The influence of depressive symptoms on
experimental smoking and intention to smoke in a
diverse youth sample
Elahe Nezami, Jennifer Unger, Sylvia Tan, Caitlin Mahaffey, Anamara Ritt-
Olson, Steve Sussman, Selena Nguyen-Michel, Lourdes Baezconde-Garbanati,
Stan Azen, C. Anderson Johnson
[Received 27 August 2003; accepted 16 October 2004]
Numerous studies have shown associations between smoking and depression, but the generalizability of the
relationship across ethnic groups remains unknown. The present study assessed the association between depression
and smoking intention and experimentation among adolescents from four ethnic groups in the Los Angeles area—
Chinese/Chinese American, Latino/Hispanic, Persian/Iranian, and White. Over 800 7th graders in the Los Angeles
area completed measures of depressive symptoms, experimentation with smoking, intention to smoke, and
sociodemographic covariates. Chinese/Chinese American students had the lowest levels of depressive symptoms,
whereas Latinos/Hispanics had the highest levels. Latinos/Hispanics also were the most likely to intend to smoke in
the next year and were the most likely to have started experimenting with cigarette smoking. Depressive symptoms
were significantly associated with intention to smoke even after controlling for language use acculturation,
socioeconomic status, gender, and ethnicity. The association between depressive symptoms and intention to smoke
did not vary significantly across ethnic groups. These results indicate that the association between depressive
symptoms and adolescent smoking generalizes across diverse ethnic groups.
Many studies have identified an association between
depression and smoking, although the directionality
of this relationship remains unclear (McMahon,
1999). Other studies have found ethnic variation in
both smoking (Mermelstein, 1999) and depression
(Roberts, Roberts, & Chen, 1997; U.S. Department of
Health and Human Services [USDHHS], 2001).
Among adolescents in California, smoking is typically
more prevalent among Whites and Latinos/Hispanics
than among Asian Americans and African Americans
(Gilpin et al., 2003). Adolescent depression typi-
cally is most prevalent among Latinos/Hispanics
(USDHHS, 2001). Few data about smoking habits
or mental health are available for other ethnic groups
of adolescents (e.g., Ahmadi & Hasani, 2003).
Because both ethnicity and depression correlate
with smoking habits, and because some ethnic
groups (Latinos/Hispanics) demonstrate high levels
of depression and high levels of smoking behavior,
ethnicity may be a confounder or a moderator of the
association between depression and smoking observed
in previous studies. The present study investigated the
association between depressive symptoms and smok-
ing behavior in adolescents across ethnic groups. We
assessed ethnic variation in this association, control-
ling for potential confounders including language use
acculturation and socioeconomic status (SES). The
present study focuses on three of the largest ethnic
minority immigrant groups in Southern California:
Latinos/Hispanics, Chinese/Chinese Americans, and
ISSN 1462-2203 print/ISSN 1469-994X online # 2005 Society for Research on Nicotine and Tobacco
Elahe Nezami, Ph.D., Jennifer Unger, Ph.D., Sylvia Tan, M.S., Caitlin
Mahaffey, B.A., Anamara Ritt-Olson, Ph.D., Steve Sussman, Ph.D.,
Selena Nguyen-Michel, B.A., Lourdes Baezconde-Garbanati, Ph.D.,
Stan Azen, Ph.D., C. Anderson Johnson, Ph.D., Institute for
Prevention Research, Keck School of Medicine, University of
Southern California, Los Angeles.
Correspondence: Elahe Nezami, Ph.D., Institute for Health
Promotion & Disease Prevention Research, Keck School of
Medicine, University of Southern California, 3375 South Hoover
Street, UVI E210, Los Angeles, CA 90089-7798, USA. Tel: +1 (213)-
821-1600; Fax: +1 (213)-821-1733; E-mail: firstname.lastname@example.org
Nicotine & Tobacco Research Volume 7, Number 2 (April 2005) 243–248
This paper reports the results of a cross-sectional,
school-based study of psychological correlates of
smoking across cultural groups of adolescents in
Southern California conducted during the 2001–2002
school year. The participants were 7th-grade students
attending middle schools in the Los Angeles area. The
survey assessed tobacco use and related psychosocial
variables among Chinese/Chinese American, Latino/
Hispanic, Persian/Iranian, and White students.
