University of Southern California Dissertations and Theses

The role of depression symptoms on social information processing and tobacco use among adolescents

(USC Thesis Other)

The role of depression symptoms on social information processing and tobacco use among adolescents

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content

THE ROLE OF DEPRESSION SYMPTOMS ON SOCIAL INFORMATION PROCESSING AND
TOBACCO USE AMONG ADOLESCENTS

by
Kari-Lyn Kobayakawa Sakuma

A Dissertation Presented to the
FACULTY OF THE GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
DOCTOR OF PHILOSOPHY
(PREVENTIVE MEDICINE-HEALTH BEHAVIOR RESEARCH)

August 2009

COPYRIGHT 2009 Kari-Lyn Kobayakawa Sakuma

ii

DEDICATION

To Brandon who shows me how to be a better person.
To Dad who worked hard so I could pursue my dreams.
To Mama and Papa who showed me what hard work and love brings.
To Alissa, Jesssica, and Channa who let me be me.
Above all else, to Mom, who never let me think I should be any less.

iii

ACKNOWLEDGMENT
I would like to thank my committee members Drs. Jennifer Unger, Stan Azen,
Alan Stacy, and Larry Palinkas for their positive support and guidance. I would like to
especially thank my mentor and committee chairman, Dr. C. Anderson Johnson for his
faith in my abilities, the many opportunities he laid before me, his tolerance for my
many scattered ideas, and of course his guidance in my academic and professional
development. Special thanks go to Dr. Unger for her patience in reading many drafts
and leading me with insightful questions when times were especially tough. I am
grateful to Drs. Jean Richardson, Mary Ann Pentz, Lourdes Baezconde-Garbanati, and
Paula Palmer who have all challenged me and supported me, and have lived their lives
modeling how to be the change we want to see in the world. I would also like to
acknowledge Drs. Chih-Ping Chou, Ping Sun, Steven Cen, and Peggy Gallaher who were
always tirelessly and patiently there for all of my questions. To Drs. Nathaniel Riggs and
Joel Milam, who have both shown me great optimism and positivity and have served as
my role models. It took a village to get me to where I am today and I can only express
my deepest gratitude to all who have helped me along the way. I have had the privilege
of being among the IPR family, of whom I would especially like to thank Jolanda Lisath,
Marny Barovich, and my fellow classmates for making this journey feel like home.
This research was supported by the Transdisciplinary Tobacco and Alcohol Use
Research Center (NIH grant number: 5P50-CA 084735), the Sidney R. Garfield
Endowment, and the USC Provost Dissertation Completion Award.

iv

TABLE OF CONTENTS

Dedication ii
Acknowledgements iii
List of Tables v
List of Figures vi
Abstract vii
Chapter 1: Introduction 1
Chapter 2: Cross Sectional Analyses on Depression Symptoms, Social
Influences, and Smoking Relationships Among Chinese Adolescents

18
Chapter 3: Evaluating Depressive Symptom Interactions on Smoking
Prevention Program Mediators: A Mediated Moderation Analysis

36
Chapter 4: Exploring Social Competence, Depression Symptoms, and
Smoking Related Social Influence Factors

67
Chapter 5: Summary and Discussion 95
Bibliography 114

v

LIST OF TABLES

Table 1: Baseline Sample Characteristics

27
Table 2: Simple Correlations with the Wuhan Male Baseline Sample

27
Table 3: Wave 1 (Baseline) Models with Wuhan Males

29
Table 4: Mediation of Depression Symptoms on 30-Day Smoking
Through Social Norm Beliefs and Perceptions of Friend Use

30
Table 5: Wave 1 (Baseline) Moderation Models with Wuhan Males

31
Table 6: Sample Characteristics

53
Table 7: Wave 2 30-Day Smoking

54
Table 8: Wave 2 Social Influence Models

55
Table 9: Wave 2 Social Influence Models

57
Table 10: Mediation and Moderation on Wave 2 30-day Smoking

58
Table 11: Mediated Moderator Equations for Wave 2 30-Day
Smoking with Perceived Friend Prevalence (Mediator) and CoM
(Moderator)

