Title: Cognitive Mediators and Disparities in the Relation
Between Teen Depressiveness and Smoking
Author: Ritesh Mistry Giridhara R. Babu Tanmay Mahapatra
William J. McCarthy
Reference: DAD 5108
To appear in: Drug and Alcohol Dependence
Received date: 13-9-2013
Revised date: 4-2-2014
Accepted date: 18-3-2014
Please cite this article as: Mistry, R., Babu, G.R., Mahapatra, T., McCarthy,
W.J.,Cognitive Mediators and Disparities in the Relation Between Teen
Depressiveness and Smoking, Drug and Alcohol Dependence (2014),
This is a PDF ﬁle of an unedited manuscript that has been accepted for publication.
As a service to our customers we are providing this early version of the manuscript.
The manuscript will undergo copyediting, typesetting, and review of the resulting proof
before it is published in its ﬁnal form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that
apply to the journal pertain.
Page 1 of 26
Cognitive Mediators and Disparities in the Relation Between Teen Depressiveness and Smoking
Ritesh Mistry1, 2, Giridhara R. Babu3, 4, Tanmay Mahapatra3, William J. McCarthy1
1. Center for Cancer Prevention and Control Research, University of California, Los Angeles
2. Department of Health Behavior and Health Education, University of Michigan
3. Department of Epidemiology, University of California, Los Angeles
4. Indian Institute of Public Health-Hyderabad (Bangalore Wing), Public Health Foundation of India
Department of Health Behavior and Health Education
University of Michigan School of Public Health
1415 Washington Heights
SPH I, Room 3806
Ann Arbor, MI 48109-2029
Page 2 of 26
BACKGROUND: Depressiveness and tobacco use in adolescents are linked, however, there is limited evidence about the cognitive
mediators involved and how the role of mediators may differ by gender and racial/ethnic subgroups. METHODS: We used a
racially/ethnically diverse population-based cross-sectional sample of middle and high school students (n=24,350). Logistic regression
models measured the associations of depressiveness with tobacco smoking status, and whether smoking-related knowledge and
attitudes (KA) and smoking refusal self-efficacy (SE) attenuated the associations indicating preliminary evidence of mediation.
RESULTS: Depressiveness was associated with intention to smoke (OR=2.41; 95% CI=2.22, 2.61), experimental smoking (OR=1.93;
95% CI=1.72, 2.17) and established smoking (OR=1.85; 95% CI=1.57, 2.18). The percent attenuation of these associations due to the
inclusion of smoking-related KA and smoking refusal SE was 58% for intention to smoke (p<0.001), 68% for experimental smoking
(p<0.001) and 86% for established smoking (p<0.001). The association of depressiveness with established smoking did not remain
statistically significant (OR=1.16; CI=0.97, 1.40) after including smoking-related KA and smoking refusal SE. Attenuation was more
pronounced in males and white students. CONCLUSIONS: The results suggest that smoking-related KA and smoking refusal SE
attenuated the relation between depressiveness and smoking, indicating that they may serve as mediators of the link between
depressiveness and smoking. Tobacco use prevention programs targeting teens with the aim of increasing anti-smoking KA and
smoking refusal SE may benefit from addressing depressiveness, particularly by using gender and racially/ethnically tailored
strategies. The cross-sectional nature of the data precludes causal inferences.
KEYWORDS: Adolescent; Smoking; Mediation; Depression; Gender; Race/ethnicity
Page 3 of 26
The rates of adolescent tobacco use in the United States remain unacceptably high. It is estimated that about 20% of
adolescents are current tobacco smokers (i.e., those who smoked on at least 1 day in the past month) and about 7% are frequent
smokers (i.e., those who smoked on 20 or more days in the past month; Centers for Disease Control and Prevention, 2010). Depression
and depressive symptoms have been identified as both antecedents (Fergusson et al., 2003) and consequences (Boden et al., 2010;
Goodman and Capitman, 2000) of adolescents smoking (Chaiton et al., 2009; Steuber and Danner, 2006), and contribute to the
initiation of and transition into regular smoking (Audrain-McGovern et al., 2004; Fergusson et al., 2003). Prolonged feelings of
sadness or hopelessness, common depression symptoms, are linked with adolescent smoking (Mistry et al., 2009). With the prevalence
of current or recent depression among adolescents in the United States at 6% (Costello et al., 2006) and the prevalence of prolonged
recent sadness and hopelessness at 29% (Eaton et al., 2012), it is important to understand potential pathways how depressiveness
impacts smoking, especially because smoking is an important risk factor for major chronic diseases (US Department of Health and
Human Services, 2010).
A number of theories posit that health risk behaviors such as tobacco use are partially determined by cognitive factors such as
perceived risks and benefits of a health risk behavior as well as self-efficacy (Health Belief Model, Social Cognitive Theory) to not
engage in unhealthful behaviors despite other influences. Social Cognitive Theory additionally posits that there is interplay between
cognition (e.g., beliefs, knowledge, attitudes, perceived self-efficacy) and affect (e.g. depressive symptoms) that influences behavior.
