Health Psychology

© 2019 American Psychological Association 2019, Vol. 38, No. 2, 133–142

0278-6133/19/$12.00 http://dx.doi.org/10.1037/hea0000703

Child and Adolescent Predictors of Smoking Involvement in

Emerging Adulthood

Jennifer M. Jester Jennifer M. Glass

University of Michigan University of Michigan and Eastern Michigan University

Kipling M. Bohnert Joel T. Nigg

University of Michigan Oregon Health and Science University
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
This document is copyrighted by the American Psychological Association or one of its allied publishers.

Maria M. Wong Robert A. Zucker

Idaho State University University of Michigan

Objective: This study examined the differential relationship of externalizing behavior, internalizing
behavior, social context, and their interactions to three developmental indicators of smoking involve-
ment: onset (age), amount of smoking, and dependence symptomatology. Method: Participants (n ⫽ 504,
73% male) from a high-risk community-based longitudinal study were followed from age 12–14 to young
adulthood (18 –20). Smoking involvement was conceptualized as a process involving differences in
(a) age of onset of smoking, (b) amount of smoking at age 18 –20, and (c) level of nicotine dependence
symptomatology at age 18 –20. Survival analysis was used to predict onset of smoking, regression for
smoking level, and zero-inflated Poisson regression for nicotine dependence. Results: Externalizing
(teacher report) and internalizing behavior (youth self-report), prior to the onset of smoking, predicted
different components of smoking and nicotine dependence in young adulthood. Parental smoking
predicted all levels of smoking involvement. Peer smoking was related to early onset of smoking, but not
higher levels of smoking involvement. Externalizing and internalizing behavior interacted to predict
nicotine dependence level, with higher levels of internalizing behavior predicting higher levels of
dependence symptoms, even at low levels of externalizing behavior. Conclusions: Externalizing and
internalizing behavior and social context are independent and interacting risk factors that come into play
at different points in the developmental process occurring between smoking onset and dependence. This
study provides important information for theoretical models of smoking progression and shows that
different types of risk should be targeted for prevention at different points in smoking progression.

Keywords: smoking, emerging adult, nicotine dependence, peer influence, parent influence

Supplemental materials: http://dx.doi.org/10.1037/hea0000703.supp

Young adult cigarette smoking remains a major public health
concern because of the potential for nicotine dependence and
long-term health consequences of chronic smoking. In 2013,
27% of young adults reported smoking cigarettes in the past
month (Substance Abuse and Mental Health Services Admin-
istration, 2014). Although rates of smoking have decreased in
recent years, according to the Surgeon General “Smoking re-
mains the leading preventable cause of premature disease and
death in the United States” (U.S. Department of Health and
Human Services, 2014). Understanding the risk factors associ-
ated with young adult smoking and the progression to depen-
dence is essential for improvement of early prevention pro-
gramming, and also may be useful in clarifying the mechanistic
underpinnings guiding the behavior.
Nicotine dependence is an important outcome because of its
relation to the ability to quit smoking (Breslau, Johnson, Hiripi, &

Jennifer M. Jester, Department of Psychiatry, University of Michigan;
Jennifer M. Glass, Department of Psychiatry, University of Michigan, and
Office of Research Development and Administration, Eastern Michigan
University; Kipling M. Bohnert, Department of Psychiatry, University of
Michigan; Joel T. Nigg, Department of Psychiatry, Oregon Health and
Science University; Maria M. Wong, Department of Psychology, Idaho
State University; Robert A. Zucker, Department of Psychiatry, University
of Michigan.
This article is based on a longitudinal study. The analyses in this article
are unique contributions. This research was supported by grant DA021032
to Jennifer M. Glass from the National Institute on Drug Abuse and grants
AA007065 and AA012217 to Robert A. Zucker from the National Institute
on Alcohol Abuse and Alcoholism.
Correspondence concerning this article should be addressed to Jennifer
M. Jester, Department of Psychiatry, University of Michigan, 4250 Plym-
outh Road, Ann Arbor, MI 48109. E-mail: jjester@umich.edu

