Skip to main content
NCBI home page
Nicotine & Tobacco Research logoLink to Nicotine & Tobacco Research
. 2014 Mar 3;16(6):855–863. doi: 10.1093/ntr/ntu018

Influences of Tobacco Advertising Exposure and Conduct Problems on Smoking Behaviors Among Adolescent Males and Females

Darren Mays 1,✉, Stephen E Gilman 2, Richard Rende 3, George Luta 4, Kenneth P Tercyak 1, Raymond S Niaura 1,5,6
PMCID: PMC4015102  PMID: 24590388

Abstract

Introduction:

Adolescents with conduct problems are more likely to smoke, and tobacco advertising exposure may exacerbate this risk. Males’ excess risk for conduct problems and females’ susceptibility to advertising suggest gender-specific pathways to smoking. We investigated the associations between gender, conduct problems, and lifetime smoking and adolescents’ exposure to tobacco advertising, and we examined prospective relationships with smoking behaviors.

Methods:

Adolescents completed baseline (2001–2004; n = 541) and 5-year follow-up (2007–2009; n =320) interviews for a family study of smoking risk. Baseline interviews assessed conduct problems and tobacco advertising exposure; smoking behavior was assessed at both timepoints. Generalized linear models analyzed gender differences in the relationship between conduct problems, advertising exposure, and smoking behavior at baseline and longitudinally.

Results:

At baseline, among males, conduct problems were associated with greater advertising exposure independent of demographics and lifetime smoking. Among females at baseline, conduct problems were associated with greater advertising exposure only among never-smokers after adjusting for demographics. In longitudinal analyses, baseline advertising exposure predicted subsequent smoking initiation (i.e., smoking their first cigarette between baseline and follow-up) for females but not for males. Baseline conduct problems predicted current (i.e., daily or weekly) smoking at follow-up for all adolescents in adjusted models.

Conclusions:

The findings of this study reinforce that conduct problems are a strong predictor of subsequent current smoking for all adolescents and reveal important differences between adolescent males and females in the relationship between conduct problems, tobacco advertising behavior, and smoking behavior. The findings suggest gender-specific preventive interventions targeting advertising exposure may be warranted.

INTRODUCTION

Preventing adolescent smoking is a public health priority in the United States (U.S. Department of Health & Human Services [U.S. DHHS], 2012). Consistent with socio-ecological frameworks for health behaviors, adolescent smoking is influenced by factors at the individual level and within the broader social environment (Turner, Mermelstein, & Flay, 2004). In the social environment, tobacco advertising is a robust predictor of adolescent smoking. Combined, the five major U.S. cigarette manufacturers spent nearly 10 billion dollars in 2010 on advertising and promotions (U.S. DHHS, 2012), and prospective studies provide strong evidence that tobacco advertising exposure increases the likelihood that young people will smoke (Lovato, Watts, & Stead, 2011; U.S. DHHS, 2012).

At the individual level, conduct problems (e.g., oppositional and aggressive behaviors) are also a strong predictor of adolescent smoking (McMahon, 1999). Prospective studies demonstrate conduct problems early in life predict subsequent smoking behavior beyond the effects of other risk factors (Fergusson, Horwood, & Ridder, 2007; Fischer, Najman, Williams, & Clavarino, 2012). Although the reasons for this association are unclear, research suggests cigarette smoking may be part of a constellation of youth risk behaviors that co-occur (Donovan & Jessor, 1985). Moreover, adolescents with a history of problem behaviors may be susceptible to the influence of tobacco advertising, heightening their risk of smoking. Because youth problem behaviors tend to cluster together, a history of conduct problems may indicate a predilection for risk taking, where tobacco advertisements could act as a trigger for smoking (Donovan & Jessor, 1985). Adolescents with conduct problems are prone to poor impulse control and experience comorbidities, such as depression (Miller-Johnson, Lochman, Coie, Terry, & Hyman, 1998; Wolff & Ollendick, 2006). Consequently, the mood-enhancing effects of nicotine may offer positive rewards for young people with conduct problems, potentially increasing the likelihood they will act on ads (Weinstein, Mermelstein, Shiffman, & Flay, 2008). Finally, conduct problems may also be markers of social pathways of increased risk; young people with a history of conduct problems may be more likely to be immersed in social environments that are saturated with tobacco advertisements and are conducive to smoking (Turner et al., 2004). Currently, research examining the degree to which adolescents with a history of conduct problems are exposed to tobacco advertising and how this exposure may influence subsequent smoking remain limited.

Evidence also points to individual characteristics that may moderate an association between conduct problems, advertising exposure, and smoking behavior. There are substantial differences between males and females in the prevalence and course of conduct problems. Conduct problems are more common among males (Colman et al., 2009; Lahey et al., 2000), but among females, conduct problems may be identified at earlier ages, comorbidities are more common, and the course of conduct problems tends to be more persistent (Keenan et al., 2010; Wolff & Ollendick, 2006). Gender-specific longitudinal studies show conduct problems influence smoking among both males (Odgers et al., 2007) and females (Bardone et al., 1998).

Although adolescent males are more likely to smoke, females have been shown to experience stronger emotional reactions to tobacco ads, especially ads portraying positive emotion (Dirocco & Shadel, 2007; Shadel, Niaura, & Abrams, 2004). Adolescent females have also been shown to identify more strongly with gender-targeted tobacco ads (Shadel et al., 2004), which may influence gender differences in ad reactivity (Amos, Greaves, Nichter, & Bloch, 2012). Finally, adolescents who have already tried smoking tend to react positively to cigarette advertising (Arnett & Terhanian, 1998; Hawkins & Hane, 2000), suggesting experiences with smoking early in life may affect the influence of advertising on further smoking behavior over time. These data point to gender and early smoking experiences as potential moderators of the relationship between conduct problems and exposure to tobacco advertising.