We used data from the California Board of Educa-
tion to identify schools with at least 10% Latino/
Hispanic, Chinese/Chinese American, or Persian/
Iranian students. We approached four school dis-
tricts and all four agreed to participate. Within these
districts, we asked eight schools to participate, and
six schools agreed.
All 7th-grade students in the participating schools
were invited to participate in the study. As required
by California law and the University of Southern
California’s institutional review board, students were
eligible to participate only if they provided written
assent and their parents provided active written
consent. In total, 1,836 students were invited. Of these,
1,123 (61%) provided both active parental consent and
student assent. A total of 1,041 students (57% of those
invited, 93% of those who provided consent) com-
pleted the survey. A comparison between the sample
and the demographic composition of the schools
revealed that boys and Latinos/Hispanics were slightly
under-represented in the sample.
Depression. Depressive symptoms were assessed with
the Centers for Epidemiological Studies Depression
(CES-D) scale. On the CES-D scale, 20 items ask the
respondent to think of the past 7 days and report on
how many days they experienced emotive states such
as ‘‘I felt sad’’ and ‘‘I felt hopeful about the future.’’
Respondents can choose from four options: ‘‘Less
than 1 day,’’ ‘‘1–2 days,’’ ‘‘3–4 days,’’ or ‘‘5–7 days.’’
Higher scores indicate higher levels of depressive
Chinese/Chinese-American, .88 for Latino/Hispanic,
.91 for Persian/Iranian, and .88 for White students).
a5.88overall; .88 for
smoking was assessed with a single dichotomous
behavior andintentions. Experimental
(yes/no) question about lifetime smoking: ‘‘Have you
ever tried smoking cigarettes, even a few puffs?’’
Intention to smoke in the next year was assessed with
a single item: ‘‘In the next 12 months, do you think
you will smoke a cigarette?’’ Students who provided
any response other than ‘‘definitely not’’ were coded
with a 1; otherwise, students were coded with a 0.
This procedure is consistent with those outlined by
Pierce and others in earlier adolescent smoking
research (Pierce & Gilpin, 1996).
Covariates. Covariates included age, gender, English
language usage, and SES. Age and gender were self-
reported. English language usage, a proxy measure of
acculturation, was assessed with three items on
language use in different contexts, adapted from
the Short Acculturation Scale for Hispanics (Marı ´n
& Gamba, 1996; Marı ´n, Sabogal, Marı ´n, Otero-
Sabogal, & Pe ´rez-Stable, 1987). The questions
asked students which language they thought in,
used at home, and used with friends. Response
options ranged from ‘‘Only English’’ to ‘‘Only
another language’’ (Cronbach’s a5.78 overall; .72
for Chinese/Chinese American, .71 for Latino/
Hispanic, .72 for Persian/Iranian, and .80 for White
students). Although the acculturation experience is
more complex than language usage alone, measures
of language usage typically correlate with more
extensive acculturation measures (Epstein, Botvin,
Dusenbury, & Diaz, 1996).
Socioeconomic status. An index of overcrowding
was used as a proxy SES measure (Myers, Baer, &
Choi, 1996). This measure, defined as the ratio of the
number of rooms to number of people living in a
home, is highly correlated with median ZIP code-
level household income in U.S. Census data.
White) was primarily self-identified. The only excep-
tion is for the category Persian/Iranian, which also
includes those who indicated in open-ended items
that they had family born in Iran or spoke Farsi,
excluding those of multiethnic background. A total
of 68 students self-identified as Persian/Iranian, and
an additional 31 students endorsed criteria for
cultural ties to Iran.
Respondents who identified themselves as both
Latino/Hispanic and White were categorized as
Latino/Hispanic. Those who identified themselves
both as Chinese and as Other Asian (Vietnamese,
Korean, Pacific Islander) were considered Chinese.