133
 134 JESTER ET AL.
Kessler, 2001). However, some risk factors have stronger associ-
ations with nicotine dependence than with smoking level and vice
versa. For example, negative affect is frequently implicated in
nicotine dependence and cessation, but not always found to be a
strong predictor of onset or amount of smoking (Baker, Brandon,
& Chassin, 2004; Chassin, Presson, Pitts, & Sherman, 2000).
Questions remain about whether the same factors operate at dif-
ferent levels of involvement, particularly regarding the interplay of
externalizing and internalizing behaviors (Audrain-McGovern,
Nigg, & Perkins, 2009). This study addresses risk factors (exter-
nalizing behavior, internalizing behavior, and social context) pre-
dicting different levels of involvement (smoking onset, level of
smoking, and nicotine dependence).
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
This document is copyrighted by the American Psychological Association or one of its allied publishers.
Externalizing Behavior
The link between childhood externalizing behavior and smoking
is well-established (Audrain-McGovern et al., 2009; Kollins, Mc-
Clernon, & Fuemmeler, 2005; Whalen & Henker, 1999). A major
component of externalizing behavior involves rule breaking. Due
to societal proscription of young people smoking, especially in
school, rule breaking is expected to be associated with earlier onset
of smoking. The rule-breaking pathway would also support con-
tinued and heavier use, which would likely lead to nicotine de-
pendence as well. Externalizing behavior is hypothesized to be a
sustained component of smoking progression.
Negative Affect/Internalizing Behavior
The parallel to externalizing behavior is internalizing behavior
(acting or feeling anxious, depressed, fearful). The underlying
mechanism for smoking risk is believed to be the higher order trait
negative affect, which reflects the tendency to experience anxiety,
sadness, fear, and anger. In the present study, a truncated version
of this trait, internalizing behavior, was utilized to exclude anger,
which is an externalizing behavior. This allows us to study both
behaviors using similar methods. The Achenbach Youth Self-
Report (Achenbach, 1991a) assessed internalizing behavior, which
includes withdrawn, somatic complaints and anxious/depressed
symptomatology. Internalizing behavior, specifically depression,
has been associated with smoking, although the relationship ap-
pears weaker than that of externalizing behavior (Audrain-
McGovern et al., 2009; Baker, Brandon, et al., 2004; Whalen,
Jamner, Henker, & Delfino, 2001). Weaker effects may be due to
concurrent protective effects. For example, anxiety may protect
against smoking in early adolescence, due to fear of negative
consequences and increased social withdrawal, leading to fewer
interactions with substance-using peers (Colder et al., 2013).
The relationship between depression and smoking has been
explained by self-medication to relieve depressed affect (Sher,
1991). Indeed, heavy smokers report that controlling negative
moods is an important motivation for smoking (Audrain-
McGovern, Leventhal, & Strong, 2015; Baker, Piper, McCarthy,
Majeskie, & Fiore, 2004). Self-medication effects appear to
emerge in late adolescence or early adulthood (Colder, Chassin,
Lee, & Villalta, 2010). Internalizing behavior is hypothesized to be
related to volume of smoking and increased probability of devel-
oping addiction. In addition, youth who display high levels of both
internalizing and externalizing behavior may be at highest risk,
because in addition to the underlying proneness to negative feel-
ings, they may develop increased internalizing behaviors (in par-
ticular, depression and demoralization) secondary to consequences
of externalizing behavior (Hussong, Jones, Stein, Baucom, &
Boeding, 2011).
From the preceding argument, internalizing and externalizing
behaviors were hypothesized to interact to amplify risk for all
aspects of smoking. Support for this hypothesis comes from stud-
ies that indicate that those with comorbid internalizing and exter-
nalizing behavior are at higher risk for smoking than those with
one or the other disorder (Turner, Mermelstein, & Flay, 2004;
Whalen et al., 2001). However, another study indicates that inter-
nalizing behavior is a protective factor for smoking in adolescence
so that comorbid groups have lower risk than those with either risk
factor alone (Colder et al., 2013).
Social Context
Social context influences most relevant to smoking behavior are
parent and peer smoking (Baker, Brandon, et al., 2004; Turner et
al., 2004). Children of smokers are at greater risk of becoming
smokers themselves (Mayhew, Flay, & Mott, 2000). Parent smok-
ing will affect early offspring smoking due to behavioral modeling
and parent smokers may be more likely to allow their children
opportunities to smoke. For these reasons, in addition to genetic
influences, children of smokers are also likely to develop nicotine
addiction earlier. Chassin and colleagues (2000) found that parent
smoking was most highly related to one particular trajectory of
offspring smoking, characterized by early onset as well as a rapid
rise to heavy use. In addition, twin studies suggest that heritability
is stronger for nicotine addiction than for smoking initiation (Vink,
Willemsen, & Boomsma, 2005). These findings suggest that parent
smoking may be related to early onset, a higher level of smoking,
a higher likelihood of dependence, as well higher levels of depen-
dence. Influence by smoking peers is consistent with findings on
parent effects, although mechanisms are unclear (Kobus, 2003;
Turner et al., 2004).
Gender
Although rates of smoking are similar between young men and
women, the incidence of certain risk factors varies. Attention
Deficit Hyperactivity Disorder (ADHD) is more common among
boys of all ages (Gaub & Carlson, 1997) whereas depression is
more common in girls starting around age 14 (Cyranowski, Frank,
Young, & Shear, 2000). Therefore, interactions between sex and
behavior problems may be important. Moreover, Whalen et al.
(2001) found that depression in boys, but not girls, reduced the risk
for smoking associated with externalizing behavior, implying a
three-way interaction between internalizing behavior, externaliz-
ing behavior, and gender. Studies of ADHD symptoms have also
shown interactions between externalizing behavior and gender
(Galera, Fombonne, Chastang, & Bouvard, 2005). These complex
findings highlight the need to examine sex differences in predict-
ing the different components of smoking behavior (Burke, Loeber,
White, Stouthamer-Loeber, & Pardini, 2007). Therefore, in this
study two- and three-way interactions of internalizing and exter-
nalizing behavior and gender were also examined.
 SMOKING INVOLVEMENT IN EMERGING ADULTHOOD
Current Study
A long-term longitudinal study from childhood through early
adulthood was used to examine these relationships. Onset of any
smoking was assessed between the emergence of adolescence and
young adulthood and level of smoking and symptoms of nicotine
dependence were measured at ages 18 –20 years. Survival analysis
was used to predict onset of smoking. Regression was used to
predict smoking level with variables measured prior to onset of
smoking. Zero inflated Poisson regression models were used to
predict both the presence and level of nicotine dependence at the
same age. This gives a unique opportunity to compare factors that
predict a range of smoking involvement.
Hypotheses based on prior research were that externalizing
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
behavior, peer and parent smoking, and the interaction between
This document is copyrighted by the American Psychological Association or one of its allied publishers.
externalizing and internalizing behavior would be related to all
aspects of smoking involvement, whereas internalizing behavior
would only be related to amount of smoking and development of
dependence.
Method
Participants
Data come from a community-based sample of an ongoing,
long-term, longitudinal study of substance abuse risk. The sample
was designed to overrepresent youth at high risk for substance
abuse and substance use disorder by targeting community families
with fathers who have alcoholism (Zucker et al., 2000). The
overall goal of the larger study was to characterize the develop-
ment of risk for substance use, and once it had begun, to charac-
terize the development of substance using behavior in a population
sample enriched for substance abuse. Highest risk families were
recruited utilizing court records of fathers convicted of drunk
driving (N ⫽ 220 youth). A comparison/control group of families
resided in the same neighborhoods as the court-recruited families
but had no substance abuse history; they were recruited using
door-to-door canvassing (N ⫽ 171 youth). An intermediate risk
group was provided by recruiting all families with a father with an
alcohol abuse/dependence diagnosis found during the community
canvas (N ⫽ 113 youth). The original sample was almost entirely
Caucasian, followed by a minority sample recruited at a later time.
Due to the later start in the study, that sample size is quite limited
and, therefore, only Caucasians were included in this study. For a
more detailed description, see Zucker et al. (2000). This research
was approved by the University of Michigan Medical School
Institutional Review Board (HUM00039806). Written informed
consent was obtained from the participants at each wave of as-
sessment, with parents/caregivers providing consent for underage
participants.
Design
In-home assessments of children and parents began at child
age 3–5 and were repeated every 3 years. In addition, children
were assessed annually on key areas over ages 11–17. In this
paper, behavior at age 12–14 was used to predict onset of
smoking, level of smoking, and nicotine dependence at age
18 –20 years (Wave 6).
135
Information was obtained from parents, teachers, and youth.
Data collection at Wave 6 ranged from 1994 to 2009; however,
80% of the data at Wave 6 was collected from 2001 to 2007.
Dummy variables indicate cohort year of birth: 1975–1980 (n ⫽
14), 1981–1985 (n ⫽ 204), 1986 –1989 (n ⫽ 208); 1990 –1996
(n ⫽ 78).
Measures
Outcome measures. At each assessment, smoking and nico-
tine dependence were measured using the Drinking and Drug
History questionnaire, beginning at age 6 – 8. This questionnaire
was designed specifically for this longitudinal study. It is com-
posed of items drawn from a number of well-used measures, which
together cover the domains of interest to the study (Zucker,
Fitzgerald, & Noll, 1990). At the initial assessment, participants
were asked about lifetime use of cigarettes. At follow-up, partic-
ipants answered questions regarding cigarette use in past 3 years
and past year. Possible responses were: Never, Once or twice,
Occasionally, but not regularly, Regularly for a while, but not
now, or Regularly now. If response was Never, respondents
skipped out of the following question: “How frequently have you
smoked cigarettes during the past 30 days?” These measures are
the same as those used in the highly cited Monitoring the Future
study (Johnston, Bachman, & O’Malley, 1979). Usage was based
on the three items at age 18 –20 years. If smoking in the past 30
days was indicated, quantities were estimated based on the re-
sponse choices: Less than one cigarette per day (coded 0.75); One
to five cigarettes per day (coded 3); About one-half pack per day
(coded 10); About one pack per day (coded 20); About one and
one-half packs per day (coded 30); Two packs or more per day
(coded 40).
For the 17 participants (3.3%) who endorsed smoking in the past
year but not in the past 30 days, and seven who endorsed smoking
in the past 3 years but not the past year (1.3%), smoking was
estimated based on their responses to the stem question (see online
supplemental material).
Nicotine dependence. Nicotine dependence was evaluated by
the Fagerstrom score (Heatherton, Kozlowski, Frecker, & Fager-
strom, 1991), a well-validated instrument that includes three ques-
tions such as smoking first thing in the morning (up to 5 points),
difficulty refraining from smoking in forbidden places (1 point),
and smoking if sick in bed (1 point), and a quantity question
(1 point for one half to one pack a day, 2 points for one to one and
one-half packs, and 3 points for more than one and one-half
packs). Scores were classified by level of dependency (Fager-
strom, Heatherton, & Kozlowski, 1990). Table 1 shows the distri-
bution of Fagerstrom scores for smokers at Wave 6 (18 –20 years).
Reliability for this measure was 0.78 (Cronbach’s alpha).
Smoking onset was calculated by taking the youngest age at
which participants reported any smoking, across all assessments. If
the participant reported smoking in the past 30 days, smoking age
was set to assessment age minus .08; if he or she reported smoking
in the past year, onset age was set to assessment age minus .5; if
smoking reported in past 3 years, onset age was set to assessment
age minus 1.5.
Predictors.
Externalizing behavior problems. Externalizing child behav-
ior at age 12–14 was rated by teachers on the Teacher Report Form
 136 JESTER ET AL.

