors of anticipated use. The statistically signifi-
cant variables in the model included current sub-
stance use, self-efficacy to say, ”no,“ positive atti-
tudes toward substance use, drug-resistance skills,
critical decision-making skills, and weapon-carrying
and fighting behavior. When a reduced model was
tested using these significant variables and control-
ling for gender and family structure, the model
continued to account for 72.3% of the variation in
anticipated substance use.
Discussion
The 1998 National Household Survey on Drug
Abuse found that for youth aged 12–13 years, 9.7% of
students reported cigarette use in the past 30 days
[17]. Our study found a much higher rate (14.1%) of
cigarette use. However, our finding, that about 19%
of seventh graders used alcohol in the past month, is
similar to the findings from the National Household
Survey on Drug Abuse. This survey also indicated
that 12–13 year olds had a 3% lifetime prevalence of
marijuana use and inhalant use, with a 0.1% preva-
lence of cocaine use [16]. By contrast, our data
indicated that about 7% of seventh graders used
inhalants and 0.6% used cocaine. Additionally, our
findings support the theory that high-risk behaviors
cluster [4]. We found that substance use had a strong
association with weapon-carrying and physical fight-
ing. Subsequently, for prevention programs, ad-
dressing related behaviors in addition to substance
use might be useful.
It is no surprise that the strongest association of
anticipated use is current use. This finding highlights
the importance of intervening before adolescents
choose to use substances. Our data indicate that
seventh grade students have significant variability in
self-efficacy skills to aid in drug-resistance. Self-
efficacy to say ”no“ is a particularly important skill
that has a strong association with current use and
anticipated use. Broadening the student’s skill set to
include other drug-resistance skills was significantly
associated with anticipated use. Although students
recognize the importance of these skills, implement-
ing them is not as simple as acquiring them. This is
where knowledge and behavior are not closely re-
lated. Our findings, in accordance with prior studies
[13–15], indicate that teaching self-efficacy skills in
addition to building a broader knowledge base
June 2002
might be warranted. Prior curricula have empha-
sized teaching students the health effects of using
substances. Our data reinforce Botvin’s research and
emphasizes the importance of changing the ap-
proach to focus on changing behavior rather than
knowledge [13–15].
Moreover, students who reported a more positive
attitude toward substance use were more likely to
report current and anticipated use. Changing atti-
tudes is also not as simple as arming students with
more health-related knowledge. Many studies indi-
cate that attitudes and behavior are closely related,
but much less strongly associated with knowledge.
This may prove to be our greatest challenge. Social
modeling from peers and family plays a strong role
in what adolescents believe are normative behaviors
[1,3,5– 8,18 –22]. Our data indicate that adolescents
perceive that their peers are using substances more
than the students’ reports actually demonstrate. Ei-
ther there is socially desirable reporting from respon-
dents or perceptions do not match reality. In the
latter case, it is possible that students perceive that
substance use is more common and, therefore, con-
sidered normative behavior. It would be difficult to
determine if perceptions determine use or if use
determines perceptions. Only 10% of students re-
ported viewing substance use as “grown-up,”
“cool,” “a way to have more fun,” or “make more
friends.” So, perhaps, it is not so much the image of
being “cool” that drives adolescents to use sub-
stances, but the idea that everyone else is doing it
and they do not want to be left out of what “normal”
adolescents do. More research needs to be conducted
to understand what defines students’ perceptions
and shapes their attitudes toward substance use.
Although many prior studies find a strong asso-
ciation between substance use and gender, ethnicity,
and one-parent households [5– 8,17,18], our findings
indicate that gender, ethnicity, and family structure
play only a minor role in understanding adolescent
substance use. Clearly, these elements are immuta-
ble. Our findings indicate that more alterable ele-
ments, such as self-efficacy skills, seem to make the
most difference in determining current and antici-
pated drug use.
Our study had limitations. We used a cross-
sectional design; therefore, only associations could
be explored and no claim to causality could be made.
Also, because adolescents only completed surveys,
we were unable to validate their reports. The YRBS
uses this same methodology and is considered to be
a standard approach to gather information on ado-
lescent high-risk behaviors [2].
YOUNG ADOLESCENTS AND SUBSTANCE USE 453
Conclusions
In today’s world where social acceptance of drug use
is common, teaching adolescents drug-resistance
skills, emphasizing self-efficacy to say “no,” and
enhancing their critical decision-making skills might
successfully reduce the level of illegal substance use
by adolescents. Because younger adolescents are
already beginning to use substances, prevention
needs to occur early and focus on enhancing skills
that will directly affect behavior.