The cognitive processes involved during adolescence in the link between depressive symptoms and smoking have been studied
empirically to a limited extent, while further research is required particularly using diverse population-based samples. Studies in adults
Page 4 of 26
suggest that expectations held about whether smoking reduces negative emotions (McChargue et al., 2004; Schleicher et al., 2009) and
self-efficacy to refuse tobacco when offered by others (Kear, 2002) mediate the relation. In addition, the research in adolescent
samples suggests that mediators of the link between depressive symptoms and smoking include social influences such as peer approval
of tobacco use (Ritt-Olson et al., 2005), smoking refusal self-efficacy (Minnix et al., 2011), perceived smoking reward (Audrain-
McGovern et al., 2012) and risk (Rodriguez et al., 2007) as well as outcome expectancies (Spruijt-Metz et al., 2005). These studies
have helped elucidate the potential pathways, however, data are needed about the underlying heterogeneity in the role of potential
mediators across important socio-demographic factors such as gender and race/ethnic identification.
Research that helps to illuminate the cognitive pathways involved in the link between depressive symptoms and smoking in
adolescents could be useful for designing prevention programs. The literature suggests uneven efficacy of affect based adolescent
tobacco use prevention programs (Flay, 2009) and that the efficacy of prevention programs may be moderated by depressive
symptoms (Johnson et al., 2007; Sun et al., 2007). Knowledge of the cognitive pathways involved may help identify which factors to
focus upon and how to design programs to better tailor intervention strategies for gender and race/ethnic subgroups of adolescents
based on whether they are experiencing depressive symptoms.
In this study, we used data from a racially/ethnically diverse large population-based sample to examine the association between
feelings of depressiveness and smoking in adolescents, and tested two cognitive factors (Figure 1) for cross-sectional evidence of
mediation. Though cross-sectional data are not ideal for making causal inferences and testing mediation, the results can provide
hypothesis-generating evidence that could inform further research using longitudinal and experimental designs. We chose to examine
smoking-related knowledge and attitudes (KA) and smoking refusal self-efficacy (SE) as potential mediators because these constructs
Page 5 of 26
are key factors that many tobacco use prevention education programs, particularly school-based program, aim to modify in order to
reduce youth tobacco use risk. In addition, depression and depressive symptoms have been shown to affect information processing
(Schwarz et al., 1991; Trope et al., 2001) and decision-making (Bechara et al., 2000), which may impact skills and knowledge gained
from tobacco use prevention education programs.
We anticipated that prolonged sadness or hopelessness, commonly reported symptoms of depression and depressive symptoms,
would impair the ability of adolescents to process tobacco use prevention information and acquire tobacco refusal skills. As a result,
adolescents who reported depressiveness may have also reported less anti-tobacco knowledge and attitudes as well as lower tobacco
refusal self-efficacy. Hence, we hypothesized that there would be positive associations between depressiveness and adolescent
smoking status (intention to smoke, experimental smoking and established smoking) in the sample, and that these associations would
be attenuated by smoking-related KA and smoking refusal SE indicating evidence of potential mediation. We also hypothesized that
the attenuation in the relation between depressiveness and smoking due to smoking-related KA and smoking refusal SE would differ
by gender and race/ethnicity, because there is substantial differential risk of depressive symptoms and tobacco use between adolescent
males and females (Fryar et al., 2009), and categories of race/ethnicity (Centers for Disease Control and Prevention, 2010; Kafilat
Page 6 of 26
We used population-based cross-sectional survey data from the 2003-2004 California Student Tobacco Survey (CSTS), a
stratified two-stage cluster sampling in-school survey of 25,868 middle and high schools students in 226 California schools
(McCarthy et al., 2008). The CSTS participants were sampled from 12 regions (strata), which were formed on the basis of
county demographic and socioeconomic characteristics. First, schools in each stratum were randomly selected with probabilities
proportional to size of enrollment. Intact classes of required courses were randomly sampled from selected schools. Active parental
consent to participate was obtained, which resulted in a student participation rate of 69.6%, for a total of 24,350 students in our study.
Intention to smoke was measured using the question “Do you think you will smoke tobacco at any time during the next year?”
Having an intention to use tobacco was defined as responses of “Definitely yes” or “Probably yes” and not having an intention to use
was defined as “Definitely no” or “Probably no.” Experimental smoking and established smoking were assessed using two items:
“About how many cigarettes have you smoked in your entire life,” and “During the past 30 days, on how many days did you smoke
cigarettes.” Experimental smoking was defined as having smoked having smoked less than 100 cigarettes in one’s lifetime and having
smoked during the last 30 days. Established smoking was defined as having smoked 100 or more cigarettes in one’s lifetime and
having smoked in the last 30 days.