Frameworks for adolescent health promotion emphasize understanding how individual- and environmental-level factors interact to influence risk behaviors, which can ultimately inform targeted preventive interventions (National Research Council, 2009). Guided by this perspective, interventions focusing on youths’ reactions to tobacco advertising could be targeted toward high-risk groups based on an understanding of how conduct problems and advertising exposure jointly influence smoking risk. To date, no study has investigated the relationships among adolescent conduct problems, tobacco advertising exposure, and smoking behavior, including potential gender differences. Our study sought to fill this research gap by examining these relationships among adolescent participants in a prospective family study of smoking risk. We did so by first examining interacting relationships among conduct problems, gender, and lifetime smoking status with adolescents’ exposure to tobacco advertising. We then examined prospective relationships among conduct problems, gender, and tobacco advertising exposure with subsequent smoking behaviors.

METHODS

Setting and Sample

This analysis is based on third-generation participants in the New England Family Study (NEFS) (Abrams, Leslie, Mermelstein, Kobus, & Clayton, 2003; Gilman et al., 2009; Graham et al., 2008). The NEFS was established to interview adult (second-generation) offspring of pregnant women enrolled between 1959–1964 at the Boston, MA and Providence, RI sites of the National Collaborative Perinatal Project, a prospective birth cohort to study the effects of in-utero and early environment on child health (Hardy, 2003). Third-generation participants were enrolled between 12 and 17 years of age as part of an NEFS study on the intergenerational transmission of smoking conducted from 2001 to 2009 (Gilman et al., 2009).

NEFS adults who resided within 100 miles of Providence, RI and had age-eligible children were invited to participate in the third-generation adolescent cohort. In total, 726 eligible adolescents were invited to participate; 559 (72%) enrolled and completed a baseline interview between 2001 and 2004 (Gilman et al., 2009). Of those, 371 adolescents aged 18 and older at the time of follow-up were invited to complete a follow-up interview. In total, 331 completed a follow-up interview (89%) between 2007 and 2009, which occurred by phone on average 5.2 years (SD = 0.66) after the baseline. Written parental consent and adolescent assent were obtained at baseline, and consent was re-established at follow-up. The Brown University Institutional Review Board approved the study protocol.

This study analyzed data from adolescents’ baseline and 5-year follow-up interviews. The baseline analytic sample for this analysis consisted of a total of 541 adolescents with complete data on smoking behavior and tobacco advertising exposure. Among these 541 adolescents, there were 180 singletons (33%), 143 sibling pairs (n = 286, 53%), and 25 sibling triads (n = 75, 14%). The longitudinal sample included 320 participants with complete data at baseline and smoking behavior outcomes at follow-up. Table 1 displays characteristics of participants in the analytic sample at baseline, those who were contacted for follow-up, and those completing a follow-up interview.

Table 1.

Baseline Sample Characteristics by Completion of the 5-Year Follow-Up Interview

Baseline characteristic Baseline sample Contacted for follow-up Follow-up sample
n = 541 n =374 Yes (n =320) No (n = 54)
Gender
 Male 47.1% (255) 46.5% (174) 44.1% (141) 61.1% (33)
 Female 52.9% (286) 53.5% (200) 55.9% (179) 38.9% (21)
Race
 White 85.4% (463) 85.6% (320) 85.3% (273) 87.0% (47)
 Non-White 14.6% (78) 14.4% (54) 14.7% (47) 13.0% (7)
Baseline age (M, SD) 14.0 (1.7) 14.3 (1.6) 14.3 (1.6) 14.2 (1.6)
Parents’ marital status
 Married 72.1% (390) 70.7% (118) 72.8% (233) 72.2% (39)
 Unmarried 27.9% (151) 29.3% (49) 27.2% (87) 27.8% (15)
Parents’ educational attainment
 Greater or equal to college education 19.6% (106) 19.5% (73) 17.8% (57) 29.6% (16)
 Less than college education 80.4% (435) 80.5% (301) 82.2% (263) 70.4% (33)
Household income
 ≤$60,000/year 45.6% (247) 45.4% (170) 45.9% (147) 42.6% (22)
 >$60,000/year 54.9% (294) 54.5% (204) 54.1% (173) 57.4% (31)
Conduct problems
 Yes 29.8% (161) 32.6% (122) 31.2% (100) 40.7% (22)
 No 70.2% (380) 67.4% (252) 68.8% (220) 69.3% (32)
Advertising exposure (M, SD) 3.4 (1.2) 3.4 (1.2) 3.4 (1.2) 3.8 (1.2)
Lifetime smoking
 Ever smoked 27.5% (149) 28.9% (108) 28.7% (92) 29.6% (16)
 Never smoked 72.5% (392) 71.1% (266) 71.3% (228) 70.4% (38)
Open in a new tab

Note. The original study sample included 559 participants with data on smoking at baseline (Gilman et al., 2008, 2009). A total of 18 participants were excluded from analyses due to missing data on baseline variables analyzed in this study. Some cells do not add up to the total sample size due to sporadic missing data (<5% of participants) for individual variables.

Measures

Demographics

Demographics assessed included age, gender, and race. Family characteristics obtained from parent interview data included household income, parents’ educational attainment, and marital status.