60
Table 12: Mediated Moderator Slope Summary

61
Table 13: Sample Characteristics

78
Table 14: Simple Correlations for Chengdu Sample

85
Table 15: Stratified Male and Female Simple Correlations

86
Table 16: Depression Models

87
Table 17: Social Competence Models

88

vi

LIST OF FIGURES

Figure 1: Multiple Mediation Model 29
Figure 2: Mediated Moderator Model 52
Figure 3: Depression x Deficit Interaction on Social Norm Beliefs 89

vii

ABSTRACT

The current studies provide evidence of the psychosocial processes involved in
generating risk for smoking behaviors among those who exhibit high levels of
depression symptoms. Study 1 examined the relationship between depressive
symptoms, smoking social influences, and smoking behaviors among a sample of
adolescents. Study 1 results supported the hypothesis that the relationship between
depression symptoms and smoking behaviors were at least partially mediated by both
pro-smoking social norm beliefs and perceived friend smoking prevalence. There was
no evidence that depression symptoms moderated the relationship between social
influences and smoking. Study 2 examined whether changes in social influence
cognitions (via a social influences based smoking prevention program) would affect
smoking behaviors of students with high levels of depression symptoms more so than
those with low or no symptoms. Study 2 results provided evidence that the smoking
prevention program changed perceptions of friend smoking prevalence rates among
adolescents who had high scores of depression and who have previously experimented
with smoking. It was this change in perception that was responsible for the observed
reduction in 30-day smoking one year after program implementation. While
perceptions of the social environment might differ due to underlying cognitive
processes between depressed adolescents and non-depressed adolescents, an
important question to ask is whether social competencies may be a better predictor of

viii

social influence factors related to risk behaviors like smoking. Study 3 explored the
relationships between depression symptoms, social competence (sociability and/or
deficits), and smoking-related psychosocial perceptions. Results of study 3 were
consistent with the findings of studies 1 and 2 in that depression symptoms were
associated with higher perceived friend prevalence. In addition, study 3 found social
competence and depression to have complex relationships with psychosocial risk factors
for smoking. Sociability was associated with lower pro-smoking social norm beliefs but
only when depression is controlled for in the models. Social deficits modified the
relationship between depression and perceptions of social norms. The interaction
suggested that depression was negatively associated with pro-smoking norm
perceptions only among those with social deficits. Collectively, all three studies suggest
that perceptions of adolescent social environments are important determinants in
smoking behaviors and interventions which target changing perceptions of adolescent
social environments would benefit from understanding how dispositional phenotypes,
such as risk for depression and social competence, would modify program effects.
1

CHAPTER 1
SPECIFIC AIMS
Many of the risk and protective factors for adolescent depression appear to be
related to social relationship and perceptions of self within a social context. Positive
self-concept, family “connectedness”, and peer acceptance were found to be protective
against risk for depressive episodes (MacPhee & Andrews, 2006; Van Voorhees et al.,
2008), while peer rejection and perception of number of friends were associated with
risk for depression (Reijntjes, Stegge, & Terwogt, 2006; Ueno, 2005). Cognitive theories
of depression suggest that individuals who are vulnerable to depression may perceive or
interpret their social environment in a way that inhibits processing of positive affect or
promotes depressive affect (Lakdawalla, Hankin, & Mermelstein, 2007; van Beek &
Dubas, 2008). Adolescents who exhibit depression symptoms may be sensitive to
socially threatening cues such as those that confer rejection (Joormann & Gotlib, 2007;
van Beek & Dubas, 2008) or negative emotional information (Ladouceur et al., 2005).
Depressed adolescents process social information in a way that may manifest as
sensitivity to social influences.
Social influence is among the most pervasive determinants of smoking onset in
adolescents (Kobus, 2003). Depressed adolescents may be at increased risk for smoking
due to their sensitivity for pro-smoking social influence messages (Allen, Porter, &
McFarland, 2006). This research examines how depression symptoms affect
perceptions of smoking related social norms and whether it accounts for higher levels of

2

smoking behaviors. Depressed adolescents may perceive or process the information
regarding social norms of smoking differently than do their non-depressed counterparts,
resulting in differences in smoking risk. With a social influences-based smoking
prevention program, these studies will investigate how targeted changes in social
information are perceived and processed, and whether there are differences between
those with and without depression symptoms. Deficits in social competence may also
play an important role in depression and social influences. Poor social competence may
be an antecedent to peer rejection and may contribute to depression symptoms
associated with social relationships in adolescence. Social competence and depression
will be investigated together to parse differences in social perception of smoking norms.
The specific aims of the study are:
To understand whether those with higher depression symptoms will have higher
scores on social influences, as measured by perceived friend (30-day smoking)
prevalence and perceived social benefits/risks of smoking;
To study whether depression will moderate the relationship between social
influences and smoking among adolescents, such that those with higher
symptoms would exhibit a stronger positive relationship between social
influences and smoking behavior;
To evaluate whether depression moderates how a social normative prevention
program is received and how it affects changes in smoking behaviors;