Depressiveness was assessed using the commonly used United States Youth Behavioral Risk Survey question: “During the last
12 months, did you feel sad or hopeless almost every day for 2 or more weeks?” This is a widely used surveillance measure of
Page 7 of 26
depressive symptoms (Eaton et al., 2012) and has been associated with measures of negative affect such as unrealistic fatalism
(Jamieson and Romer, 2008) and suicidal ideation (Jamieson and Romer, 2008). Smoking-related KA was assessed through
standardized scores on an ad-hoc 12-item scale (Cronbach’s alpha=0.71). The items included in the scale asked about whether
smoking: reduces weight gain; is a way to lose friends; results in shorter lives; makes one appear more grown up; makes one more
relaxed; makes one look cool; cigars are as bad as cigarettes; if those who smoke have more friends, look more cool or fit in; risk
harming themselves if they smoke 1 to 5 cigarettes a day; and whether it is safe to smoke for a year or two as long as one quits. There
were four Likert-type response categories from “Definitely yes” to “Definitely not,” which were coded as 0, 33, 66 or 100 with higher
scores reflecting more accurate knowledge about and more negative attitudes toward tobacco use. An average score across all items
was created and then standardized. Smoking refusal SE was defined as low if the response to the question “How hard would it be to
refuse or to say “no” to a friend who offered you a cigarette?” was “Very hard” or “Hard” and high if the response was “Easy” or
Study covariates included age (categorized as 12 yrs or less, 13-14 yrs, 15-16 yrs and 17 yrs or more), gender, race/ethnicity
(categorized as white, Hispanic, Asian/Pacific Islanders, African American and American Indian/Alaskan Natives), academic
performance (categorized as Superior=Mostly A’s; Good=A’s and B’s; Fair=Mostly B’s, and B’s and C’s; Poor=Mostly C’s or
worse), perceived peer smoking (defined as “Low” if respondents reported that <20% of peers smoke, “Moderate” if 21-60% and
“High” if >61%) and perceived ease of obtaining tobacco (categorized as “Easy” and “Hard”). Figure 1 shows that age, gender,
race/ethnicity and educational performance were expected to be potential confounders of the relation between depressiveness and
smoking (Brunswick and Messeri, 1984; Klungsoyr et al., 2006; Young and Rogers, 1986). Because perceived peer smoking and ease
Page 8 of 26
of access to tobacco are not likely to be associated with depressiveness (Hover and Gaffney, 1988), they were not hypothesized to
cofound the relation between depressiveness and smoking, but were hypothesized to confound the association of smoking-related KA
and smoking refusal SE (i.e., the hypothesized mediators) with smoking outcomes.
First, we conducted descriptive analyses of the study variables. Second, we used the causal steps approach to measure evidence
of mediation (MacKinnon and Fairchild, 2009). The percent attenuation in the association between depressiveness and smoking after
inclusion of the cognitive cofactors (anti-smoking KA and smoking refusal SE) was used to measure preliminary mediation. We used
logistic regression to estimate the association of depressiveness with each smoking measure controlling for age, educational
performance, gender and race/ethnicity (Model Set 1). Because we hypothesized that peer smoking and perceived ease of access to
tobacco would confound the association of smoking status with smoking-related KA and smoking refusal SE, but not likely to be
associated with depressiveness, we did not include them as covariates in Model Set 1. In Model Set 2, we added smoking-related KA,
perceived peer smoking and perceived ease of access to tobacco to the variables included in Model Set 1. In Model Set 3, the fully
controlled models, we added smoking refusal SE to the variables included in Model Set 2. Attenuation after inclusion of smoking-
related KA was measured as percent reductions of the Model Set 1 coefficients for the association between depressiveness and each
smoking measure when compared to the corresponding coefficients in Model Set 2. Similarly, attenuation after inclusion of both
smoking-related KA and smoking refusal SE was obtained as percent reductions in the Model Set 1 coefficients for the association
between depressiveness and each smoking measure, when compared to the corresponding coefficients in Model Set 3. We next
analyzed whether the fully controlled association between depressiveness and smoking status (Model Set 3) and percent attenuation
Page 9 of 26
due to inclusion of smoking-related KA and smoking refusal SE differed based on gender and race/ethnicity by stratifying the
regression analysis by these two demographic variables. Finally, we stratified the analysis by gender and race/ethnicity. Statistical
analysis was conducted using SAS 9.1. Results were adjusted for the survey sampling design, and weighted to represent California
middle and high school students.
Table 1 shows the distribution of study variables and cross-tabulations of depressiveness and smoking status with the
covariates. Nearly 50% of participants were 15 or more years of age, and there were more females (53%) than males (47%); the
gender distribution was similar across race/ethnicity, i.e. 52-55% female and 45-48% male. The participants were mostly white (42%)
or Hispanic (34%), but Asian/Pacific Islanders (16%), African Americans (7%) and American Indians/Alaskan natives (2%) were also
included in the sample. Depressiveness was reported by about 28% of the participants, and was most frequently reported by Hispanics
(30%) and African Americans (30%). Intention to smoke, experimental smoking and established smoking were reported by 15%, 6%
and 3% of the participants, respectively, with Hispanics most frequently reporting intention to smoke (17%) and experimental
smoking (8%) while whites most frequently reporting established smoking (5%).
Table 2 shows results from the logistic regression analysis using Model Sets 1, 2 and 3 as described above and the estimated
attenuation in the association between depressiveness and smoking status due to the inclusion of smoking-related KA and smoking
refusal SE. Results from Model Set 1 show that students who reported depressiveness were at 2.41 times higher odds of reporting
intention to smoke (95% CI: 2.22, 2.61), 1.93 times higher odds of reporting experimental smoking (95% CI: 1.72, 2.17) and 1.85
times higher odds of reporting established smoking (95% CI: 1.57, 2.18). When the measure for smoking-related KA was added to
Page 10 of 26
Model Set 1 the odds ratios for each outcome were attenuated (Model Set 2) by 54% for intention to smoke (p<0.001), 68% for
experimental smoking (p<0.001) and 75% for established smoking (p<0.001). When the measures of both smoking-related KA and
smoking refusal SE (Model Set 3) were added to Model Set 1, the odd ratios were attenuated even further as compared to Model Set 1
for intention to smoke (58%, p<0.001) and established smoking (86%, p<0.001), but not for experimental smoking (68%, p<0.001).