Conduct Problems

Baseline conduct problems were assessed using 16 items from the adolescent version of the Composite International Diagnostic Interview Version 3.0 (Merikangas et al., 2010). This measure has been validated and used in population-based epidemiological studies of adolescent mental health and related comorbidities, including the National Comorbidity Study Replication Adolescent Supplement (Merikangas et al., 2010). Items were introduced by stating “Many kids go through times when they do things adults don’t want them to do, like skipping school, destroying property, and breaking rules. Was there ever a time …” Items capture behaviors such as rule breaking, lying, property damage, and aggression using a yes/no response format. Population-level data show that the prevalence of conduct problems in the general population of adolescents is low (Merikangas et al., 2010). Consistent with these data, in this community-based study, a majority (70%) of participants reported no lifetime conduct problems, and the range of possible conduct problems was strongly truncated (possible range = 0–16, M = 0.48 [SD = 0.89], median = 0, maximum = 5). As a result of this feature of the study sample, lifetime history of conduct problems was operationalized as a dichotomous variable indicating a history of one or more problem behaviors at baseline.

Tobacco Advertising Exposure

Baseline tobacco advertising exposure was assessed using three valid items from the National Youth Tobacco Survey, a survey monitoring youth tobacco use and risk factors (Centers for Disease Control and Prevention, 2012). Items assessed how often adolescents were exposed to smoking advertisements through billboards, magazines, and newspapers with response options for never, a few times a year, every other month, about twice a month, and once a week or more. The items had good internal consistency (Cronbach’s α = 0.76); responses were averaged to create an overall score (range 1–5) with higher values indicating greater advertising exposure. This measure of tobacco advertising exposure was developed and used in the National Youth Tobacco Survey prior to and shortly after the 1998 Master Settlement Agreement between U.S. tobacco companies and 46 states, which restricted certain forms of tobacco advertising, including outdoor billboards. Research conducted shortly after the agreement indicated continued exposure to these advertising channels despite the new restrictions (Celebucki & Diskin, 2002; Chung et al., 2002). Although this measure includes advertising channels that were restricted by the Master Settlement Agreement, the content was valid at the time baseline data were collected when tobacco company advertising practices were rapidly evolving.

Smoking Behavior

The valid and reliable Lifetime Inventory of Smoking Trajectories was administered at each interview to assess adolescent smoking behavior (Colby et al., 2012). For analyses, smoking behaviors were operationalized to reflect developmentally appropriate stages of smoking initiation (i.e., the first cigarette smoked) and progression to regular use from adolescence into young adulthood over the course of the 5-year follow-up period (Turner et al., 2004). At baseline, lifetime smoking was operationalized as whether adolescents reported ever smoking a cigarette (ever smoked/never smoked). In baseline analyses, lifetime smoking was examined to characterize the relationships among smoking status, gender, conduct problems, and exposure to tobacco advertising. At follow-up, two smoking behavior outcomes were analyzed. Smoking initiation was operationalized as participants who had never smoked a cigarette at baseline, but reported smoking by the follow-up (i.e., their first smoking experience occurred during the 5-year follow-up period). For analyses of smoking initiation between baseline and follow-up, the sample was limited to only those adolescents (n = 228) who had never smoked at baseline to prevent misclassification of baseline lifetime smokers as noninitiators at follow-up. Current smoking at follow-up was defined as smoking a cigarette on ≥1 day per week to capture regular use and was analyzed in the full follow-up sample (Turner et al., 2004).

Statistical Analysis

Bivariate tests (e.g., t-tests, chi-square tests, Pearson’s r) were first used to examine associations among baseline characteristics and advertising exposure to inform multivariable analyses. Next, a series of linear regression models was created to examine how conduct problems related to advertising exposure at baseline, including examining potential differences based on gender and lifetime smoking. Models adjusted for demographics that were associated with advertising exposure (p < .10) in bivariate analyses. The initial step was to test a three-way interaction between gender, conduct problems, and lifetime smoking in the baseline sample with and without adjustment. This model included the three-way interaction, all two-way interactions between conduct problems, gender, and lifetime smoking, and main effects. A significant three-way interaction (described in the Results section) indicated gender and lifetime smoking were moderators of the relationship between conduct problems and advertising exposure. Therefore, the final models were stratified by gender.

The next analyses examined whether baseline conduct problems and advertising exposure predicted smoking initiation and current smoking at follow-up and investigated gender differences by testing a similar series of interactions. Demographics associated with the outcomes in bivariate analyses (p < .10) were adjusted in the models. Models were parameterized as logistic regression for the binary smoking behavior outcomes (Gibbons, Hedeker, & DuToit, 2010). Two- and three-way interactions among baseline conduct problems, advertising exposure, and gender were tested for each smoking outcome. Results of the logistic regression models are reported as OR and 95% CIs. Because the NEFS included siblings, all of the models described previously used generalized estimating equations that adjusted standard errors and test statistics to account for correlated observations within families (Liang & Zeger, 1993).

RESULTS

Sample Characteristics

Sample characteristics are shown in Table 1. The baseline sample (n = 541) was nearly half female (53%), most were White race (85%), and the average age was 14 years (SD = 1.7). At baseline, 27% reported lifetime smoking and moderate levels of advertising exposure were evident (M = 3.5, SD = 1.2, range 1 = never to 5 = once a week or more). Nearly 30% (n = 161) reported lifetime conduct problems, which were more common among males (36%) than females (24%, χ 2 = 9.2, p = .002). The most commonly reported conduct problems were intentionally deceiving others (22%), telling lies (19%), breaking curfew (15%), physical fights (13%), skipping school (11%), and intentionally physically hurting someone (11%). Those reporting conduct problems were also more likely to be lifetime smokers at baseline (51% vs. 18%, χ 2 = 62.3, p < .001); however, baseline lifetime smoking did not significantly differ between males (30%) and females (26%, χ 2 = 1.2, p = .27).