3

To understand the unique contribution of depression versus social competence
in perceiving smoking related social influence information.
BACKGROUND AND SIGNIFICANCE
Peer relationships grow in importance and influence during adolescence.
Adolescents learn the rules of social engagement and choose behaviors based on their
assessments of their social environment (Bos, Sandfort, de Bruyn, &Hakvoort, 2008;
Deb, Mitra, & Mukherjee, 2001; Makri-Botsari, 2005; Nelson, Leibenluft, McClure, &
Pine, 2005). Adolescents experience great interpersonal stress due to the changing peer
hierarchies and accompanying new social rules (Rudolph & Conley, 2005), identity
formation (Montague, Enders, Dietz, Dixon, & Cavendish, 2008), and the transition in
social support from parent to peers in new social environments (Newman, Newman,
Griffen, O’Conner & Spas, 2007). These examples of social stressors have been
associated with the development of depression in adolescents (Hankin, Mermelstein, &
Roesch, 2007; Rudolph, 2008; Rudolph, Ladd, & Dinella, 2007).
The prevalence of depression has been shown to increase linearly from 2% in
prepubescent children to 5-8% in early adolescence and rising to about 20% prevalence
in the adult population (Birmaher, Ryan, Williamson, & Brent, 1996; Jellinek & Snyder,
1998; Kessler, Avenevoli, & Ries Merikangas, 2001). Disturbances in social engagement
and motivation during adolescence create vulnerabilities for future depressive episodes
(Birmaher et al., 2004; Dahl & Spear, 2004; Williamson, Forbes, Dahl, & Ryan, 2005). All

4

cognitive theories of depression are based on diathesis-stress models where a
predisposition for depression (eg. Genetic make-up, neuronal milieu) interact with
stressors in the environment to produce a depressive phenotype. These theories
generally posit that individuals differ in the way they pay attention to stimuli, perceive
and interpret stimuli, and how they remember negative events. They further posit that
these differences contribute to risk for developing depression later in life (Hermans et
al., 2008; Lakdawalla et al., 2007). Depression during adolescence has direct
implications for peer relations and future health behaviors (Andersen & Teicher, 2008;
Aseltine, Gore, & Colten, 1994; Keenan-Miller, Hammen, & Brennan, 2007). Depression
symptoms have been associated with poor relationships with parents and peers
(Brendgen, Wanner, Morin, &Vitaro, 2005), substance use (Chinet et al., 2006; Crum,
Storr, Ialongo, & Anthony, 2008; McCaffery, Papandonatos, Stanton, Llyod-Richardson,
& Niaura, 2008), and suicide ideation (Beam, Gil-Rivas, Greenberger, & Chen, 2002;
Nrugham, Larsson, & Sund, 2008). Depression and smoking during adolescence appear
to have a particularly robust association, though the debate on the direction of causality
continues (Brook, Schuster, & Zhang, 2004; Dierker, Vesel, Sledjeski, Costello & Perrine,
2007; McCaffery, et al., 2008; Rodriguez, Moss, & Audrain-McGovern, 2005).
Beck’s Cognitive Theory of Depression posits that individuals have developed a
negative or maladaptive self-schema that affects the way a person encodes, interprets,
and remembers information (Beck, 1987; Beck & Clark, 1991). Beck suggests that an
individual who is predisposed to depression would make cognitive errors in judgments

5

about themselves, the world, and their future. Repeated exposure to negative events or
their consistent attention to negative stimuli would reinforce depression symptoms, and
the individual may experience worse symptoms over time leading to major depression
disorder (Beck, Brown, Steer, Eidelson, &Riskind, 1987; Beck & Clark, 1991; Lakdawalla
et al., 2007). Consistent with Beck’s theory, several empirical studies have found
attentional biases in clinically depressed adolescents. In a study of adolescent girls
which compared those with biological mothers with recurrent Major Depressive
Disorder (MDD) to those without, researchers found that girls at elevated risk for
depression selectively attended to negative facial expressions while the control girls
selectively attended to positive facial expressions (Joormann & Gotlib, 2007). Similar
results were found in a study comparing youth diagnosed with MDD and co-morbid
MDD with anxiety to normal controls on a Go/No Go task (Ladouceur et al., 2006). Only
those youth with MDD and co-morbidity had significantly longer reaction times when
the backgrounds on the tasks held negative emotional images compared to neutral
images. Normal control youth had significantly longer reaction times when the
background held positive emotional images. The authors attribute these differences
between the MDD/Co-morbid and Control groups to altered processing of emotional
information and to interference in processes that govern how attentional resources are
allocated. Although these studies suggest that depressed youth may have an
attentional bias toward negative stimuli, other studies found no attentional biases
present (Neshat-Doost, Moradi, Taghavi, Yule & Dalgleish, 2000; Taghavi, Neshat-Doost,