The odds ratio for the relation between depressiveness and established smoking became non-significant in Model Set 3.
Table 3 shows the regression results stratified by gender and race/ethnicity in order to assess whether the fully controlled
associations between depressiveness and smoking outcomes (Model Set 3) and percent attenuation due to smoking-related KA and
smoking refusal SE were modified by these demographic variables. The results show that fully controlled associations of
depressiveness with intention to smoke and established smoking was significantly greater for female versus male students, and
Hispanic students versus white students. This is evident from the odds ratio point estimates of Model Set 3 for intention to smoke and
established smoking for one gender or race/ethnicity being outside of the corresponding 95% confidence intervals for the other gender
The results also show that percent attenuation due to smoking-related KA and smoking refusal SE for intention to smoke and
established smoking was, respectively, 53% and 73% for female students compared to 67% and 100% for males students, and 51%
and 54% for Hispanic students versus 65% and 100% for white students.
This study analyzed a large racially/ethnically diverse population-based sample of middle and high school students to
preliminarily assess mediation by two cognitive factors in the link between depressiveness and tobacco smoking. We examined
Page 11 of 26
smoking-related KA and smoking refusal SE as potential mediators because they may be negatively influence by depressive
symptoms (Schwarz et al., 1991; Trope et al., 2001) and because they are important modifiable factors that many tobacco use
prevention education programs aim to improve (Lantz et al., 2000). The results add to the current literature by finding preliminary
evidence of the mediational roles of smoking-related KA and smoking refusal SE as shown by the attenuation of the associations of
depressive symptoms with smoking status. Prior research suggests that smoking-related KA and smoking refusal SE may act as
mediators (Finney Rutten et al., 2008; Minnix et al., 2011; Rodriguez et al., 2007), however how mediating the roles of these factors
vary across gender and racial/ethnic subgroups has not been examined to our knowledge. This study shows that attenuation due to
smoking-related KA and smoking refusal SE differed based on gender and racial/ethnic identification indicating the these
demographic factors may serve as important moderators. The findings indicate that smoking-related KA and smoking refusal SE may
completely explain away the association between depressiveness and established smoking, and substantially attenuated the association
of depressiveness with intention to smoke and experimental smoking. This suggests good preliminary evidence of mediation.
Consistent with other literature (Waller et al., 2006), the association of depressiveness with intention to smoke and established
smoking was higher in female compared to male students. In Hispanic compared to non-Hispanic students, the association of
depression on intention to smoke and established smoking was also higher, but this was not consistent research in Hispanic adults
(Berg et al., 2012) despite evidence suggesting greater prevalence of depressiveness and smoking from studies in Hispanic populations
(Lorenzo-Blanco et al., 2011) and national data comparing racial/ethnic subgroups (Johnston et al., 2014). Attenuation after the
inclusion of smoking-related KA and smoking refusal SE was more pronounced in male and white students.
Page 12 of 26
The results point towards researching the integration of the management and reduction of negative affective symptoms such as
depressiveness as part of adolescent tobacco use prevention programs, especially those that aim to increase anti-tobacco KA and
smoking refusal SE. For example, identification of students with significant depressive symptoms and referral to mental health
services may enhance tobacco use prevention or other health education efforts. The literature suggests that addressing depression as
part of smoking cessation interventions in adults has positive impacts (Gierisch et al., 2012), however, research with respect to
tobacco use prevention and cessation in adolescent populations with depressive symptoms and depression is needed (DeHay et al.,
2012). Moreover, the results suggest that designing tobacco use prevention strategies to improve anti-smoking KA and smoking
refusal SE should be done in a manner that is salient to adolescents experiencing depressive symptoms and are at high risk for
depression. Such adolescents may require different modes of delivery and content of tobacco use prevention education and skill
building exercises, particularly for male and white adolescents.
Finally, adolescent depression and depressiveness is in itself of public health importance. Effective programs and services for
primary prevention and treatment of depression in adolescents are not readily available, especially for socio-economically
disadvantaged youth. Preventing adolescent depressiveness may reduce the risk of tobacco use initiation (Dudas et al., 2005; Roberts
et al., 2011) and other behavioral risks (Mistry et al., 2009). More research is, therefore, required to better understand the determinants
of adolescent depressiveness and to identify efficacious prevention and treatment strategies.
4.1 Strengths and Limitations
To our knowledge, there is little information in the scientific literature on the possible mediational roles of smoking-related
KA and smoking refusal SE in the pathways linking depressiveness and smoking in adolescent and how they vary across racial/ethnic
Page 13 of 26
and gender strata. This study adds to the current literature by examining these potential mediators using a large racially/ethnically
diverse population-based sample of adolescents. As with any cross-sectional observational study, however, the results from this study
precludes assessment of causal direction, and should be cautiously interpreted including with respect to the amount of mediation.
Nonetheless the results are useful for generating hypotheses about mediation for further testing via prospective and experimental data,
which we suggest as next steps.