At baseline, advertising exposure differed significantly by gender (males: M = 3.5, SD = 1.2; females: M = 3.3, SD = 1.2, t = 2.37, p = .018), conduct problems (lifetime conduct problems: M = 3.7, SD 1.1; no lifetime conduct problems: M = 3.3, SD = 1.2, t = 3.61, p < .001), and lifetime smoking (ever-smokers: M = 3.6, SD = 1.1; never-smokers: M = 3.3, SD = 1.2, t = 2.37, p = .019). Other demographics associated with advertising exposure included race, parents’ marital status, and household income, all of which are shown in Table 2.

Table 2.

Bivariate Associations With Baseline Sample Characteristics, Baseline Advertising Exposure, and Follow-Up Smoking Behaviors

Characteristics Baseline advertising exposure Follow-up smoking initiation Follow-up current smoking
p value Yes No p value Yes No p value
Gender .018 .767 .354
 Male 3.5 (1.2) 44.4% (48) 57.5% (69) 48.3% (42) 42.5% (99)
 Female 3.3 (1.2) 55.6% (60) 42.5% (51) 51.7% (45) 57.5% (134)
Race .002 .386 .324
 White 3.3 (1.2) 88.9% (96) 85.0% (102) 88.5% (77) 84.1% (196)
 Non-White 3.7 (1.0) 11.1% (12) 15.0% (18) 11.5% (10) 15.9% (37)
Age 0.08a .072 14.0 (1.4) 13.9 (1.5) .711 14.6 (1.6) 14.3 (1.5) .301
Parents’ marital status .002 .569 .220
 Married 3.3 (1.1) 77.8% (84) 80.8% (97) 67.8% (59) 74.7% (174)
 Unmarried 3.6 (1.2) 22.2% (24) 19.2% (23) 32.2% (28) 25.3% (59)
Parent’s education .665 .055 .014
 Greater or equal to college education 3.4 (1.1) 17.6% (19) 28.3% (34) 9.2% (8) 21.0% (49)
 Less than college education 3.4 (1.2) 82.4% (89) 71.7% (86) 90.8% (79) 79.0% (184)
Household income .025 .909 .309
 ≤ $60,000/year 3.7 (1.1) 40.7% (44) 40.0% (48) 49.4% (43) 55.8% (130)
 > $60,000/year 3.3 (1.2) 59.3% (64) 60.0% (72) 50.6% (44) 44.2% (103)
Baseline conduct problems <.001 .088 <.001
 Yes 3.7 (1.1) 26.9% (29) 17.5% (21) 49.4% (43) 24.5% (57)
 No 3.3 (1.2) 73.1% (79) 82.5% (99) 50.6% (44) 75.5% (176)
Baseline lifetime smoking .019 N/A <.001
 Ever smoked 3.6 (1.1) – – 55.2% (48) 18.9% (44)
 Never smoked 3.3 (1.2) 100% (108) 100.0% (120) 44.8% (39) 81.1% (189)
Open in a new tab

Note. Baseline: n = 541 and follow-up: n = 320. For smoking initiation, the analysis included only adolescents who had never smoked at baseline (n = 228). Values reflect M (SD) or % (n) unless marked with an a, which indicates Pearson’s r correlation.

Associations With Baseline Advertising Exposure

In the initial models examining factors associated with baseline advertising exposure, the three-way interaction between gender, conduct problems, and lifetime smoking was marginally significant in the unadjusted model (B = 0.85, SE = 0.44, z = 1.94, p = .052) and significant after adjusting for age, race, parents’ marital status, and household income (B = 0.97, SE = 0.43, z = 2.27, p = .024). Gender-specific models are shown in Table 3. For males, the interaction between conduct problems and lifetime smoking was not significant (B = 0.18, SE = 0.32, z = 0.59, p = .557) and was excluded from further analyses. In the final model for males, baseline conduct problems were associated with greater cigarette advertising exposure (B = 0.34, 95% CI = 0.02, 0.67, p = .042) after adjusting for demographic factors and lifetime smoking (Table 3).

Table 3.

Baseline Conduct Problems and Lifetime Smoking as Predictors of Cigarette Advertising Exposure by Gender

Males (n = 255) Females (n = 286)
B (95% CI) p value B (95% CI) p value
Demographics
 Age 0.01 (−0.09, 0.11) .845 0.01 (−0.07, 0.09) .821
 White race (vs. non-White) −0.06 (−0.48, 0.36) .777 −0.46 (−0.83, −0.10) .012
 Parents married (vs. not married) −0.18 (−0.54, 0.18) .331 −0.31 (−0.67, 0.06) .099
 Annual household income > $60,000 (vs. ≤$60,000) 0.16 (−0.18, 0.50) .349 0.12 (−0.21, 0.44) .476
Lifetime smoking (vs. never smoking) 0.08 (−0.28, 0.43) .675 –
Conduct problems (vs. no conduct problems) 0.34 (0.01, 0.67) .042 –
Lifetime smoking × conduct problems interaction – –
Never smoked –
 No conduct problems – Reference
 Conduct problems – 0.48 (0.14, 0.83) .006
Ever smoked –
 No conduct problems – 0.43 (−0.08, 0.94) .101
 Conduct problems – 0.12 (−0.30, 0.53) .581
Open in a new tab

Note. Total n = 541. For males, the interaction between baseline conduct problems and lifetime smoking was nonsignificant (p = .557) and was excluded from the final model shown in the table. Among females, the interaction between conduct problems and lifetime smoking was significant (p = .014), and the interaction term based on lifetime smoking status and history of conduct problems was included in the model shown previously.