Table 1 The 54-item self-report Peer Behavior Profile (Bingham,

Level of Nicotine Dependence for Smokers at Age 18 –20
Level of dependence Fagerstrom score Frequency
No dependence symptoms 0 95
Very low dependence 1 10
2 14
Low dependence 3 15
4 9
Medium dependence 5 8
High dependence 6 6
7 4
Very high dependence 8 1
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
Fitzgerald, & Zucker, 1995) provided a measure of peer smoking:
“How many of your friends smoke cigarettes?”. The responses are
Percent Almost none, A few, Half, Most, Almost all. Thirty-five percent of
58.6 the sample reported A few or more friends who smoked. The most
6.23 recent response prior to the respondent’s onset of smoking was
8.6 used. Therefore, if no smoking was reported at age 15–17, the peer
9.3 measure was from this age, otherwise the peer measure from age
5.6
4.9 12–14 was used. Overall, this measure has been found to have
3.7 adequate internal reliability (Cronbach’s alpha ⫽ .78; Trucco,
2.5 Villafuerte, Burmeister, & Zucker, 2017).
.62 Choice of reporter for different types of behavior. Theory
and practical reasons guided the choice of reporter for each type of

behavior problem (see Jester et al, 2005). Self-report is the most

This document is copyrighted by the American Psychological Association or one of its allied publishers.

(Achenbach, 1991c), a widely used and recognized empirical mea-
sure of child behavior problems, with excellent reliability, and
strong face, construct (factor structure), and predictive validity
(Achenbach, 1991b; Ivanova, 2007). Test–retest reliability for 15
days was reported to be 0.92 for externalizing behavior (Achen-
bach, 1991b). Items are rated on a 3-point scale: Not True;
Somewhat or Sometimes True; Very True or Often True. Exter-
nalizing behavior is the sum of the Aggression and Delinquency
subscales. The Aggression subscale (20 items) involves items
focused on antisocial and aggressive behaviors. The Delin-
quency subscale (9 items) involves items reflecting both delin-
quent and rule breaking activity.
Internalizing behavior problems. Negative affect was in-
dexed via the internalizing behavior score of the Youth Self-
Report Form (Achenbach, 1991c), at age 12–14. Internalizing
involves nine items from Withdrawn subscale, nine items from
Somatic complaints subscale, and 14 items from the Anxious/
Depressed subscale. One week test–retest reliability for internal-
izing behavior for children age 11–14 was reported to be 0.67
(Achenbach, 1991c).
Social influence. Mothers and fathers reported cigarette use
on the Drinking and Drug History (Zucker et al., 1990). This was
dichotomized into any/no smoking at child age 12–14. Mother and
father alcoholism were quantified at baseline by the Lifetime
Alcohol Problems Score (Zucker, Davies, Kincaid, Fitzgerald, &
Reider, 1997), a continuous measure that combines measures of
onset, breadth of symptomatology, and percentage of lifetime
affected.
reliable for internalizing behavior by early adolescence, as studied
here via the measure of internalizing symptoms (van der Ende,
Verhulst, & Tiemeier, 2012), but not for externalizing behavior. In
order to have an assessment that would more aptly reflect behavior
outside the family context, teacher reports were used. Parents and
children/youth self-reported on their own smoking.
Data Analysis
The analysis was implemented in Mplus, V.8 (Muthén &
Muthén, 1998 –2007), using a robust maximum likelihood estima-
tor (MLR). Regression models, controlling for sex, age, and co-
hort, were estimated for each predictor of interest. In addition, four
interactions models were also estimated based on theory and prior
research (see Table 2). Variables were centered and the product of
centered terms as well as centered variables entered into the
regression equation.
The Vuong test (Greene, 1994) demonstrated that the zero-
inflated Poisson model (Atkins, Baldwin, Zheng, Gallop, & Neigh-
bors, 2013) was preferred over the Poisson model (Akaike ad-
justed ⫽ 18, p ⫽ .003) for nicotine dependence, a count variable
with a preponderance of zeros. The ZIP regression model includes
two processes. The presence of any dependence symptoms is
governed by a binary distribution that generates structural zeros,
and the count of the number of nicotine dependence symptoms was
a Poisson distribution that generated counts.
Survival analysis was performed to predict onset of smoking
across all assessments. Continuous time survival analysis was

Table 2
Relationship of Predictors to Smoking Onset, Smoking Level and Nicotine Dependence at Age 18 –20

Smoking onset Nicotine dependence Nicotine dependence

Predictor (survival analysis) Smoking level yes/no (Odds ratio) level (Incident rate)

Internalizing 1.009 (ns) .10 (.03) 1.20 (.012) 1.65 (.047)