The study also aimed to take into account the most relevant risk factors of adolescent tobacco use in the analysis. Although we
could not include all potential confounders, the inclusion of perceived peer smoking and perceived ease of access to tobacco may
serve as proxies for many factors commonly associated with smoking-related KA, smoking refusal SE and smoking status, such as
community norms about tobacco use, social acceptance/stigma around smoking, as well as actual availability, financial access and
social access to tobacco products. By controlling for perceived peer smoking and perceived ease of access to tobacco and relevant
demographic and other covariates, we aimed to minimize confounding. However, as always there remained the possibility of
confounding due to unmeasured variables. We also were limited to using self-reported data on smoking status, which are not ideal, but
the data were collected confidentially to limit biases. The procedures used to collect these data have been widely used and have shown
good validity and reliability when self-administered anonymously within classroom settings (Stanton et al., 1996). Finally, we used
single-item measures of depressiveness, intention to smoke and smoking refusal SE. These, however, are commonly used measures in
national and state level surveillance of youth behavioral risk factors. In addition, the measure of depressiveness has been consistently
associated with mental health outcomes such as suicidal ideation (Jamieson and Romer, 2008) and health risk behaviors (Mistry et al.,
2009). In summary, due to the cross-sectional nature of the data and the limitations due to the measurement of key study constructs,
Page 14 of 26
our results were not conclusive, but hypothesis generating and require further validation through studies using more detailed measures
and rigorous study designs.
Depressiveness was reported by a significant proportion of the sample and was consistently associated with various measures
of smoking status, particularly in female and Hispanic students. The associations between depressiveness and smoking outcomes were
attenuated by smoking-related KA and smoking refusal SE, indicating preliminary evidence of mediation. Despite the limitations of
using cross-sectional data to evaluate mediation, these results suggest that tobacco use prevention programs targeting middle and high
school aged students may benefit from addressing depressiveness, particularly by using gender and racially/ethnically tailored
Page 15 of 26
Figure 1. Conceptual framework of factors involved in the link between depressiveness and smoking
Page 16 of 26
Audrain-McGovern, J., Lerman, C., Wileyto, E.P., Rodriguez, D., Shields, P.G., 2004. Interacting effects of genetic predisposition and
depression on adolescent smoking progression. Am. J. Psychiatry 164, 1224-1230.
Audrain-McGovern, J., Rodriguez, D., Rodgers, K., Cuevas, J., Sass, J., Riley, T., 2012. Reward expectations lead to smoking uptake
among depressed adolescents. Drug Alcohol Depend. 120, 181-189.
Bechara, A., Damasio, H., Damasio, A.R., 2000. Emotion, decision making and the orbitofrontal cortex. Cereb. Cortex 10, 295-307.
Berg, C.J., Kirch, M., Hooper, M.W., McAlpine, D., An, L.C., Boudreaux, M., Ahluwalia, J.S., 2012. Ethnic group differences in the
relationship between depressive symptoms and smoking. Ethn. Health 17, 55-69.
Boden, J.M., Fergusson, D.M., Horwood, L.J., 2010. Cigarette smoking and depression: tests of causal linkages using a longitudinal
birth cohort. Br. J. Psychiatry 196 440-446
Brunswick, A.F., Messeri, P.A., 1984. Origins of cigarette smoking in academic achievement, stress and social expectations: Does
gender make a difference. J. Early Adolesc. 4, 353 -370.
Centers for Disease Control and Prevention, 2010. Cigarette use among high school students—United States, 1991–2009. MMWR 59,
Chaiton, M.O., Cohen, J.E., O'Loughlin, J., Rehm, J., 2009. A systematic review of longitudinal studies on the association between
depression and smoking in adolescents. BMC Public Health 9, 356.
Page 17 of 26
Costello, J.E., Alaattin, E., Adrian, A., 2006. Is there an epidemic of child or adolescent depression? J. Child Psychol. Psychiatry 47,
DeHay, T., Morris, C., May, M.G., Devine, K., Waxmonsky, J., 2012. Tobacco use in youth with mental illnesses. J. Behav. Med. 35,
Dudas, R.B., Hans, K., Barabas, K., 2005. Anxiety, depression and smoking in schoolchildren-implications for smoking prevention. J.
Roy. Soc. Promot. Health 125, 87-92.
Eaton, D.K., Kann, L., Kinchen, S., Shanklin, S., Flint, K.H., Hawkins, J., Harris, W.A., Lowry, R., McManus, T., Chyen, D., Whittle,
L., Lim, C., Wechsler, H., 2012. Youth risk behavior surveillance - United States, 2011. MMWR Surveill. Summ. 61, 1-162.
Fergusson, D.M., Goodwin, R.D., Horwood, L.J., 2003. Major depression and cigarette smoking: results of a 21-year longitudinal
study. Psychol. Med. 33, 1357 – 1367.
Finney Rutten, L.J., Augustson, E.M., Moser, R.P., Beckjord, E.B., Hesse, B.W., 2008. Smoking knowledge and behavior in the
United States: sociodemographic, smoking status, and geographic patterns. Nicotine Tob. Res. 10, 1559-1570.
Flay, B.R., 2009. The promise of long-term effectiveness of school-based smoking prevention programs: a critical review of reviews.
Tob. Induc. Dis. 5, 7.
Fryar, C.D., Merino, M., Hirsch, R., Porter, K.S., 2009. Smoking, alcohol use, and illicit drug use reported by adolescents aged 12-17
years: United States, 1999-2004. Natl. Health Stat. Rep. 1.