For females, the interaction between lifetime smoking and conduct problems was statistically significant (B = −0.79, SE = 0.32, z = −2.47, p = .014), indicating the relationship between conduct problems and advertising exposure differed based on females’ lifetime smoking status. The final model for females suggests that, after adjusting for age, race, parents’ marital status, and household income, female nonsmokers who reported conduct problems had significantly greater exposure to cigarette advertising at baseline (B = 0.48, 95% CI = 0.14, 0.83, p = .006), compared with nonsmokers without conduct problems. This interaction can also be characterized based on differences in adjusted mean baseline advertising exposure among adolescent females based on lifetime smoking status and history of conduct problems. Female never-smokers with conduct problems had significantly greater advertising exposure (M = 3.6, SE = 0.16) compared with never-smokers without conduct problems (M = 3.1, SE = 0.09, p = .006). There were no statistically significant differences in comparison with females with (M = 3.3, SE = 0.18) and without (M = 3.6, SE = 0.24) conduct problems who had ever smoked at baseline (all p’s > .05).

Predictors of Follow-Up Smoking Behavior

At follow-up (n = 320, M age 19.6 years, SD 1.5), 34% (n = 108) of the sample were never-smokers at baseline who had initiated smoking and 27% (n = 87) were current smokers. In bivariate analyses, conduct problems at baseline and parents’ college education were associated with smoking initiation at follow-up, although these associations did not reach statistical significance (Table 2). In the logistic regression models, the two-way interaction between gender and baseline advertising was statistically significant in the model examining smoking initiation (Table 4). The results of this model indicate that among females, greater baseline advertising exposure significantly increased the odds of initiating smoking at follow-up (OR = 1.85, 95% CI = 1.17, 2.94, p = .009), after adjusting for baseline age, parental education, and conduct problems.

Table 4.

Gender, Baseline Conduct Problems, and Advertising Exposure as Predictors of Follow-Up Smoking Behavior

Smoking initiation Current smoking
OR (95% CI) p value OR (95% CI) p value
Demographics
 Female gender 0.13 (0.03, 0.63) .012 0.86 (0.51, 1.46) .585
 Baseline age 0.98 (0.82, 1.18) .862 1.01 (0.85, 1.20) .885
 Parent’s college education (vs. < college) 0.52 (0.26, 1.02) .058 0.42 (0.19, 0.93) .033
Advertising exposure 0.75 (0.54, 1.05) .099 0.97 (0.78, 1.22) .823
Female gender × advertising exposure 1.85 (1.17, 2.94) .009 – ns
Conduct problems (vs. no conduct problems) 1.76 (0.90, 3.42) .099 2.90 (1.67, 5.03) <.001
Open in a new tab

Note. OR = odds ratio; CI = confidence interval; ns = nonsignificant. Logistic regression models with generalized estimating equations examining whether conduct problems and advertising exposure at baseline predict smoking initiation (i.e., first cigarette smoked between baseline and 5-year follow-up) and current smoking (i.e., smoking ≥1 day per week) at 5 years. n = 228 for smoking initiation, which was limited to only participants who were never-smokers at baseline. n = 320 for current smoking, which included all participants completing a 5-year follow-up interview. All two-way interactions between gender, advertising exposure, and conduct problems were tested, along with a three-way interaction between conduct problems, baseline advertising exposure, and gender. Only the interaction between gender and advertising exposure for smoking initiation was statistically significant.

In bivariate analyses, current smoking at follow-up was more common among adolescents whose parents had less than a college education and among those reporting conduct problems (Table 2). In the logistic regression models, none of the two- or three-way interactions involving gender, advertising exposure, and conduct problems were statistically significant, therefore, they were excluded from the model. The final model is shown in Table 4. The odds of current smoking at follow-up were nearly three times higher among adolescents reporting conduct problems at baseline (OR = 2.90, 95% CI = 1.67, 5.03, p < .001) after adjusting for the effects of gender, baseline age, parental education, and advertising exposure. Adolescents whose parents had a college education or greater were significantly less likely to be current smokers (OR = 0.42, 95% CI = 0.19, 0.93, p = .033) after adjusting for the other variables in the model.

DISCUSSION

The findings of this study reinforce existing evidence that adolescent conduct problems are a risk factor for smoking, including progression to regular smoking (McMahon, 1999). The prospective association between baseline conduct problems and current smoking behavior at follow-up is consistent with prior research demonstrating that conduct problems during childhood and adolescence predict cigarette smoking behavior later in life (Fergusson, Horwood, & Ridder, 2007; Fischer et al., 2012). This research is in line with behavioral theory positing that youth problem behaviors, including cigarette smoking, tend to cluster together and may indicate a general predilection toward risk taking (Donovan & Jessor, 1985).

The results of this study advance research on adolescent smoking by demonstrating that a history of conduct problems is associated with greater cigarette advertising exposure and uncovering potential gender differences in this relationship. Although research on this topic is scarce, one possible explanation for the association between conduct problems and advertising exposure observed at baseline is that young people with a history of conduct problems may be drawn to social environments where tobacco advertising is more prevalent. Others have similarly noted that the relationship between cigarette smoking and internalizing symptoms (e.g., anxiety) among young people is likely influenced by adolescent developmental transitions, including exposures in the social environment (Marmorstein et al., 2010). The hypothesis that young people with a history of conduct problems may be drawn to social environments that are more saturated with tobacco advertising warrants further investigation.

Our results also show differences between males and females that suggest a potential increase in risk for smoking among females with a history of conduct problems based on their exposure to tobacco advertising. Adolescent females with conduct problems, particularly those who experience comorbidities, are at an increased risk for a host of adverse health and developmental outcomes (Keenan et al., 2010; Wolff & Ollendick, 2006). With respect to smoking, the joint effects of conduct problems and tobacco advertising exposure among female nonsmokers may increase the risk of smoking, as both risk factors are known predictors of smoking behavior (Fischer et al., 2012; Lovato et al., 2011; U.S. DHHS, 2012).