Externalizing behavior 1.04 (.000) .16 (.005) 1.21.19 (.02) 1.22 (ns)
Externalizing behavior by internalizing 1.001 (ns) .41 (ns) 2.20 (.071) .17 (.029)
Sex 1.065 (ns) ⫺.026 (ns) 1.1 (ns) .86 (ns)
Sex by externalizing behavior 1.009 (ns) .29 (ns) 1.09 (ns) 6.1 (.037)
Sex by internalizing .99 (ns) ⫺.04 (ns) 1.049 (ns) 1.13 (ns)
Externalizing behavior by internalizing by sex 1.00 (ns) .00 (ns) 1.054 (ns) .73 (ns)
Peer smoking 1.21 (.020) ⫺.13 (ns) 1.07 (ns) 1.14 (ns)
Father smoking 1.48 (.005) .11 (.031) 1.17 (.043) 1.58 (.087)
Mother smoking 1.13 (ns) .17 (.001) 1.20 (.014) 1.45 (.098)
Note. For smoking onset, hazard ratios are reported; for smoking level (Column 2) standardized betas are reported; for nicotine dependence yes/no, odds
ratios are reported, and for nicotine dependence level, incident raters are reported. p values are given in parentheses. Bold values signifies p ⬍ .10.
 SMOKING INVOLVEMENT IN EMERGING ADULTHOOD 137

implemented in Mplus, using a proportional hazard model and Results

estimating the baseline hazard parameters. Full information max-
imum likelihood was used to handle missing data for predictors in
the survival analysis (there were no missing values for censoring
or time to onset, due to the way the data were selected).
The present analysis included all children for whom there was
any onset smoking data indicating that they had responded to at
least one report on smoking between ages 12–21 and at least one
of the predictors from age 12–14. Of the 553 possible participants,
49 children who already reported smoking more than one or two
cigarettes in their lifetime at age 12–14 were omitted from the
analysis, leaving a sample size of 504 (353 males and 151 fe-
males). Successful follow-up procedures in this study have limited
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
Because approximately 60% of the youth are still residing with
their parents at age 18 –20, and only 14% report income greater
than $16,000 per year, we report the demographics of the parents.
For the parents, 16% report family income less than $30,000, 21%
between $30K and $50K, 27% between $50K and $75K, and 35%
report income higher than $75,000. About 46% of the parents had
a bachelor’s degree and 15% had a master’s or PhD. Table 3 shows
the number of adolescents smoking at each wave. By age 14, 6.2%
reported some smoking, with only 10 adolescents reporting 10 or
more cigarettes per day. By age 18 –20, however, 38% report some
smoking, with 20% smoking at least half a pack per day. There
was very strong stability in smoking; very few reported starting

the number of drop-outs and missed waves. However, because of

smoking and then quitting by Wave 6 (age 18 –20). Only two
This document is copyrighted by the American Psychological Association or one of its allied publishers.

complexities in the study design, not all children had data available
for all waves. Of the 504 participants, 17.9% were missing smok- adolescents reported smoking more than three cigarettes per day at
ing data at age 18 –20, 11.9% were missing peer smoking, 14% Wave 4 (age 12–14) and quitting by Wave 6. Four adolescents
were missing externalizing behavior, 4.5% were missing mother reported smoking more than three cigarettes per day at Wave 5
smoking, 14.1% were missing father smoking, 11.6% were miss- (15–17) and quitting by Wave 6. The level of nicotine dependence,
ing teacher report of externalizing behavior, and 4.0% were miss- for smokers at Wave 6, is shown in Table 1. Thirty percent of
ing self-report of internalizing behavior. Those missing smoking smokers displayed low levels of dependence and 12% displayed
data did not differ from those not missing on smoking level at age medium to high levels. Although 45% of females and 36% of
15–17, teacher report of externalizing behavior at age 12–14, or males had ever smoked, there were no differences in level of
self-report of internalizing behavior at age 12–14. smoking (average cigarettes per day ⫽ 4.1 for females and 3.7 for
Analyses with nicotine dependence as the dependent variable males, t(494) ⫽ .45, p ⬎ .6, or of dependence (average number of
were restricted to those for whom there was evidence of smoking dependence symptoms ⫽ 0.97 for females and 1.0 for males,
at age 18 –20. Omitting those missing cigarette smoking data as t(462) ⫽ .29, p ⬎ .7) (data not shown).
well as those not reporting smoking at age 18 –20 would have Table 4 provides descriptive statistics for each of the variables
resulted in a sample size of 200 (48% of the sample not missing used in the regression models and zero-order correlations between
smoking data). Therefore, cigarette smoking was imputed at age each of the variables. With regard to parent smoking, 39% of
18 –20 for participants with missing data. If the number of impu- fathers and 34% of mothers reported some smoking when the
tations that resulted in value greater than 0 for cigarette use was children were age 12–17. The correlation between smoking and
less than half of the imputations, then the case was omitted from nicotine dependence is high. Although the nicotine dependence
the analyses. This resulted in a sample size of 270 for analysis of measure contains a quantity of smoking item, the correlation
nicotine dependence (53.4% of the total sample). For the nicotine remains high even when removing this item (r ⫽ .75). The corre-
dependence analysis, multiple imputation with 30 imputations was lation between externalizing behavior and internalizing behavior is
implemented in SAS, to handle missing data for the parameter lower than might be expected because these measures are from
estimates. Trace plots showed evidence of convergence for each of different reporters. In contrast, the correlation is much higher when
the variables, using 500 burn-in iterations. Fully conditional spec- looking at measures that use the same reporter. For instance, the
ification with discriminant function method (FCS discrim) was correlation between internalizing and externalizing behavior is
used to impute categorical variables, including Fagerstrom score, 0.58 at age 12–14 when the measures are both youth self-report.
mother and father smoking.
After creating the imputed data set, the data were broken into 30 Main Effect Predictors of Smoking Involvement
separate data files, one for each imputation. Analysis was done in Table 2 shows the results of the three sets of analyses. Each
Mplus, using TYPE ⫽ IMPUTATION. Mplus generates multiple entry represents one model, as each of the predictors was tested
analyses for each imputation and then combines the results, as per
Rubin (1987). Imputation models included all of the variables in Table 3
the analyses and the parallel measures from the time points earlier Smoking at Each Wave of Data Collection
and later than the predictor and outcomes measures in the impu-
tation. For example, internalizing behavior at age 12–14 is the Percent age Percent age Percent age
predictor variable in the regression equation; therefore, the impu- # Cigarettes 12–14 15–17 18 –20
per day (Wave 4) (Wave 5) (Wave 6)
tation model included internalizing behavior at 12–14 as well as
internalizing at age 9 –11 and internalizing at age 15–17 as auxil- None 93.8 79.5 61.7
iary variables. Any 6.2 20.5 38.3
1 2.3 6.5 11.1
The 504 children were from 266 families: two families had five 3 2.3 4.5 7.1
participants, 10 families had four participants, 41 families had 10 1.3 5.8 8.1
three, 117 had two, and 97 had one child in the study. For each of 20 .33 2.9 9.9
the analyses, TYPE ⫽ COMPLEX in Mplus controlled for the 30 .0 .7 1.6
40 .0 .2 .6
nonindependence of residuals.
 138 JESTER ET AL.

separately. The first column shows the results of the survival

.15 (p ⴝ .001)
.1 (p ⫽ .06)

.16 (p ⴝ .02)
ⴚ.16 (p < .001) ⴚ.21 (p < .001) ⴚ.2 (p < .001) ⴚ.14 (p ⴝ .003) ⴚ.12 (p ⴝ .05)