Gierisch, J.M., Bastian, L.A., Calhoun, P.S., McDuffie, J.R., Williams, J.W., 2012. Smoking cessation interventions for patients with
depression: a systematic review and meta-analysis. J. Gen. Intern. Med. 27, 351-360.
Page 18 of 26
Goodman, E., Capitman, J., 2000. Depressive symptoms and cigarette smoking among teens. Pediatrics 106, 748-755.
Hover, S.J., Gaffney, L.R., 1988. Factors associated with smoking behavior in adolescent girls. Addict. Behav. 13, 139 -145.
Jamieson, P.E., Romer, D., 2008. Unrealistic fatalism in US youth ages 14 to 22: prevalence and characteristics. J. Adolesc. Health 42,
Johnson, C.A., Cen, S., Gallaher, P., Palmer, P.H., Xiao, L., Ritt-Olson, A., Unger, J.B., 2007. Why smoking prevention programs
sometimes fail. Does effectiveness depend on sociocultural context and individual characteristics? Cancer Epidemiol.
Biomarkers Prev. 16, 1043-1049.
Johnston, L.D., O'Malley, P.M., Miech, R.A., Bachman, J.G., Schulenberg, J.E., 2014. 2013 Overview: Key Findings on Adolescent
Drug Use. Institute for Social Research, Ann Arbor, MI. p. 85.
Kafilat Tolani, J., 2012. Ethnic differences in susceptibility to smoking and intention to smoke on smoking behavior among
adolescents. J. Community Med. Health Educ. 2, 143.Kafilat Tolani, J., 2012. Ethnic differences in susceptibility to smoking
and intention to smoke on smoking behavior among adolescents. J. Community Med. Health Educ.
Kafilat Tolani, J., 2012. Ethnic differences in susceptibility to smoking and intention to smoke on smoking behavior among
adolescents. J. Community Med. Health Educ. 2, 143
Kear, M.E., 2002. Psychosocial determinants of cigarette smoking among college students. J. Community Health Nurs. 19, 245-257.
Klungsoyr, O., Nygard, J.F., Sorensen, T., Sandanger, I., 2006. Cigarette smoking and incidence of first depressive episode: an 11-
year, population-based follow-up study. Am. J. Epidemiol. 163, 421–432.
Page 19 of 26
Lantz, P.M., Jacobson, P.D., Warner, K.E., Wasserman, J., Pollack, H.A., Berson, J., Ahlstrom, A., 2000. Investing in youth tobacco
control: a review of smoking prevention and control strategies. Tob. Control 9, 47-63.
Lorenzo-Blanco, E.I., Unger, J.B., Ritt-Olson, A., Soto, D., Baezconde-Garbanati, L., 2011. Acculturation, gender, depression, and
cigarette smoking among U.S. Hispanic youth: the mediating role of perceived discrimination. J. Youth Adolesc. 40, 1519-
MacKinnon, D.P., Fairchild, A.J., 2009. Current directions in mediation analysis. Curr. Direct. Psychol. Sci. 18, 16-20.
McCarthy, W.J., Dietsch, B., Hanson, T.L., Zheng, C.H., 2008. Evaluation of the In-School Tobacco Use Prevention Education
Program, 2003-2004. California Department of Public Health, Sacremento, CA.
McChargue, D.E., Spring, B., Cook, J.W., Neumann, C.A., 2004. Reinforcement expectations explain the relationship between
depressive history and smoking status in college students. Addict. Behav. 29 991–994.
Minnix, J.A., Blalock, J.A., Marani, S., Prokhorov, A.V., Cinciripini, P.M., 2011. Self-efficacy mediates the effect of depression on
smoking susceptibility in adolescents. Nicotine Tob. Res. 13, 699-705.
Mistry, R., McCarthy, W.J., Yancey, A.K., Lu, Y., Patel, M., 2009. Resilience and patterns of health risk behaviors in California
adolescents. Prev. Med. 48, 291-297.
Ritt-Olson, A., Unger, J., Valente, V., Nezami, E., Chou, C., Trinidad, D., Milam, J., Earleywine, M., Tan, S., Johnson, C.A., 2005.
Exploring peers as a mediator of the association between depression and smoking in young adolescents. Subst. Use Misuse 40,
Page 20 of 26
Roberts, C., Williams, R., Kane, R., Pintabona, Y., Cross, D., Zubrick, S., Silburn, S., 2011. Impact of a mental health promotion
program on substance use in young adolescents. Advances in Mental Health 10, 72-82.
Rodriguez, D., Romer, D., Audrain-McGovern, J., 2007. Beliefs about the risks of smoking mediate the relationship between exposure
to smoking and smoking. Psychosom. Med. 69, 106-113.
Schleicher, H.E., Harris, K.J., Catley, D., Nazir, N., 2009. The role of depression and negative affect regulation expectancies in
tobacco smoking among college students. J. Am. Coll. Health 57, 507-512.
Schwarz, N., Bless, H., Bohner, G., 1991. Mood and persuasion: affective states influence processing of persuasive communications.
Adv. Exp. Soc. Psychol. 24, 161-199.
Spruijt-Metz, D., Gallaher, P., Unger, J.B., Johnson, C.A., 2005. Unique contributions of meanings of smoking and outcome
expectancies to understanding smoking initiation in middle school. Ann. Behav. Med. 30, 104-111.