References
1. Irwin CE, Igra V, Eyre S, Millstein S. Risk-taking behavior in
adolescents: The paradigm. Ann N Y Acad Sci 1997;817:1–35.
2. Kann L, Kinschen SA, Williams BI, et al. Youth Risk Behavior
Surveillance — United States 1999. MMWR CDC Surveill
Summ 2000;49:1–96.
3. DuRant RH, Smith JA, Kreiter SR, Krowchuk DP. The
relationship between early age of onset of initial substance
use and engaging in multiple health risk behaviors among
young adolescents. Arch Pediatr Adolesc Med 1999;153:
286 –91.
4. Jessor R. Risk behavior in adolescence: A psychosocial frame-
work for understanding and action. J Adolesc Health 1991;12:
597– 605.
5. DuRant RH, Altman D, Wolfson M, et al. Exposure to violence
and victimization, depression, substance use and the use of
violence by young adolescents. J Pediatr 2000;137:707–13.
6. DuRant RH, Knight J, Goodman E. Factors associated with
aggressive and delinquent behavior among patients attending
an adolescent medicine clinic. J Adol Health 1997;21:303– 8.
7. Woods ER, Lin YG, Middleman A, et al. The associations of
suicide attempts in adolescents. Pediatrics 1997;99:791– 6.
8. DuRant RH, Krowchuk DP, Kreiter S, et al. Weapon-carrying
on school property among middle school students. Arch
Pediatr Adolesc Med 1999;15:21– 6.
9. Hayes CD. Risking the Future: Adolescent Sexuality, Preg-
nancy and Childbearing. Washington, DC: National Academy
Press, 1987.
10. Mosher WD, McNally JW. Contraceptive use at first premar-
ital intercourse: United States, 1965–1985. Fam Plann Perspect
1991;23:108 –28.
11. Escobedo LG, Reddy M, DuRant RH. Relationship between
cigarette smoking and health risk and problem behaviors
among US adolescents. Arch Pediatr Adolesc Med 1997;151:
66 –71.
12. Jackson C, Henriksen L, Dickinson D, Levine D. The early
use of alcohol and tobacco: Its relation to children’s com-
petence and parent’s behavior. Am J Public Health 1997;87:
359 – 64.
13. Botvin GJ, Dusenbury LS, Botvin EM, Diaz T. Long-term
follow-up results of a randomized drug abuse prevention trial
in a white population. JAMA 1995;273:1106 –112.
14. Botvin GJ, Baker E, Filazzola A, Botvin EM. A cognitive-
behavioral approach to substance abuse prevention: A one-
year follow-up. Addict Behav 1990;15:47– 63.
15. Botvin GJ, Baker E, Dusenbury L, et al. Preventing adolescent
drug abuse through a multimodal cognitive-behavioral ap-
454 BARKIN ET AL.
proach: Results of a three-year study. J Consult Clin Psychol
1990;58:437– 46.
16. Bandura A. A Social Cognitive Approach to the Exercise of
Control over AIDS Infection in Adolescents and AIDS: A
Generation in Jeopardy. Newbury Park, CA: Sage Publica-
tions, 1992.
17. National Household Survey on Substance Abuse: Main Find-
ings 1998. DHHS Pub. No. (SMA) 98-3333. Rockville, MD:
Substance Abuse and Mental Health Services Administration,
1998.
18. Ellickson PL, McGuigan KA. Early predictors of adolescent
violence. Am J Public Health 2000;90:566 –72.
JOURNAL OF ADOLESCENT HEALTH Vol. 30, No. 6
19. Greening L, Stoppelbein L. Young drivers’ health attitudes
and intentions to drink and drive. J Adolesc Health 2000;27:
94 –101.
20. Huesmann LR, Guerra NG. Children’s normative beliefs
about aggression and aggressive behavior. J Pers Soc Psychol
1997;72:408 –19.
21. Barkin S, Kreiter S, DuRant RH. Exposure to violence and
intentions to engage in moralistic violence during early ado-
lescence. J Adolesc 2001;24:777–90.
22. DuRant RH, Treiber F, Goodman E, Woods ER. Intentions to
use violence among young adolescents. Pediatrics 1996;98:
1104 – 8.