For females, advertising exposure was greatest among those with conduct problems who had not yet smoked, suggesting an important window of opportunity for targeted interventions to offset the impact of tobacco ads. Prevention efforts could consider screening for these two important risk factors for smoking and targeting intervention approaches accordingly. Approaches such as media literacy (i.e., educating adolescents about manipulative aspects of tobacco advertising) are acceptable to youths and may be impactful to reduce smoking risk if they can be adapted to target gender-specific themes (Primack, Fine, Yang, Wickett, & Zickmund, 2009). Similarly, industry denormalization and countermarketing could be adapted to enhance their impact for male and female adolescents with conduct problems, respectively, particularly given the focus on advertising (Farrelly, Nonnemaker, Davis, & Hussin, 2009; Malone, Grundy, & Bero, 2012). Future research to discern the mediating mechanisms of advertising exposure on smoking behavior will be helpful to identify impactful intervention targets that could be implemented by clinicians and other health professionals working with this high-risk youth population.

We found an association between baseline advertising exposure and subsequent smoking initiation among adolescent females but not among males, which was a somewhat surprising pattern of results. This finding may be because the advertising exposure measure deployed for the study captured a limited range of potentially relevant media. It is possible, for example, that the measure captured advertising channels (e.g., magazines) that more often feature ads targeted toward females, which are more directly noticed by young girls than boys (Pierce et al., 2010; While, Kelly, Huang, & Charlton, 1996). If at the time of the study such ads were more likely to appear in media channels assessed by our measures (e.g., magazines), this could have affected our results. Similarly, if advertising through channels not captured by our measure differentially affect males, such as point-of-sale promotions or event sponsorships (e.g., music, sports), this could have influenced our findings as well (Hafez & Ling, 2006; Henriksen, Feighery, Wang & Fortmann, 2004; Siegel, 2001). Continued investigation into gender differences in adolescent exposure to tobacco advertising, particularly as tobacco companies shift the channels through which they promote their products in light of increasingly stronger advertising restrictions (e.g., Freeman & Chapman, 2010), is an important avenue for future research to clarify these results.

The results of this study should be interpreted in light of important study limitations. All interviews were based on adolescent self-report and may by subject to biases inherent to this data collection method. Examining gender differences in adolescent conduct problems in a community-based research sample is challenging due to the low population-level prevalence of conduct symptoms (Keenan et al., 2010). Although this study used a robust clinically oriented measure of conduct problems, the low prevalence of symptoms prevented more detailed examination of symptom levels, gender-specific features, or clinically diagnosable conduct disorder as predictors of advertising exposure and smoking outcomes. Baseline analyses used cross-sectional data; therefore, temporality in the relationships examined could not be evaluated. The statistical power to detect interactions in prospective analyses may have also been influenced by the sample size at follow-up. Future studies can build from this work by examining the relationships in larger, population-based samples to examine how a range of conduct problems affects smoking.

Finally, as noted earlier, the measure of tobacco advertising exposure used items based on the tobacco media environment prior to the Master Settlement Agreement. Tobacco companies’ advertising and marketing strategies have evolved following the agreement and in response to additional regulations to restrict advertising practices. Integrating measures of tobacco advertising exposure that capture the contemporary media environment including youth-oriented media such as mobile devices and online social networks will be important in future studies (Freeman & Chapman, 2010). In addition, because adolescents who have already tried smoking tend to react positively to cigarette advertising (Arnett & Terhanian, 1998; Hawkins & Hane, 2000), the self-report measure of advertising exposure frequency may be subject to misclassification in baseline analyses of smoking behavior.

Despite these limitations, the results of this study demonstrate that nonsmoking adolescent females with a history of conduct problems report greater exposure to tobacco advertising in their social environments, which may exacerbate their risk of initiating smoking. Advertising exposure predicted subsequent smoking initiation for females, but not males. For all adolescents, conduct problems were a strong predictor of current (i.e., daily or weekly) smoking 5 years later. These findings suggest interventions targeting youths with conduct problems should consider gender differences and highlight the need for research to understand the potential mediating role of emotional reactions to different types of tobacco ads for influencing smoking behavior.

FUNDING

This study was supported by National Institutes of Health Transdisciplinary Tobacco Research Center Award (CA084719). The study was also supported in part by the Biostatistics and Bioinformatics Shared Resource of Georgetown Lombardi Comprehensive Cancer Center through Comprehensive Cancer Center Support Grant (P30CA051008).

DECLARATION OF INTERESTS

None declared.

ACKNOWLEDGMENTS

Portions of this research were presented at the 2013 Annual Meeting of the Society of Research on Nicotine and Tobacco.

The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Cancer Institute or the National Institutes of Health.