Non-Chinese Asian Americans, those of mixed origin
(i.e., who self-identified with more than one ethnic
group), and those of unknown origin were excluded
THE INFLUENCE OF DEPRESSIVE SYMPTOMS ON EXPERIMENTAL SMOKING
from the present analysis. Self-identified African
Americans also were excluded, because they com-
prised only 0.9% of the study sample.
Of the 1,041 respondents who provided data, 186
(17.9%) were of mixed, Other Asian, African
American, or unknown origin and were excluded
from the analysis. Of the remaining 855, seven (0.8%)
were missing smoking experimentation or intention
data, 65 (7.6%) were missing CES-D data, and 29
(3.4%) were missing SES or language use accultura-
tion data. Thus a total of 754 subjects (72.4% of the
sample) were included in analyses. Smoking pre-
valence did not differ significantly between the 754
respondents with complete data and the 94 missing
CES-D, SES, or language use acculturation data
(Fisher’s exact test, p5.45 for experimentation,
p51.00 for intention).
Table1 shows the demographic characteristics
of the sample, overall and by ethnic group. The
total sample size was 754 (58.4% female and 41.6%
male). The analytic sample was 20.3% Chinese/
Chinese American, 40.3% Latino/Hispanic, 10.5%
Persian/Iranian, and 28.9% non-Latino White. The
mean age was 12.6 years (SD50.5). Over 80% of the
respondents were born in the United States (75%
of Chinese/Chinese-American respondents, 85% of
Latino/Hispanic respondents, 64% of Persian/Iranian
respondents, and 84% of White respondents).
The sample was restricted to respondents with
complete data for ethnicity, gender, smoking, depres-
sion, SES, and English language usage. Univariate
analyses (one-way analyses of variance [ANOVAs]
and chi-square tests) were performed to identify demo-
graphicdifferences acrossethnic groups.ANOVA and
chi-square analyses also were performed to assess
variation in smoking, smoking intentions, and
depression across ethnic and gender groups.
Associations between depression and smoking were
evaluated using bivariate and multivariate logistic
regression. The multivariate logistic regression models
assessed the association between depression and
smoking experimentation and intention, adjusted for
age, gender, ethnicity, SES, and English language
usage. We used PC-SAS v. 8 for all analyses.
As shown in Table1, SES differed across ethnic
groups. Latinos/Hispanics reported the lowest SES,
and Whites and Chinese/Chinese Americans reported
the highest SES. English language usage also differed
across ethnic groups, with Whites reporting more
English language usage than the other three groups
(the other three groups did not differ significantly
from one another).
Gender and ethnic variation in smoking and depressive
Among the entire sample (N5754), 15.2% of res-
pondents had ever tried smoking. Table2 shows the
ethnic and gender variation in depressive symp-
toms and smoking behavior and intentions. Lifetime
Persians/Iranians reported the lowest smoking pre-
valence, and Latinos/Hispanics reported the highest.
Pairwise comparisons showed significant differences
between Latinos/Hispanics and each of the other
ethnic groups (p,.05) but not between any other
Among students who already had tried smok-
ing (lifetime ever-smokers), 58.3% reported inten-
tions to smoke in the next year. The only significant
ethnic difference in smoking intentions among ever-
smokers was between Chinese/Chinese Americans
and Latinos/Hispanics. Among students who had
never tried smoking (lifetime never-smokers), 11.9%
reported intentions to smoke in the next year.
Intentions to smoke among never-smokers were
significantly higher among Latinos/Hispanics than
by ethnic group.
Table 1. Demographic characteristics of analytic sample
Age in years, mean (SD)
Socioeconomic status, rooms-to-
people ratio, mean (SD)
English language usage, mean (SD)
Percentage born in United States
C-P, C-L, W-P,
W-P, W-L, W-C
C-L, P-L, P-W
Note. C, Chinese/Chinese American; L, Latino/Hispanic; P, Persian/Iranian; W, Non-Latino White. The analytic sample includes all
respondents with complete data, N5754.
*p(.05 for each, after adjusting for multiple comparisons.
NICOTINE & TOBACCO RESEARCH
among Whites and Chinese/Chinese
(p,.05), though no other significant pairwise ethnic
differences were found.