6

Moradi, Yule, & Dalgleish, 1999) and point to cognitive biases occurring at later stages of
processing. For example, in an experimental study with 142 adolescents, children with
higher depression symptoms were more likely to prefer negative feedback in response
to negative peer evaluations (Reijntjes, Dekovic, Vermande, &Telch, 2007). This
suggests that the bias may not occur during initial attention allocation but in seeking
information that would confirm their negative cognitions.
Consistent with cognitive theories, depressed adolescents may process social
information differently than non-depressed adolescents. In a study with 606 youth,
researchers found that those who had a difficult time decoding subtle nonverbal cues
were more depressed than those who were accurate at decoding such subtle cues (van
Beek & Dubas, 2008). The authors found that the perception of anger and joy was
related to depression symptoms, but perception of sadness or fear was not. They
attributed this to the possibility that anger and joy might communicate emotional cues
representing rejection or acceptance (van Beek & Dubas, 2008). This finding is
distinguished from attentional bias in that those who were depressed did not selectively
attend to negative facial expressions but rather they may have misinterpreted or
inaccurately decoded those emotional faces indicating that processing of the social
information may be different among depressed adolescents. Another study with 72
children found evidence for selective processing among depressed and normal control
children (Timbremont & Braet, 2005). Within group comparisons showed that both
control and depressed children rated more positive words as self-descriptive than

7

negative words. However, control children rated significantly more positive words as
self-descriptive overall. More importantly, a difference in selective processing was
apparent through a recall task. The control group recalled more positive self-descriptive
words than the depressed group, and the depressed group recalled more negative self-
descriptive words than the control group (Timbremont & Braet, 2005). This study
supports cognitive theories that propose either cognitive biases in processing
information or biased recall of negative events that contribute to a negative schema for
depressed adolescents. Depression symptoms have been associated with discrepancies
between self and peer evaluations of social competencies. In a longitudinal study,
researchers provided children with a class roster and asked them how much they liked
each classmate using a scale of 1 to 5, with high scores indicating higher liking (Kistner,
David-Ferdon, Repper, & Joiner, 2006). Children were also asked to predict how each of
those students would rank them, using the same scale. The accuracy of the children’s
perceptions was then assessed by taking the mean difference between pairs of
predicted and received ratings. Depression symptoms predicted decreasing accuracy in
peer acceptance and inaccurate perceptions predicted increases in depression
symptoms (Kistner, et al., 2006). Another study with undergraduates found that
discrepancies between self evaluation on a self-esteem scale and their roommates’
evaluations were associated with higher depression symptoms (Joiner, Kistner,
Stellrecht, Merrill, 2006). This study found that discrepancies, whether they were self-
enhancing or self-deprecating, were both associated with higher depression symptoms.

8

Most research only focused on negative evaluation of self being related to depression
symptoms. These studies provide support for biased social cognitions among depressed
adolescents.
Identity formation, self-concept, and self-worth develop and strengthen during
adolescence due to peer relationships and maturation (Cole et al., 2001; Davey, Yucel, &
Allen, 2008; Gregory et al., 2007). Perceptions of self that are consistent with a negative
schema of self have been associated with depression symptoms. Social evaluative
concerns were associated with concurrent and future levels of depression, especially
among adolescent girls (Rudolph & Conley, 2005). In a longitudinal study with 228
adolescents, low self-concept was associated with higher depression symptoms
(Montague, et al., 2008). Authors found growth trajectories that illustrated that as self-
concept increased depression symptoms decreased (Montague, et al., 2008). Another
longitudinal study supported the role of self-concept in depressive symptom
development (Van Voorhees et al., 2008). Van Voorhees and colleagues (2008) found
that active coping and positive self-concept predicted lower risk for depression
symptoms a year later. These studies suggest that those adolescents with poor
individual identity formation may have high levels of depression symptoms. These
adolescents may develop or continually reinforce negative cognitions when interacting
with their social environments. These studies are not able to demonstrate causal
mechanisms that would indicate whether depression leads to perceptions of low social
ability or low concept of self, actual poor social skills lead to actual social rejection, or if