REFERENCES

  1. Abrams D. B., Leslie F., Mermelstein R., Kobus K., Clayton R. R.(2003). Transdisciplinary tobacco use research. Nicotine & Tobacco Research, 5, S5–S10. 10.1080/14622200310001625519 [DOI] [PubMed] [Google Scholar]
  2. Amos A., Greaves L., Nichter M., Bloch M.(2012). Women and tobacco: A call for including gender in tobacco control research, policy and practice. Tobacco Control, 21, 236–243. 10.1136/tobaccocontrol-2011–050280 [DOI] [PubMed] [Google Scholar]
  3. Arnett J. J., Terhanian G.(1998). Adolescents’ respones to cigarette advertisements: Links between exposure, liking, and appeal of smoking. Tobacco Control, 7, 129–133. 10.1136/tc.7.2.129 [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Bardone A. M., Moffitt T. E., Caspi A., Dickson N., Stanton W. R., Silva P. A.(1998). Adult physical health outcomes of adolescent girls with conduct disorder, depression, and anxiety. Journal of the American Academy of Child and Adolescent Psychiatry, 37, 594–601. 10.1097/00004583-199806000-00009 [DOI] [PubMed] [Google Scholar]
  5. Celebucki C. C., Diskin K. (2002). A longitudinal study of externally visible cigarette advertising on retail storefronts in Massachusettes before and after the Master Settlement Agreement. Tobacco Control, 11, ii47–ii53. 10.1136/tc.11.suppl_2.ii47 [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Centers for Disease Control and Prevention. (2012). Current tobacco use among middle and high school students--United States, 2011. MMWR Mobidity and Mortality Weekly Report, 61, 581–585 Retrieved from www.cdc.gov/mmwr/preview/mmwrhtml/mm6131a1.htm [PubMed] [Google Scholar]
  7. Chung P. J., Garfield C. F., Rathouz P. J., Lauderdale D. S., Best D., Lantos J. (2002). Youth targeting by tobacco manufacturers since the Master Settlement Agreement. Health Affairs, 21, 254–263. 10.1377/hlthaff.21.2.254 [DOI] [PubMed] [Google Scholar]
  8. Colby S. M., Clark M. A., Rogers M. L., Ramsey S., Graham A. L., Boergers J, … Abrams D. B. (2012). Development and reliability of the lifetime interview on smoking trajectories. Nicotine & Tobacco Research, 14, 290–298. 10.1093/ntr/ntr212 [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Colman I., Murray J., Abbott R. A., Maughan B., Kuh D., Croudace T. J., Jones P. B. (2009). Outcomes of conduct problems in adolescence: 40 year follow-up of national cohort. BMJ, 338, a2981. 10.1136/bmj.a2981 [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Dirocco D. N., Shadel W. G. (2007). Gender differences in adolescents’ responses to themes of relaxation in cigarette advertising: Relationship to intentions to smoke. Addictive Behaviors, 32, 205–213. 10.1016/j.addbeh.2006.03.035 [DOI] [PubMed] [Google Scholar]
  11. Donovan J. E., Jessor R. (1985). Structure of problem behavior in adolescence and young adulthood. Journal of Consulting and Clinical Psychology, 53, 890–904. 10.1037/0022-006X.53.6.890 [DOI] [PubMed] [Google Scholar]
  12. Farrelly M. C., Nonnemaker J., Davis K. C., Hussin A. (2009). The influence of the national truth campaign on smoking initiation. American Journal of Preventive Medicine, 36, 379–384. 10.1016/j.amepre.2009.01.019 [DOI] [PubMed] [Google Scholar]
  13. Fergusson D. M., Horwood L. J., Ridder E. M.(2007). Conduct and attentional problems in childhood and adolescence and later substance use, abuse and dependence: Results of a 25-year longitudinal study. Drug and Alcohol Dependence, 88, S14–S26. 10.1016/j.drugalcdep.2006.12.011 [DOI] [PubMed] [Google Scholar]
  14. Fischer J. A., Najman J. M., Williams G. M., Clavarino A. M. (2012). Childhood and adolescent psychopathology and subsequent tobacco smoking in young adults: Findings from an Australian birth cohort. Addiction, 107, 1669–1676. 10.1111/j.1360-0443.2012.03846.x [DOI] [PubMed] [Google Scholar]
  15. Freeman B., Chapman S. (2010). British American Tobacco on Facebook: Undermining article 13 of the global World Health Organization Framework Convention on Tobacco Control. Tobacco Control, 19, e1–e9. 10.1136/tc.2009.032847 [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Gibbons R. D., Hedeker D., DuToit S. (2010). Advances in analysis of longitudinal data. Annual Review of Clinical Psychology, 6, 79–107. 10.1146/annurev.clinpsy.032408.153550 [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Gilman S. E., Martin L. T., Abrams D. B., Kawachi I., Kubzansky L., Loucks E. B, … Buka S. L. (2008). Educational attainment and cigarette smoking: A causal association? International Journal of Epidemiology, 37, 615–624. 10.1093/ije/dym250 [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Gilman S. E., Rende R., Boergers J., Abrams D. B., Buka S. L., Clark M. A, … Niaura R. S.(2009). Parental smoking and adolescent smoking initiation: An intergenerational perspective on tobacco control. Pediatrics, 123, e274–e281. 10.1542/peds.2008–2251 [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Graham A. L., Papandonatos G. D., DePue J. D., Pinto B. M., Borrelli B., Neighbors C. J, … Abrams D. B. (2008). Lifetime characteristics of participants and non-participants in a smoking cessation trial: Implications for external validity and public health impact. Annals of Behavioral Medicine, 35, 295–307. 10.1007/s12160-008-9031-1 [DOI] [PubMed] [Google Scholar]
  20. Hafez N., Ling P. M. (2006). Finding the Kool Mixx: How Brown & Williamson used music marketing to sell cigarettes. Tobacco Control, 15, 359–366. 10.1136/tc.2005.014258 [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Hardy J. B. (2003). The Collaborative Perinatal Project: Lessons and legacy. Annals of Epidemiology, 13, 303–311. 10.1016/S1047-2797(02)00479-9 [DOI] [PubMed] [Google Scholar]
  22. Hawkins K., Hane A. C. (2000). Adolescents’ perceptions of print cigarette advertising: A case for counteradvertising. Journal of Health Communication, 5, 83–96. 10.1080/108107300126786 [DOI] [PubMed] [Google Scholar]