Stanton, W.R., McClelland, M., Elwood, C., Ferry, D., Silva, P.A., 1996. Prevalence, reliability and bias of adolescents' reports of
smoking and quitting. Addiction 91, 1705-1714.
Steuber, T.L., Danner, F., 2006. Adolescent smoking and depression: which comes first. Addict. Behav. 31, 133–136.
Sun, P., Unger, J.B., Guo, Q., Gong, J., Ma, H., Palmer, P.H., Chou, C.-P., Li, Y., Sussman, S., Ritt-Olson, A., 2007. Comorbidity
between depression and smoking moderates the effect of a smoking prevention program among boys in China. Nicotine Tob.
Res. 9, S599-S609.
Trope, Y., Ferguson, M., Raghunathan, R., 2001. Mood as a resource in processing self-relevant information. In: Forgas, J.P. (Ed.),
Handbook of Affect and Social Cognition.
Page 21 of 26
US Department of Health and Human Services, 2010. How Tobacco Smoke Causes Disease: The Biology and Behavioral Basis for
Smoking-Attributable Disease:: A Report of the Surgeon General. Department of Health and Human ServicesCenters for
Disease Control and PreventionNational Center for Chronic Disease Prevention and Health PromotionOffice on Smoking and
Health. Atlanta, GA.
Waller, M.W., Hallfors, D.D., Halpern, C.T., Iritani, B.J., Ford, C.A., Guo, G., 2006. Gender differences in associations between
depressive symptoms and patterns of substance use and risky sexual behavior among a nationally representative sample of U.S.
adolescents Arch. Women’s Ment. Health 9, 139-150.
Young, T.L., Rogers, K.D., 1986. School performance and characteristics preceding onset of smoking in high school students. Am. J.
Dis. Child. 140, 257-225.
Page 22 of 26
Table 1. Descriptive analysis of study participants (n=24350)
Smoking Outcomes, % (n)
% (n) Intention Experimental Established
self-efficacy, % (n)
(n) Yes No Yes No Yes No Yes No Low High Mean SE
Column sample size 100
(20429) -0.01 0.00
Age p<0.001 p<0.001 p<0.001 p<0.001 p<0.010 p<0.001
12 yrs or less 20.9 20.2 79.8 4.3 95.7 1.8 98.2 0.3 99.8 17.2 82.8 0.08 0.01
13-14 yrs 30.5 27.7 72.3 13.6 86.4 5.8 94.2 1.3 98.7 17.4 82.6 -0.04 0.01
15-16 yrs 30.0 30.6 69.4 19.2 80.9 8.1 31.9 4.5 95.5 11.0 89.0 -0.03 0.01
17 yrs or more 18.7 29.4 70.6 21.9 78.1 9.1 90.9 7.1 92.9 8.6 91.4 -0.01 0.01
Gender p<0.001 p<0.001 p<0.001 p<0.001 p<0.001 p<0.001
Female 53.0 30.8 69.2 14.1 85.9 6.0 94.0 2.4 97.6 12.8 87.2 0.05 0.00
Male 47.0 24.0 76.1 15.7 84.3 6.6 93.4 4.0 96.0 14.7 85.3 -0.07 0.01
Race/ethnicity p<0.001 p<0.001 p<0.001 p<0.001 p<0.001 p<0.001
White Caucasian 42.0 25.2 74.8 15.3 84.7 6.0 94.0 4.6 95.4 11.5 88.5 0.05 0.01
Hispanic 33.5 30.1 70.0 17.3 82.7 7.8 92.2 2.0 98.0 16.5 83.5 -0.07 0.01
Asian/Pacific Islander 15.5 28.8 71.3 10.8 89.2 4.3 95.7 2.3 97.7 13.8 86.2 0.01 0.02
African American 7.2 30.0 70.0 11.3 88.7 5.1 94.9 1.7 98.3 12.0 88.0 -0.06 0.01
American Indian/AN 1.9 26.2 73.8 12.2 87.8 7.4 92.6 2.8 97.2 18.9 81.1 -0.05 0.03
Academic performance p<0.001 p<0.001 p<0.001 p<0.001 p<0.05 p<0.001
Superior 25.0 22.8 77.2 9.2 90.8 3.2 96.8 2.0 98.0 12.0 88.0 0.09 0.02
Good 33.2 25.4 74.6 11.5 88.5 4.7 95.3 2.2 97.8 12.0 88.0 0.05 0.01
Fair 27.9 29.5 70.6 18.4 81.6 7.9 92.1 3.7 96.3 13.5 86.5 -0.04 0.01
Poor 13.9 37.6 62.4 29.4 70.6 14.0 86.0 7.3 92.7 20.1 79.9 -0.21 0.01
Peers smoking p<0.001 p<0.001 p<0.001 p<0.001 p<0.001 p<0.001
Low 49.5 22.3 77.7 6.6 93.4 2.7 97.3 0.9 99.1 13.7 86.3 0.06 0.00
Moderate 39.7 31.4 68.6 19.5 80.5 8.4 91.6 4.2 95.8 12.2 87.8 -0.03 0.01
High 10.8 37.7 62.4 36.8 63.2 15.9 84.2 9.8 90.2 19.6 80.4 -0.24 0.01
Tobacco ease of access p<0.001 p<0.001 p<0.001 p<0.001 p<0.001 p<0.001
Hard 64.3 24.2 75.9 8.9 91.1 4.0 96.0 1.5 98.5 13.1 86.9 0.06 0.00
Easy 35.7 33.7 66.3 26.0 74.0 10.4 89.6 6.1 93.9 14.7 85.3 -0.13 0.01
Page 23 of 26
Table 2. Association between depressiveness and smoking measures, and mediation due to smoking-related knowledge and attitudes
(KA) and smoking refusal self-efficacy (SE) (n=24,350)
Model Set 11Model Set 22Model Set 33Percent mediation in relation to
Model Set 1
Smoking outcomes OR 95% CI OR 95% CI OR 95% CI Smoking KA Smoking KA &
smoking refusal SE
Intention to smoke 2.41*** 2.22, 2.61 1.84*** 1.68, 2.01 1.73*** 1.58, 1.90 54.35*** 57.80***
Experimental smoking 1.93*** 1.72, 2.17 1.47*** 1.30, 1.67 1.47*** 1.30, 1.67 68.03*** 68.03***
Established smoking 1.85*** 1.57, 2.18 1.34** 1.12, 1.60 1.16 0.97, 1.40 74.63*** 86.21***