9

these deficits in social interaction lead to depression symptoms (Hoffman, Cole, Martin,
Tram & Seroczynski, 2000).
Interpersonal theories of depression may further inform these complex relationships
between biased perceptions and social interactions. Coyne’s Interpersonal Theory of
Depression (1976) deviates from cognitive models in that he places emphasis on a
dysfunctional relationship between reassurance-seeking individuals and their frustrated
friends as the impetus for developing depression. These frustrated friends continue to
verbally reassure but simultaneously exhibit negative non-verbal cues that reflect a level
of rejection. Coyne’s assertion directly challenges cognitive theories in that the
depressed individual perceives social rejection accurately and it is the interaction
between the two players that produces depression. Although there has not been broad
empirical support for interpersonal theories (Marcus & Nardone, 1992), peer interaction
remains an important factor in understanding the etiology of depression and associated
problem behaviors. In a longitudinal study of a large cohort of adolescents, researchers
found that number of friends was negatively associated, but produced the strongest
predictor of, depression symptoms (Ueno, 2005). Adolescents who associated with very
popular students (those receiving the most nominations) received some protective
benefit against depression symptoms although popular students were not any more or
less likely to have depression symptoms than the average student. The authors found
that a sense of belonging mediated the effect between number of friends and level of

10

depression symptoms suggesting that the positive affect associated with social
integration protects against depression (Ueno, 2005).
Depression has been found to be strongly interrelated with smoking (Covey,
Glassman, & Stetner, 1998; Fergusson, Goodwin, & Horwood, 2003; Fergusson, Lynskey,
& Horwood, 1996; Glassman & Covey, 1996; Johnson et al., 2007; Johnson et al., 2005;
Munafo, Hitsman, Rende, Metcalfe, & Niaura, 2008; Needham, 2007; Sun et al., 2007).
In a twin study with 16-year-olds, there was significant correlation between depression
symptoms and smoking involvement, which, for the male sample, was solely
attributable to non-shared environmental or nonfamilial effects (McCaffery, et al.,
2008). For females, smoking and depression symptoms were also significantly
associated and attributable to non-shared environments, but an additive genetic
contribution was also present. This study provided support that non-shared
environments, such as peer interactions and peer expectations, may be responsible for
the association between smoking and depression symptoms. Another study that
followed an adolescent cohort found distinct trajectories for depression and its
association with smoking (Rodriguez et al., 2005). Students were stratified by 9
th
grade
baseline depression symptom levels and followed through the 12
th
grade. Smokers with
the highest depression scores in 9
th
grade had lower scores of depression at 12
th
grade
compared to those with only moderate levels of depression which showed increased
depression scores at 12
th
grade (Rodriguez et al., 2005). Among those with the lowest
depression scores, smoking at 9
th
grade did not show any acceleration or deceleration

11

pattern of depression scores at 12
th
grade. These findings further support that
depression and smoking are significantly associated and that the trajectories of those
associations during adolescence are complex. Causal direction could not be inferred
from this study.
Although there is some support for direct causal relationships between smoking and
depression, other studies suggest indirect relationships (Patton et al., 1998; Rao, 2006;
Ritt-Olson et al., 2005). Specifically, social perceptions and peer influences may be an
important mediator between depression and smoking. Peers play an important role in
establishing expectations and norms about behaviors, and, during adolescence, are
especially influential in that respect. Smoking onset and experimentation is largely a
socially driven behavior (Patton et al., 1998; Unger et al., 2001). It has been well
established that social perceptions and peer influences are primary factors in adolescent
smoking (Kobus, 2003). Depression may impact how the individual processes social
information and how they choose to act based on that information.
For depressed adolescents, it appears that peer deviant behaviors are largely
overestimated. In a study with a mid-adolescent sample, researchers found that those
who overestimate peer involvement in deviant behaviors, such as tobacco use, alcohol
use, serious fighting, or stealing, were also participants in those behaviors and reported
the most depression symptoms and lowest self-esteem (Jacobs & Johnston, 2005).
Specific to smoking outcomes, another study with adolescents assessed the mediating
role of peers between depression and smoking (Ritt-Olson et al., 2005). Peer influence