Chinese Americans. Significant differences in depres-
sive symptoms were found between Latinos/His-
panics and each of the other ethnic groups (p,.05),
though no other significant pairwise ethnic differ-
ences were found.
Boys were significantly more likely than girls to
report lifetime smoking (p,.05), but no significant
gender difference in intentions to smoke was
observed among lifetime never-smokers or lifetime
ever-smokers (Table2). Girls reported significantly
higher levels of depressive symptoms, compared with
Associations between depressive symptoms and
In bivariate logistic regression models (not shown
in the tables), depressive symptoms were associated
with an increased probability of smoking experi-
mentation (OR51.04, 95% CI51.02–1.05, p,.0005)
and smoking intentions (OR51.05, 95% CI51.03–
To testthe consistency
smokers and lifetime ever-smokers, we assessed the
interaction between lifetime smoking status and
depression in the prediction of smoking intentions.
The interaction term was significant (p,.0001). There-
fore, we conducted the analysis of intentions separately
among lifetime ever-smokers and never-smokers.
To determine whether the association between
depressive symptoms and smoking varied across
ethnic or gender groups, we added depression6
ethnicity and depression6gender interaction terms
to the model after the main effects. Neither interac-
tion term was significant. Therefore, we present
analyses of the entire sample instead of stratifying by
ethnicity or gender.
Table3 shows the association between depres-
sive symptoms and smoking, controlling for gender,
ethnicity, age, SES, and language use acculturation.
The analysis of smoking experimentation includes
the entire sample (N5754). In this analysis, depres-
sive symptoms were significantly associated with
smoking experimentation (adjusted OR51.03, 95%
CI51.01–1.05). Because the depression6lifetime
smoking interaction was significant in the prediction
of smoking intention, analyses of smoking intention
are presented separately for lifetime ever-smokers
(n5115) and lifetime never-smokers (n5639). In both
groups, the odds ratio for the association between
depressive symptoms and smoking intention was
significant (ever-smokers: OR51.05, 95% CI51.01–
1.09; never-smokers OR51.04, 95% CI51.01–1.06).
The present study assessed the association between
depressive symptoms and smoking behavior and inten-
tions in an ethnically diverse adolescent sample. Con-
sistent with previous research (reviewed by McMahon,
1999), depressive symptoms were associated with an
increased probability of smoking experimentation and
smoking intention. The association was consistent
across ethnic and gender groups, and it remained
significant after controlling for ethnicity, gender, age,
SES, and language use acculturation.
The results of this study may have been con-
founded by SES or English language usage; however,
the association between depressive symptoms and
Table 2. Gender and ethnic differences in smoking and depressive symptoms
Gender comparisonEthnic comparison
Lifetime smoking, percent (n)
Smoking intentions among
Smoking intentions among
(CES-D score), mean (SD)
11.5 (25) L-C, L-P, L-W
11.9 (76)13.6 (52)9.4 (24)4.2 (6)9.2 (7)21.0 (48)7.8 (15)L-C, L-W
L-C, L-P, L-W
Note. CES-D, Centers for Epidemiological Studies Depression scale; C, Chinese/Chinese American; L, Latino/Hispanic; P, Persian/
Iranian; W, Non-Latino White.
aSignificant gender difference, p,.05.
bSignificant gender difference, p,.005.
*p,.05 for each, after adjusting for multiple comparisons.
THE INFLUENCE OF DEPRESSIVE SYMPTOMS ON EXPERIMENTAL SMOKING
smoking remained significant across ethnic groups
after controlling for SES and English language usage.
It remains possible that the measures of SES and
English language usage in this study were not valid
measures and, therefore, did not control adequately
for these potential confounders. SES and accultura-
tion are difficult constructs to operationalize and
measure among adolescents. The rooms-per-person
measure correlates well with household income in
U.S. Census data, however, and previous studies
have found that brief language measures correlate
highly with more comprehensive acculturation scales
(Epstein et al., 1996). Therefore, although rooms-to-
people ratio and English language usage cannot
capture the qualitative complexity of SES and
acculturation, they tend to be highly correlated with
more extensive measures and therefore at least
partially control for their confounding effects.