.09 (p ⴝ .04)

.11 (p ⴝ .05)
Peer smoking
(Wave 4)
analysis predicting onset of smoking across age 12–21. The second

.08 (ns)

.07 (ns)

.00 (ns)
.07 (ns)

.05 (ns)
column shows the prediction of level of smoking at age 18 –20, and
the third and fourth columns show the results for the models
predicting presence and level of nicotine dependence symptoms at
.17 (p < .001)
.19 (p < .001)

.25 (p < .001)

.42 (p < .001)

.21 (p < .001)
.42 (p < .001)
Father smoking

.09 (p ⴝ .05)
(Wave 5)

age 18 –20. The results for the zero-inflated Poisson regression are

.00 (ns)
.07 (ns)
given in terms of odds ratios and incident rates, which are unstan-
dardized effect sizes.
.2 (p < .001)
.24 (p < .001)

.24 (p < .001) .21 (p < .001)

.39 (p < .001) .31 (p < .001)

.3 (p < .001)
Internalizing behavior predicted amount of smoking, in addition
.09 (p ⴝ .05)
smoking
Mother

to presence and level of nicotine dependence, but was not signif-

.00 (ns)
.05 (ns)
icantly related to onset of smoking. Externalizing behavior, on the
other hand, predicted smoking onset and level of smoking in late
.16 (p ⴝ .002)
.12 (p ⴝ .02)

.1 (p ⴝ .04)

adolescence as well as the presence of any nicotine dependence,

Mother alc.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

but did not predict the level of nicotine dependence. Sex did not
.02 (ns)
.06 (ns)

Higher values indicate higher levels for those whose parent smokes. Bold values signifies p ⬍ .05.
This document is copyrighted by the American Psychological Association or one of its allied publishers.

predict any aspect of smoking or dependence. Father and mother

smoking were both strong predictors of smoking across the spec-
.15 (p ⴝ .0007)
.14 (p ⴝ .004)
.21 (p < .001)

trum of involvement. Only mother smoking did not predict onset

Father alc.

of smoking. However, peer smoking only predicted onset of smok-

.03 (ns)

⫺.03 (ns)
.08 (ns)

ing and did not predict any further involvement.

.14 (p ⴝ .002)

Interactive Predictors of Smoking Involvement

.12 (p ⴝ .03)
Age

.0 (ns)

.02 (ns)
.11 (p ⴝ .05) ⫺.05 (ns)
.0 (ns)

Two interactive effects were predictors of nicotine dependence.

There was a significant interaction between externalizing and
internalizing behavior in predicting level of nicotine dependence.
.17 (p ⴝ .03)

ⴚ.33 (p < .001) ⴚ.15 (p ⴝ .04)

This interaction term was probed and plotted in Figure 1. Among

.02 (ns)
ⴚ.2 (p < .001) ⫺.01 (ns)
Sex

those with lower levels of internalizing behavior, higher external-

izing behavior predicted more dependence symptoms. However,
ⴚ.19 (p ⴝ .006)

among those with higher levels of internalizing behavior, exter-

ⴚ.13 (p ⴝ .06)
control
Behav.

nalizing behavior was not related to dependence symptoms. Num-

ber of symptoms was consistently high regardless of the level of
externalizing behavior.
(Youth report) (Teacher report)

In addition, the effect of externalizing behavior on level of

.21 (p < .001)
.2 (p < .001)

.11 (p ⴝ .04)

nicotine dependence was moderated by sex. This is illustrated in

Ext

Figure 2, which shows a stronger relationship between externaliz-

ing behavior and nicotine dependence for males. At low levels of
Descriptive Statistics and Correlations for Variables Used in Models

.17 (p < .001)

externalizing behavior, males had slightly lower dependence;

.81 (p < .001) .11 (p ⴝ .01)

whereas at high levels of externalizing behavior, males had more

Int

dependence symptoms than females. Note that the analysis was

performed with continuous measures of internalizing and external-
dependence

izing behaviors. However, to graphically display the interaction,

Nicotine
Std error (% for binary) Skew

⫺.11

⫺.06
.18
.54
2.2
1.9

1.2

2.5

2.0
b
Higher value indicates higher level in females.
71% male

39% Yes
34% yes
Mean

3.8
1.0

8.7

5.6
5.1

19.7
9.9
9.9

1.4
.34
.09

.27

.36
.04

.04
.09
.08

.03
0–32.3

17.2–22.5
5.3–18.7
9.9–17.5
2.5–7.7

yes/no
yes/no
Range

0–40

0–55
0–8

0–1

1–5
teacher report (t-score)
Externalizing behavior
self-report (t-score)
Nicotine dependence
Internalizing youth

Mother alcoholism
Behavioral control
Cigarettes per day

Father alcoholism
Variables

b
Mother smoking
b
Father smoking
age 18–20

Peer smoking
Table 4

Figure 1. Interaction of internalizing and externalizing behavior for the

Sexa
Age

prediction of nicotine dependence symptoms.