12

was measured by asking the students to think about their five best friends and indicate
how many would approve of smoking. The authors found that when peer influence was
considered in the model, the relationship between depression and smoking became
non-significant, indicating statistical mediation (Ritt-Olson et al., 2005). In other words,
perceptions about friends’ approval of smoking behaviors explained the relationship
between depression and smoking. However, this study was not able to address if
depression magnified perceptions of peer approval or what mechanism would explain
the association between depression and peer approval.
A longitudinal study with 106 adolescents, their parents, and a close peer of their
choosing were assessed for susceptibility to peer influence and depression (Allen et al.,
2006). Susceptibility to peer influences was measured by interviewing the target
adolescent and asking them to make a decision regarding a hypothetical situation (i.e.
they are stuck on another planet and must choose 7 of 12 fictional characters to return
to earth) and then separately interviewing their close peer with the same situation.
Both children were then asked to meet together and make a decision together for the
final answer. Peer influences were assessed as percentage of instances that the target
adolescent changed their initial answer to match their peer. Allen and colleagues (2006)
did not find evidence for concurrent associations between depression and susceptibility
to peer influence and could not assess whether depression led to higher levels of
susceptibility to peer influence. However, they found that susceptibility to peer
influence at baseline predicted an increase in depression symptoms at one year

13

reassessment (Allen et al., 2006). Notably, those susceptible to peer influences were
more likely to be pressured into negative problem behaviors (e.g. pick fights, smoke, cut
class, teasing), be engaged in actual higher levels of problem behaviors, and for those
most susceptible to peer influences, a stronger relationship was found between peer
drug use and own difficulties associated with drug use. This study illustrates the
complex interactions between individual susceptibility to peer influence, depression,
and problem behaviors. More importantly, this study demonstrates a relationship
between depression and social influences in adolescent behaviors.
A longitudinal cohort study found that among adolescents, depression and anxiety
increased risks associated with peer smoking and predicted experimentation only in the
presence of peer smoking (Patton et al., 1998). Peer smoking was assessed by whether
adolescents reported none, some, or most of their friends were smokers. Researchers
found a peer smoking X psychiatric morbidity (depression and anxiety) interaction for
both experimental smoking and daily smoking. This indicates that the relationship
between depression and smoking behaviors was modified by peer influence, thus
researchers chose to stratify their analyses by different levels of peer smoking. Among
those who reported most of their friends smoked, high depression and anxiety scorers
were almost 3 times more likely to become smokers; while in those who reported that
none of their friends smoked, depression and anxiety did not increase risk for smoking
initiation (Patton et al., 1998). These findings are consistent with the complex peer
influence and tobacco use literature which emphasizes immediate social contexts (i.e.

14

perceived friend use in this case) in determining the norms that influence individuals’
behaviors (Henry & Kobus, 2007; Kobus, 2003).
Finally, a cross-sectional study with 1123 9
th
graders found that adolescents who
scored high on advertising receptivity and depressed moods were most vulnerable to
experimenting with smoking and their risk was even greater when exposed to peer
smoking (Tercyak, Goldman, Smith, & Audrain, 2002). The effects of advertising on
smoking behavior were heightened among the depressed adolescents, but exposure to
other smokers (e.g. family, peers) by itself did not increase the risk for being a smoker in
the presence of depression symptoms. This suggests that simply being around others
that smoke may not increase the risk of smoking onset for depressed adolescents, but
that the increased risk may, in fact, be due to how depressed adolescents perceive
smoking, such as providing positive attention from peers or more overall interaction
with peers. Perceptions of social information from advertising, from social contexts, or
from peer interactions may influence smoking behaviors and, among depressed
adolescents, it may strengthen that influence.
RATIONALE FOR STUDIES
Bandura’s (Bandura, 1977) Social Cognitive Theory/Social Learning Theory and
Festinger’s (Festinger, 1954) Social Comparison Theory generally suggest that individuals
learn from and are influenced by others in their social environment. Social Cognitive
Theory is more explicit in stipulating how individuals learn through observing others’
modeling behaviors and the reinforcement or punishment one would receive as a