The results of the present study are consistent with
smoking, further documenting that the association
between smoking and depression exists in early
adolescence (McMahon, 1999). Because these results
are cross-sectional, they could support either the self-
medication theory (i.e., adolescents smoke to relieve
depressive symptoms; Breslau, Peterson, Schultz,
Chilcoat, & Andreski, 1998) or the theory that
smoking increases depressive symptoms (Goodman
& Capitman, 2000; Wu & Anthony, 1999). In the
present study, because the association was observed
in a sample with a very low smoking prevalence, the
respondents’ depressive symptoms were unlikely to
have been caused by nicotine exposure. However,
longitudinal studies will be necessary to clarify the
direction of causality.
Few epidemiological studies of depression and
smoking have explored the role that ethnicity plays
in the link between depression and the early stages
of smoking intention and experimentation. The
results of the present study suggest that the associa-
tion between depressive symptoms and smoking
experimentation and intention exists across ethnic
groups of adolescents, as indicated by the nonsigni-
ficant depressive symptoms6ethnicity interaction
term. The study adds to the growing literature by
documenting this association among Persian/Iranian
Americans in addition to Whites, Latinos/Hispanics,
and Chinese Americans. The present study also
found an association between depressive symptoms
and smoking experimentation and intention among
both boys and girls, further supporting the link
between depression and smoking among adolescents.
These results are limited by the cross-sectional design
of the study. Previous studies have indicated that
the association between depression and smoking
is bidirectional: Depressive symptoms can increase
increase depressive symptoms (McMahon, 1999).
Longitudinal studies are needed to examine the
temporal patterns in these associations.
These results also are limited by the measures used.
Although the CES-D is a widely used measure of
depressive symptoms, it may not accurately measure
true levels of depression. It has been suggested that
Latinos/Hispanics report high scores on the CES-D
because of high levels of culturally determined
emotional expressiveness (Posner, Stewart, Marı ´n,
& Pe ´rez-Stable, 2001). Similarly, some aspects of
Chinese culture or the Chinese-American experience
may influence Chinese/Chinese-American subjects’
self-reports of depressive symptoms in a way that
affects their scores on instruments such as the CES-D
(Chen, Roberts, & Aday, 1998; Noh, Kasper, &
Chen, 1998). Little is known about the validity
Table 3. Covariate-adjusted associations between depressive symptoms and smokinga
Smoking intentions (lifetime
Smoking intentions (lifetime never-
Depressive symptoms (CES-D)
Model R2without CES-D
1.03(1.01–1.05)**** 1.05(1.01–1.09)** 1.04 (1.01–1.06)***
aModels simultaneously adjusted for age, socioeconomic status, and language use acculturation.
*Wald’s p(.10; **Wald’s p(.05; ***Wald’s p(.005; ****Wald’s p(.0005.
NICOTINE & TOBACCO RESEARCH
of the CES-D among Persian/Iranian-American
Further methodological research is needed to
examine the reasons for nonparticipation in school-
based surveys. In the present study, a post-hoc
comparison between our sample and the ethnic
compositions of the schools revealed that Latino
Therefore, these results may not generalize to
students who did not participate in the study. Some
youth who would have reported relatively higher
levels of depressive symptoms and greater intention
to smoke may not have been assessed in the study
(e.g., Dent et al., 1993). Unfortunately, lack of
parental consent to participate prevented us from
gathering data about students who did not ultimately
participate. However, the results showed an associa-
tion between smoking and depression even when
adjusted to account for ethnicity and gender.
explained only a small proportion of the variance
in smoking behavior. This finding is not surprising
because depression is only one of many psycho-
logical, social, environmental, and social risk and
protective factors that contribute to adolescents’
smoking behavior. Continued research is needed to
understand the other factors that influence adoles-
cents’ smoking. Nevertheless, even a small effect size
contributes to the growing body of knowledge about
adolescent smoking, which ultimately could contri-
bute to the development of improved prevention and
in our sample.