a
 SMOKING INVOLVEMENT IN EMERGING ADULTHOOD
Figure 2. Interaction of externalizing behavior and sex predicts nicotine
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
dependence symptoms.
This document is copyrighted by the American Psychological Association or one of its allied publishers.
behavior problems were categorized into low, medium, and high
levels.
Full Models
Finally, full models were estimated, with all of the main effect
predictors for each of the smoking outcomes. Peer smoking (only
related to onset in the prior analysis) was tested in the model for
onset of smoking with the rest of the predictors. It was not found
to be significant, and was omitted from the table. The results are
shown in Table 5. Internalizing behavior remained a predictor for
smoking level as well as the presence of nicotine dependence, but
not level of dependence. Externalizing behavior, when controlling
for internalizing and parent smoking, predicted onset and level of
smoking. Father smoking remained a predictor of onset of smoking
only, and mother smoking predicted level of smoking when con-
trolling for the other main effect predictors. Figure 3 depicts the
progression of smoking levels and the points at which each pre-
dictor is involved.
Discussion
This study prospectively examined predictors of degree of
smoking involvement in young adulthood. One of the major
strengths of this study is that the predictors were measured prior to
onset of any smoking. Therefore, the predictor variables cannot be
caused by the onset of smoking. This longitudinal study is essential
to elucidate factors that affect smoking progression.
As would be anticipated from the study design, smoking rates in
this sample are higher than national averages. Smoking at age 18
for subjects assessed in the years 2004 –2006 was compared with
data from the Youth Risk Behavior Surveillance System (YRBSS)
2005; 38.7% of 18 year-olds in the sample reported smoking in the
Table 5
Full Models, Including All Main Effect Predictors
Smoking onset
Predictor (survival analysis)
Internalizing .009 (ns)
Externalizing behavior .04 (.000)
Father smoking .31 (.064)
Mother smoking .12 (ns)
139
last 30 days. In the YRBSS data for 2005, 26.9% reported smoking
in the last 30 days (Centers for Disease Control and Prevention,
2005). This is most likely due to the recruitment of a high-risk
sample, accessed by way of the father’s alcoholism.
By young adulthood, nearly 40% of the sample reported some
smoking and about 20% reported half a pack a day of smoking or
more. It was noteworthy that very few adolescents reported begin-
ning smoking and then quitting. Only six out of 410 young
adulthoods reported smoking at some time between age 12 and 17
and then quitting entirely. With regard to dependence symptoms,
59% of those who were smoking at age 18 –20 reported no depen-
dence symptoms, whereas 30% reported very low to low depen-
dence and 12% reported moderate to high levels of dependence.
Externalizing behavior was hypothesized to predict all aspects
of smoking involvement, due to the relationship of rule breaking
with higher levels of smoking (Lynskey & Fergusson, 1995).
Externalizing behavior was related to earlier onset, heavier use in
young adulthood, and a greater likelihood of developing nicotine
dependence. This may be related to higher levels of impulsivity, so
that these young adults are more susceptible to the desire to smoke;
alternatively, once started, they may be more likely to experience
the biological rewards of smoking (Audrain-McGovern et al.,
2009). Impulsivity also would make it more difficult to sustain the
intention to cut back or to quit. Thus, the higher level of smoking
could then lead to presence of dependence symptoms. However,
externalizing behavior did not predict the level of nicotine depen-
dence.
Internalizing behavior was hypothesized to be related to level of
smoking and to higher probability of nicotine dependence symp-
toms. This was supported by the findings, which showed that
internalizing behavior did not predict onset of smoking but pre-
dicted all the later levels of involvement. The lack of relationship
with onset may be related to the protective aspects of internalizing
behavior in early adolescence (Colder et al., 2013), leading to
increased fear of consequences and lower socialization with peers.
However, once a person is exposed to smoking, their level of
internalizing behavior is related to higher levels of smoking, in-
creased likelihood of dependence, as well as a higher level of
dependence symptoms. As described earlier, for many smokers,
self-medication or the attempt to control negative emotion is a
primary motivation to smoke (Baker, Brandon, et al., 2004). Based
on the findings, such associations seem to be relatively robust
throughout the development of dependence.
Internalizing behavior and externalizing behavior interacted to
predict nicotine dependence, although not as hypothesized. Instead
of finding that the highest risk was for individuals high in both
internalizing behavior and externalizing behavior, among those
with lower levels of internalizing behavior there was a stronger
Nicotine Nicotine
Smoking level dependence yes/no dependence level
.088 (.057) .16 (.012) .41 (ns)
.14 (.014) .14 (ns) ⫺.013 (ns)
.006 (ns) .06 (ns) .35 (ns)
.15 (.00) .13 (ns) .37 (ns)
 140 JESTER ET AL.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
Figure 3. Range of effect for predictors of different levels of smoking
This document is copyrighted by the American Psychological Association or one of its allied publishers.
behavior.
relationship between level of externalizing behavior and depen-
dence symptoms. However, those at a high level of internalizing
behavior had high levels of dependence symptoms, and the effect
of externalizing behavior was attenuated. This is similar to the
finding of Whalen et al. (2001), who found that higher levels of
internalizing behavior in boys reduced the risk for smoking among
those with higher levels of externalizing behavior. This finding
suggests that in the presence of higher levels of internalizing
behavior, there are likely to be heavier levels of nicotine depen-
dence regardless of the level of externalizing behavior, emphasiz-
ing that those with higher levels of internalizing behavior are more
vulnerable to nicotine dependence. The hypothesized three-way,
externalizing behavior by internalizing behavior by sex interaction,
based on the literature was not supported.
Based on prior research, parent smoking was expected to be
related to both presence and level of smoking and to addiction.
Father smoking was significantly related to onset of smoking, and
both mother and father smoking were related to all of the higher
levels of involvement. Both mother’s and father’s smoking fit with
the theory of modeling of smoking behavior (for smoking) and
genetic effects (for dependence) that are important in influencing
their children’s smoking behavior (Vink et al., 2005). Moreover,
although peer smoking was expected to be related to presence as
well as level of smoking, it was only related to earlier onset of
smoking. However, this study only has a self-report measure of
peer behavior, which might explain the lack of relationship to
higher involvement in smoking. In addition, the study measure of
amount of contact with peers who smoke only provided an ap-
proximate index since the peer questionnaire only inquired about
proportion of friends who used cigarettes (and did not take into
account how many friends or amount of contact with friends).
Therefore, the results regarding peer influence on smoking may be
attenuated.
Considering all predictors together, the pattern of results was
similar, with externalizing being more strongly related to earlier
characteristics of smoking, and internalizing being more strongly
related to later stages. The effect of parent smoking on dependence