  23. Henriksen L., Feighery E. C., Wang Y., Fortmann S. P. (2004). Association of retail tobacco marketing with adolescent smoking. American Journal of Public Health, 94, 2081–2083. 10.2105/AJPH.94.12.2081 [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Keenan K., Hipwell A., Chung T., Stepp S., Stouthamer-Loeber M., Loeber R., McTigue K. (2010). The Pittsburgh Girls Study: Overview and initial findings. Journal of Clinical Child and Adolescent Psychology, 39, 506–521. 10.1080/15374416.2010.486320 [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Lahey B. B., Schwab-Stone M., Goodman S. H., Waldman I. D., Canino G., Rathouz P. J, … Jensen P. S. (2000). Age and gender differences in oppositional behavior and conduct problems: A cross-sectional household study of middle childhood and adolescence. Journal of Abnormal Psychology, 109, 488–503. 10.1037/0021-843X. 109.3.488 [PubMed] [Google Scholar]
  26. Liang K. Y., Zeger S. L. (1993). Regression analysis for correlated data. Annual Review of Public Health, 14, 43–68. 10.1146/annurev.pu.14.050193.000355 [DOI] [PubMed] [Google Scholar]
  27. Lovato C., Watts A., Stead L. F. (2011). Impact of tobacco advertising and promotion on increasing adolescent smoking behaviours. The Cochrane Database of Systematic Reviews, 10, CD003439. 10.1002/14651858.CD003439.pub2 [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Malone R. E., Grundy Q., Bero L. A.(2012). Tobacco industry denormalisation as a tobacco control intervention: A review. Tobacco Control, 21, 162–170. 10.1136/tobaccocontrol-2011–050200 [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Marmorstein N. R., White H., Chung T., Hipwell A., Stouthamer-Loeber M., Loeber R. (2010). Associations between first use of substances and change in internalizing symptoms among girls: Differences by symptom trajectory and substance use type. Journal of Clinical Child and Adolescent Psychology, 39, 545–558. 10.1080/15374416.2010.486325 [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. McMahon R. J. (1999). Child and adolescent psychopathology as risk factors for subsequent tobacco use. Nicotine & Tobacco Research, 1, S45–S50. 10.1080/14622299050011801 [DOI] [PubMed] [Google Scholar]
  31. Merikangas K. R., He J. P., Burstein M., Swanson S. A., Avenevoli S., Cui L, … Sendsen J. (2010). Lifetime prevalence of mental disorders in U.S. adolescents: Results from the National Comorbidity Survey Replication–Adolescent Supplement (NCS-A). Journal of the American Academy of Child and Adolescent Psychiatry, 49, 980–989. 10.1016/j.jaac.2010.05.017 [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Miller-Johnson S., Lochman J. E., Coie J. D., Terry R., Hyman C. (1998). Comorbidity of conduct and depressive problems at sixth grade: Substance use outcomes across adolescence. Journal of Abnormal Child Psychology, 26, 221–232. 10.1023/A:1022676302865 [DOI] [PubMed] [Google Scholar]
  33. National Research Council. (2009). Preventing mental, emotional, and behavioral disorders among young people: Progress and possibilities. Washington, DC: National Academies Press; [PubMed] [Google Scholar]
  34. Odgers C. L., Caspi A., Broadbent J. M., Dickson N., Hancox R. J., Harrington H, … Moffitt T. E. (2007). Prediction of differential adult health burden by conduct problem subtypes in males. Archives of General Psychiatry, 64, 476–484. 10.1001/archpsyc.64.4.476 [DOI] [PubMed] [Google Scholar]
  35. Pierce J. P., Messer K., James L. E., White M. M., Kealey S., Vallone D. M., Healton C. G.(2010). Camle No. 9 cigarette marketing campaign targeted teenage girls. Pediatrics, 124, 619–626. 10.1542/peds.2009-0607 [DOI] [PubMed] [Google Scholar]
  36. Primack B. A., Fine D., Yang C. K., Wickett D., Zickmund S. (2009). Adolescents’ impressions of antismoking media literacy education: Qualitative results from a randomized controlled trial. Health Education Research, 24, 608–621. 10.1093/her/cyn062 [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Shadel W. G., Niaura R., Abrams D. B. (2004). Adolescents’ responses to the gender valence of cigarette advertising imagery: The role of affect and the self-concept. Addictive Behaviors, 29, 1735–1744. 10.1016/j.addbeh.2004.03.042 [DOI] [PubMed] [Google Scholar]
  38. Siegel M. (2001). Counteracting tobacco motor sports sponsorship as a promotional tool: Is the tobacco settlement enough? American Journal of Public Health, 91, 1100–1106. 10.2105/AJPH.91.7.1100 [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. 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. 10.1196/annals.1308.023 [DOI] [PubMed] [Google Scholar]
  40. U.S. Department of Health and Human Services. (2012). Preventing tobacco use among youth and young adults: A report of the surgeon general. Atlanta, GA: Centers for Disease Control and Prevention, Office on Smoking and Health; [PubMed] [Google Scholar]
  41. Weinstein S. M., Mermelstein R., Shiffman S., Flay B. (2008). Mood variability and cigarette smoking escalation among adolescents. Psychology of Addictive Behaviors, 22, 504–513. 10.1037/0893-164X.22.4.504 [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. While D., Kelly S., Huang W., Charlton A. (1996). Cigarette advertising and onset of smoking in children: Questionnaire survey. BMJ, 313, 398–399. http://dx.doi.org/10.1136/bmj.313.7054.398 [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. Wolff J. C., Ollendick T. H. (2006). The comorbidity of conduct problems and depression in childhood and adolescence. Clinical Child and Family Psychology Review, 9, 201–220. 10.1007/s10567-006-0011-3 [DOI] [PubMed] [Google Scholar]

Articles from Nicotine & Tobacco Research are provided here courtesy of Oxford University Press

RESOURCES