This research was supported by National Cancer Institute/National
Institute on Drug Abuse grant CA8475-01 (Transdisciplinary Tobacco
Use Research Center) and California Tobacco-Related Disease
Research Program grant 10DT-0162.
Ahmadi, J., & Hasani, M. (2003). Prevalence of substance use among
Iranian high school students. Addictive Behaviors, 28, 375–379.
Breslau, N., Peterson, E. L., Schultz, L. R., Chilcoat, H. D., &
Andreski, P. (1998). Major depression and stages of smoking: A
longitudinal investigation. Archives of General Psychiatry, 55,
Chen, I. G., Roberts, R. E., & Aday, L. (1998). Ethnicity and
adolescent depression: The case of Chinese Americans. The Journal
of Nervous and Mental Disease, 186, 623–630.
Dent, C. W., Galaif, J., Sussman, S., Stacy, A. W., Burton, D., &
Flay, B. R. (1993). Demographic, psychosocial and behavioral
differences in samples of actively and passively consented adoles-
cents. Addictive Behaviors, 18, 51–56.
Epstein, J. A., Botvin, G. J., Dusenbury, L., & Diaz, T. (1996).
Validation of an acculturation measure for Hispanic adolescents.
Psychological Reports, 79, 1075–1079.
Gilpin, E. A., White, M. M., White, V. M., Distefan, J. M.,
Trinidad, D. R., James, L., Lee, L., Major, J., Kealey, S., &
Pierce, J. P. (2003). Tobacco control successes in California: A focus
on young people. Results from the California Tobacco Surveys, 1990–
2002. La Jolla: University of California, San Diego.
Goodman, E., & Capitman, J. (2000). Depressive symptoms and
cigarette smoking among teens. Pediatrics, 106, 748–755.
Marı ´n, G., & Gamba, R. J. (1996). A new measurement of
acculturation for Hispanics: The Bidimensional Acculturation
Scale for Hispanics (BAS).
Sciences, 18, 297–316.
Marı ´n, G., Sabogal, F., Marı ´n, B. V., Otero-Sabogal, R., & Pe ´rez-
Stable, E. (1987). Development of a short acculturation scale for
Hispanics. Hispanic Journal of Behavioral Sciences, 9, 183–205.
McMahon, R. J. (1999). Child and adolescent psychopathology as
risk factors for subsequent tobacco use. Nicotine & Tobacco
Research, 1(Suppl. 2), S45–S50.
Mermelstein, R. (1999). Ethnicity, gender and risk factors for smoking
initiation: An overview. Nicotine & Tobacco Research, 1(Suppl. 2),
Myers, D., Baer, W. C., & Choi, S.-Y. (1996). The changing problem of
overcrowded housing. Journal of the American Planning Association
Noh, S., Kasper, V., & Chen, X. (1998). Measuring depression in
Korean immigrants: Assessing validity of the translated Korean
version of the CES-D scale. Cross-Cultural Research: The Journal of
Comparative Social Science, 32, 358–377.
Pierce, J. P., & Gilpin, E. (1996). How long will today’s new adolescent
smoker be addicted to cigarettes? American Journal of Public Health,
Posner, S. F., Stewart, A. L., Marı ´n, G., & Pe ´rez-Stable, E. J. (2001).
Factor variability of the Center for Epidemiological Studies
Depression scale (CES-D) among urban Latinos. Ethnicity &
Health, 6, 137–144.
Roberts, R. E., Roberts, C., & Chen, Y. R. (1997). Ethnocultural
differences in prevalence of adolescent depression. American Journal
of Community Psychology, 25, 95–110.
U.S. Department of Health and Human Services. (2001). Mental
health: Culture, race, and ethnicity—A supplement to mental
health: A report of the surgeon general. Rockville, MD: Author,
Public Health Service, Office of the Surgeon General.
Wu, L., & Anthony, J. C. (1999). Tobacco smoking and depressed
mood in late childhood and early adolescence. American Journal of
Public Health, 89, 1837–1840.
Hispanic Journal of Behavioral
THE INFLUENCE OF DEPRESSIVE SYMPTOMS ON EXPERIMENTAL SMOKING
Page 7 Download full-text