was no longer significant when controlling for internalizing and
externalizing behaviors in adolescence; it only operated as a pre-
dictor of the earlier stages of smoking.
These results have direct implications for prevention strategies
for smoking in young adulthood. On the prevention side, prior to
the onset of smoking behavioral factors as well as family factors
are important and should be targeted. For example, those with
higher externalizing behavior are at higher risk to onset early, as
are those in families with a father who smokes. When use starts
and then becomes more severe, the targets of intervention should
focus on internalizing as well as externalizing behavior. Those
seeking to decrease smoking in adolescents and young adults need
to remain aware that family smoking is an important risk factor.
This study has several other strengths. One, already mentioned,
is its longitudinal nature, allowing us to examine predictors of
smoking progression that were extant prior to the onset of any
smoking. Included in this was the ability to determine peer smok-
ing prior to when the young person began smoking. A second
strength is the availability of a direct measure of smoking by the
parent, rather than needing to make use of the offspring’s report,
which may be biased by the offspring’s smoking involvement. In
addition, the study provided multiple reporters of behavior, includ-
ing parent and teacher as well as self-reports. This is atypical, and
again avoids the bias of having behavioral data confounded with
the reports of the smokers themselves.
One limitation is the high-risk sample of young people, many of
whom have a parent with alcoholism. Given the association of
alcoholism with smoking, the sample involves substantially more
parental smoking than exists in the general population, which leads
to a higher than usual base rate of use among youth. However, this
sampling procedure also provides a significant strength. Due to
higher level of parental smoking, relationships to offspring smok-
ing can more accurately be estimated than would be the case in a
study less well-seeded with parent smokers, where estimates of
these effects would be attenuated.
A second limitation of the study is the substantially fewer girls
than boys. This was an unfortunate legacy of funding priorities that
were in existence when the data were first collected, and which
were skewed in the direction of lack of funding agency interest in
addictive behavior among girls, who would be expected to have
lower base rates of the problem than would boys. Data was
provided by151 females, and 45% of the females reported smok-
ing. The imbalance in males and females leads to higher power to
detect relationships for the males in the study, which may explain
the lack of interaction between sex and negative affect, which was
expected based on the hypotheses. Since the sample included more
males than females, further study is also needed to properly ex-
plore the interesting interaction between externalizing behavior,
internalizing behavior, and sex. Also, due to the recruitment of the
sample from a typical midsize Midwestern city, the original sam-
ple did not contain ethnic minorities. Although a later sample was
recruited to include them, the sample size of participants with
relevant data for this analysis was too low to include.
Although smoking outcomes were self-reported and not con-
firmed by biological measures, the smoking rates were higher than
those reported in larger national samples, as would be expected for
this high-risk sample. In addition, the smoking data were mostly
collected from 2001–2007; therefore, the current trends of less
smoking and the use of e-cigarettes is not accounted for in this
study.
In conclusion, results pinpoint which factors are most important
in designing interventions for smoking in young adulthood. Neg-
ative affect and externalizing behavior are implicated in determi-
nation of both smoking level and the development of dependence;
 SMOKING INVOLVEMENT IN EMERGING ADULTHOOD
therefore, prevention efforts should focus on both these problems.
In addition, parent smoking is a strong predictor of smoking and of
dependence, and the offspring of smokers should be a target of
intense prevention efforts. Internalizing behavior, externalizing
behavior, and social contexts are all risk factors for various com-
ponents of smoking behavior. Furthermore, the interactions of the
three types of risk factors also predict nicotine dependence. These
results provide important information for theoretical models of
smoking progression involving externalizing behavior, internaliz-
ing behavior, social context, and genetics. The fact that different
patterns of predictors relate to how early one begins to smoke, how
much one smokes, and whether one develops significant depen-
dence on cigarettes emphasizes two points that are not always
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
acknowledged in the smoking literature. First, these are different
This document is copyrighted by the American Psychological Association or one of its allied publishers.
types of behavior, and the matrix of relationships that leads to each
of these outcomes is therefore also different. Second, these behav-
iors are occurring at different developmental waypoints and there-
fore would be anticipated to have different predictors.
References
Achenbach, T. M. (1991a). Manual for the Youth Self-Report and 1991
Profile. Burlington: Department of Psychiatry, University of Vermont.
Achenbach, T. M. (1991b). Manual for the Child Behavior Checklist/4 –18
and 1991 Profile. Burlington: Department of Psychiatry, University of
Vermont.
Achenbach, T. M. (1991c). Manual for the Teacher’s Report Form and
1991 Profile. Burlington: Department of Psychiatry, University of Ver-
mont.
Atkins, D. C., Baldwin, S. A., Zheng, C., Gallop, R. J., & Neighbors, C.
(2013). A tutorial on count regression and zero-altered count models for
longitudinal substance use data. Psychology of Addictive Behaviors, 27,
166 –177. http://dx.doi.org/10.1037/a0029508
Audrain-McGovern, J., Leventhal, A. M., & Strong, D. R. (2015). The role
of depression in the uptake and maintenance of cigarette smoking.
International Review of Neurobiology, 124, 209 –243. http://dx.doi.org/
10.1016/bs.irn.2015.07.004
Audrain-McGovern, J., Nigg, J. T., & Perkins, K. (2009). Endophenotypes
for nicotine dependence risk at or before initial nicotine exposure. In
G. E. Swan, T. B. Baker, L. Chassin, D. V. Conti, C. Lerman, & K. A.
Perkins (Eds.), Phenotypes and endophenotypes: Foundations for ge-
netic studies of nicotine use and dependence (pp. 339 – 402). Washing-
ton, DC: U.S. Department of Health & Human Services, National
Institutes of Health.
Baker, T. B., Brandon, T. H., & Chassin, L. (2004). Motivational influ-
ences on cigarette smoking. Annual Review of Psychology, 55, 463– 491.
http://dx.doi.org/10.1146/annurev.psych.55.090902.142054
Baker, T. B., Piper, M. E., McCarthy, D. E., Majeskie, M. R., & Fiore,
M. C. (2004). Addiction motivation reformulated: An affective process-
ing model of negative reinforcement. Psychological Review, 111, 33–51.
http://dx.doi.org/10.1037/0033-295X.111.1.33
Bingham, C. R., Fitzgerald, H. E., & Zucker, R. A. (1995). Peer Behavior
Profile/Peer Activities Questionnaire. Unpublished questionnaire. East
Lansing: Department of Psychology, Michigan State University.
Breslau, N., Johnson, E. O., Hiripi, E., & Kessler, R. (2001). Nicotine
dependence in the United States: Prevalence, trends, and smoking per-
sistence. Archives of General Psychiatry, 58, 810 – 816. http://dx.doi
.org/10.1001/archpsyc.58.9.810
Burke, J. D., Loeber, R., White, H. R., Stouthamer-Loeber, M., & Pardini,
D. A. (2007). Inattention as a key predictor of tobacco use in adoles-
cence. Journal of Abnormal Psychology, 116, 249 –259. http://dx.doi
.org/10.1037/0021-843X.116.2.249
141