1 Adjusted for age, gender, race/ethnicity & educational performance.
2 Adjusted for smoking-related KA, peer smoking and perceived ease of access to tobacco, in addition to Model Set 1 covariates
3 Adjusted for smoking refusal SE, in addition to Model Set 2 covariates
Page 24 of 26
ing effect of gender and race/ethnicity on the association of depressiveness with smoking
outcomes and mediation due to smoking-related knowledge and attitudes (KA) and smoking refusal self-efficacy
Model Set 11Model Set 22Model Set 33Percent mediation
Smoking Outcomes OR 95% CI OR 95% CI OR 95% CI Smoking
Intention to smoke 2.56 2.28, 2.88 1.97 1.74, 2.24 1.88 1.66, 2.14 50.76 53.19
Experimental smoking 1.98 1.68, 2.33 1.49 1.25, 1.77 1.47 1.24, 1.75 67.11 68.03
Established smoking 1.98 1.53, 2.56 1.48 1.13, 1.94 1.38 1.04, 1.82 67.57 72.46
Intention to smoke 2.13 1.88, 2.42 1.61 1.40, 1.86 1.50 1.30, 1.73 62.11 66.67
Experimental smoking 1.83 1.53, 2.19 1.45 1.20, 1.75 1.44 1.20, 1.74 68.97 69.44
Established smoking 1.64 1.30, 2.06 1.18 0.92, 1.52 0.99 0.77, 1.29 84.75 100.00
Intention to smoke 2.26 1.98, 2.57 1.63 1.41, 1.89 1.55 1.34, 1.80 61.35 64.52
Experimental smoking 2.04 1.69, 2.46 1.56 1.28, 1.89 1.57 1.29, 1.91 64.10 63.69
Established smoking 1.59 1.28, 1.99 1.13 0.89, 1.43 1.00 0.78, 1.28 88.50 100.00
Intention to smoke 2.50 2.17, 2.87 2.07 1.78, 2.41 1.97 1.69, 2.30 48.31 50.76
Experimental smoking 1.80 1.48, 2.18 1.45 1.19, 1.77 1.41 1.15, 1.72 68.97 70.92
Established smoking 2.50 1.74, 3.58 2.02 1.38, 2.94 1.84 1.26, 2.71 49.50 54.35
Intention to smoke 2.35 1.84, 2.99 1.79 1.37, 2.33 1.66 1.27, 2.18 55.87 60.24
Experimental smoking 2.07 1.45, 2.96 1.50 1.03, 2.18 1.48 1.02, 2.17 66.67 67.57
Established smoking 1.75 1.07, 2.86 1.23 0.72, 2.10 1.06 0.62, 1.84 81.30 94.34
Intention to smoke 2.41 1.68, 3.47 1.79 1.20, 2.67 1.67 1.11, 2.52 55.87 59.88
Experimental smoking 2.31 1.38, 3.84 1.65 0.95, 2.88 1.58 0.90, 2.76 60.61 63.29
Established smoking 1.44 0.58, 3.55 0.72 0.25, 2.04 0.64 0.22, 1.87 - -
Intention to smoke 2.20 1.05, 4.60 1.61 0.68, 3.77 1.31 0.52, 3.33 62.11 76.34
Experimental smoking 0.92 0.33, 2.59 0.90 0.30, 2.73 0.86 0.28, 2.67 - -
Established smoking 1.88 0.40, 8.90 1.80 0.30, 10.68 1.32 0.18, 9.46 55.56 75.76
1 Adjusted for age, gender, race/ethnicity & educational performance.
2 Adjusted for smoking-related KA, peer smoking and perceived ease of access to tobacco, in addition to Model Set 1
3 Adjusted for smoking refusal SE, in addition to Model Set 2 covariates
Page 25 of 26
This study was supported with funding from the California Tobacco Related Disease Research
RM originated the study concept and wrote the paper. GB assisted with writing the paper and
provided critical reviews. TM conducted the data analysis and provided critical reviews. WM
ensured the scientific integrity of the research and provided critical reviews.
Conflict of Interest
The authors declare that there are no conflicts of interest in relation to the information presented
in this manuscript.
Page 26 of 26
Ease of access to