Centers for Disease Control and Prevention. (2005). Youth Risk Behavior
Surveillance System (YRBSS). Retrieved from https://www.cdc.gov/
healthyyouth/data/yrbs/index.htm
Chassin, L., Presson, C. C., Pitts, S. C., & Sherman, S. J. (2000). The
natural history of cigarette smoking from adolescence to adulthood in a
Midwestern community sample: Multiple trajectories and their psycho-
social correlates. Health Psychology, 19, 223–231.
Colder, C. R., Chassin, L., Lee, M. R., & Villalta, I. K. (2010). Develop-
mental perspectives: Affect and adolescent substance use. In J. D. Kassel
(Ed.), Substance abuse and emotion (pp. 109 –135). Washington, DC:
American Psychological Association.
Colder, C. R., Scalco, M., Trucco, E. M., Read, J. P., Lengua, L. J.,
Wieczorek, W. F., & Hawk, L. W., Jr. (2013). Prospective associations
of internalizing and externalizing problems and their co-occurrence with
early adolescent substance use. Journal of Abnormal Child Psychology,
41, 667– 677. http://dx.doi.org/10.1007/s10802-012-9701-0
Cyranowski, J. M., Frank, E., Young, E., & Shear, M. K. (2000). Adoles-
cent onset of the gender difference in lifetime rates of major depression:
A theoretical model. Archives of General Psychiatry, 57, 21–27. http://
dx.doi.org/10.1001/archpsyc.57.1.21
Fagerstrom, K. O., Heatherton, T. F., & Kozlowski, L. T. (1990). Nicotine
addiction and its assessment. Ear, Nose, and Throat Journal, 69, 763–
765.
Galéra, C., Fombonne, E., Chastang, J. F., & Bouvard, M. (2005). Child-
hood hyperactivity-inattention symptoms and smoking in adolescence.
Drug and Alcohol Dependence, 78, 101–108. http://dx.doi.org/10.1016/
j.drugalcdep.2004.10.003
Gaub, M., & Carlson, C. L. (1997). Gender differences in ADHD: A
meta-analysis and critical review. Journal of the American Academy of
Child & Adolescent Psychiatry, 36, 1036 –1045. http://dx.doi.org/10
.1097/00004583-199708000-00011
Greene, W. (1994). Accounting for excess zeros and sample selection in
Poisson and negative binomial regression models (NYU working paper
no. EC-94-10). Available at https://ssrn.com/abstract⫽1293115
Heatherton, T. F., Kozlowski, L. T., Frecker, R. C., & Fagerstrom, K. O.
(1991). The Fagerstrom Test for Nicotine Dependence: A revision of the
Fagerstrom Tolerance Questionnaire. British Journal of Addiction, 86,
1119 –1127.
Hussong, A. M., Jones, D. J., Stein, G. L., Baucom, D. H., & Boeding, S.
(2011). An internalizing pathway to alcohol use and disorder. Psychol-
ogy of Addictive Behaviors, 25, 390 – 404. http://dx.doi.org/10.1037/
a0024519
Ivanova, M. Y., Achenbach, T. M., Dumenci, L., Rescorla, L. A.,
Almqvist, F., Weintraub, S., . . . Verhulst, F. C. (2007). Testing the
8-syndrome structure of the child behavior checklist in 30 societies.
Journal of Clinical Child and Adolescent Psychology, 36, 405– 417.
http://dx.doi.org/10.1080/15374410701444363
Jester, J. M., Nigg, J. T., Adams, K., Fitzgerald, H. E., Puttler, L. I., Wong,
M. M., & Zucker, R. A. (2005). Inattention/hyperactivity and aggression
from early childhood to adolescence: Heterogeneity of trajectories and
differential influence of family environment characteristics. Develop-
ment and Psychopathology, 17, 99 –125. http://dx.doi.org/10.10170/
S0954579405050066
Johnston, L. G., Bachman, J. G., & O’Malley, P. M. (1979). Drugs and the
class of ’78: Behaviors, attitudes, and recent national trends. Rockville,
MD: National Institute on Drug Abuse. Retrieved from http://catalog
.hathitrust.org/api/volumes/oclc/6862517.html
Kobus, K. (2003). Peers and adolescent smoking. Addiction, 98(Suppl. 1),
37–55. http://dx.doi.org/10.1046/j.1360-0443.98.s1.4.x
Kollins, S. H., McClernon, F. J., & Fuemmeler, B. F. (2005). Association
between smoking and attention-deficit/hyperactivity disorder symptoms
in a population-based sample of young adults. Archives of General
Psychiatry, 62, 1142–1147. http://dx.doi.org/10.1001/archpsyc.62.10
.1142
 142 JESTER ET AL.
Lynskey, M. T., & Fergusson, D. M. (1995). Childhood conduct problems,
attention deficit behaviors, and adolescent alcohol, tobacco, and illicit
drug use. Journal of Abnormal Child Psychology, 23, 281–302. http://
dx.doi.org/10.1007/BF01447558
Mayhew, K. P., Flay, B. R., & Mott, J. A. (2000). Stages in the develop-
ment of adolescent smoking. Drug and Alcohol Dependence, 59(Suppl.
1), 61– 81. http://dx.doi.org/10.1016/S0376-8716(99)00165-9
Muthén, L., & Muthén, B. (1998 –2007). Mplus user’s guide (5th ed.). Los
Angeles, CA: Author.
Rubin, D. B. (1987). Multiple imputation for nonresponse in surveys. New
York, NY: Wiley. http://dx.doi.org/10.1002/9780470316696
Sher, K. J. (1991). Children of alcoholics: A critical appraisal of theory
and research. Illinois: University of Chicago Press.
Substance Abuse and Mental Health Services Administration. (2014).
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
Results from the 2013 National Survey on Drug Use and Health:
Summary of national findings (NSDUH Series H-48, HHS Publication
This document is copyrighted by the American Psychological Association or one of its allied publishers.
No. SMA 14 – 4863). Rockville, MD: Author.
Trucco, E. M., Villafuerte, S., Burmeister, M., & Zucker, R. A. (2017).
Beyond risk: Prospective effects of GABA receptor subunit alpha-2
(GABRA2) ⫻ positive peer involvement on adolescent behavior. De-
velopment and Psychopathology, 29, 711–724. http://dx.doi.org/10
.1017/S0954579416000419
Turner, L., Mermelstein, R., & Flay, B. (2004). Individual and contextual
influences on adolescent smoking. Annals of the New York Academy of
Sciences, 1021, 175–197. http://dx.doi.org/10.1196/annals.1308.023
U.S. Department of Health and Human Services. (2014). The health
consequences of smoking: 50 years of progress. A report of the Surgeon
General. Atlanta, GA: Author, Centers for Disease Control and Preven-
tion, National Center for Chronic Disease Prevention and Health Pro-
motion, Office on Smoking and Health. Retrieved from https://www
.surgeongeneral.gov/library/reports/50-years-of-progress/full-report.pdf
van der Ende, J., Verhulst, F. C., & Tiemeier, H. (2012). Agreement of
informants on emotional and behavioral problems from childhood to
adulthood. Psychological Assessment, 24, 293–300. http://dx.doi.org/10
.1037/a0025500
Vink, J. M., Willemsen, G., & Boomsma, D. I. (2005). Heritability of
smoking initiation and nicotine dependence. Behavior Genetics, 35,
397– 406. http://dx.doi.org/10.1007/s10519-004-1327-8
Whalen, C. K., & Henker, B. (1999). The child with attention-deficit/
hyperactivity disorder in family contexts. In H. C. Quay & A. E. Hogan
(Eds.), Handbook of disruptive behavior disorders (pp. 139 –155). New
York, NY: Kluwer Academic/Plenum Press. http://dx.doi.org/10.1007/
978-1-4615-4881-2_6
Whalen, C. K., Jamner, L. D., Henker, B., & Delfino, R. J. (2001).
Smoking and moods in adolescents with depressive and aggressive
dispositions: Evidence from surveys and electronic diaries. Health Psy-
chology, 20, 99 –111. http://dx.doi.org/10.1037/0278-6133.20.2.99
Zucker, R. A., Davies, W. H., Kincaid, S. B., Fitzgerald, H. E., & Reider,
E. E. (1997). Conceptualizing and scaling the developmental structure of
behavior disorder: The Lifetime Alcohol Problems Score as an example.
Development and Psychopathology, 9, 453– 471. http://dx.doi.org/10
.1017/S0954579497002125
Zucker, R. A., Fitzgerald, H. E., & Noll, R. B. (1990). Drinking and drug
history. Unpublished instrument. East Lansing: Michigan State Univer-
sity.
Zucker, R. A., Fitzgerald, H. E., Refior, S. K., Puttler, L. I., Pallas, D. M.,
& Ellis, D. A. (2000). The clinical and social ecology of childhood for
children of alcoholics: Description of a study and implications for a
differentiated social policy. In H. E. Fitzgerald, B. M. Lester, & B. S.
Zuckerman (Eds.), Children of addiction: Research, health and policy
issues (pp. 109 –141). New York, NY: Routledge Falmer.
Received December 4, 2017
Revision received September 22, 2018
Accepted September 30, 2018 䡲