15

consequence to such behaviors. Social Comparison Theory similarly suggests that
individuals look to others to evaluate their own behaviors.
Deutsch and Gerard (1955) distinguished between two types of social influences
which may give insight on why an individual may choose to participate in a specific
behavior. They identified normative social influence as “an influence to conform to the
positive expectations of another” (Deutsch & Gerard, 1955, p. 629) and informational
social influences as “an influence to accept information obtained from another as
evidence about reality” (Deutsch & Gerard, 1955, p. 629)(p. 629). The distinction
between these types of influence is at the root of operationalizing social influences
based prevention programming constructs. Normative social influence connotes a
certain level of motivation to comply with others (or perception of others’ beliefs) in
order to gain social rewards or avoid punishment (Campbell & Fairey, 1989; Deutsch &
Gerard, 1955). By disagreeing or acting in a way contrary to the surrounding social
group, an individual may be socially ridiculed and risk being rejected (Gerard & Rotter,
1961). An individual who is influenced by informational social influence processes is
driven by the desire to be accurate and look to others to provide a “reality check.” That
individual may choose to conform to a group norm because they believe the rest of the
group is correct (Campbell & Fairey, 1989). Generally, social influences based
prevention programs rely on both types of social influence to enact behavior change.
The program provides informational social influence by correcting overestimations of
smoking prevalence estimates but relies on normative social influence to build an anti-

16

smoking environment within the classroom and encourage motivation to comply with
the “correct” norm. The more uncertain an individual is about the accuracy of their
judgment, the more susceptible they become to both types of social influence. By
showing students the discrepancy between their estimation of peer smoking and actual
prevalence rates, the program creates greater acceptance for the establishment of the
new anti-smoking norms (Informational social influence) and increases the pressure for
conformity with the new group (i.e. classroom) judgments (Normative Social Influence).
This dissertation hypothesizes that adolescents who have higher levels of depression
symptoms are generally more influenced by normative social influences than those with
lower scores. A study with clinically depressed adults demonstrated that depressed
individuals may be more prone to conformity under social pressure than controls
(Katkin, Blum Sasmor, & Tan, 1966). Similarly, we would expect depressed adolescents
to be more sensitive to influences than their non-depressed counterparts. Specifically,
depressed adolescents may be more motivated to comply with the social group
expectations due to their heightened sensitivity to rejection. Social norms are
subjective and as such can be considered as an uncertain judgment. When an
intervention brings that depressed individual’s judgment into question and replaces it
with 1) the accurate numbers and beliefs and 2) allows discussion within the classroom
to socially confirm the new reality, we can expect to see depression moderating
program effects. These theories suggests that individuals may smoke if they perceive

17

high prevalence rates of smoking among their friends and if they perceive that smoking
provides positive social benefits.
This research investigated the role of depression symptoms, social influences,
and smoking behaviors among adolescent youth. Each of these studies provides
support for the argument that adolescents who exhibit depression symptoms are more
sensitive to their social context as it relates to smoking and will process social norms
regarding smoking in a way that it results in higher risk for smoking behaviors. Study 1
investigates the relationships present between depression, social influences, and
smoking using cross-sectional data. Study 2 researches how manipulated changes in
perceptions of smoking related social influences (via social-influences based smoking
prevention curriculum) might be most effective among adolescents exhibiting
depression symptoms. Finally, Study 3 explores whether depressive symptomatology
alone is responsible for biased processing of social information, or if levels of social
competence affect how adolescents are able to perceive social information. Study 3
provides clarity on the nature of social interactions and how social competence during
adolescence may contribute to depression symptoms and perceptions of the social
environment.

18

CHAPTER 2: Cross-Sectional Analyses on Depression symptoms, Social Influences, and
Smoking Relationships Among Chinese Adolescents.

ABSTRACT
Smoking and depression during adolescence is a complex and dynamic
relationship that warrants closer examination. The current study investigates smoking,
depression, and social influence relationships among 1391 male adolescents in central
China. In a multiple mediator model using cross-sectional data, social normative beliefs
(i.e. smoking makes others look cool, attractive) and perceived friend smoking
prevalence were found to partially mediate the relationship between depression
symptoms and smoking behaviors. Depression symptoms were not found to modify the
relationship between social influence factors and smoking. The conclusions drawn from
this study suggest that adolescents at risk for depression may be biased in their
perception or more influenced by their social environment. Biased social perceptions
among depressed adolescents may be responsible for higher risk for smoking behaviors.

19

INTRODUCTION
Depression has been found to be strongly interrelated with smoking (Covey,
Glassman, & Stetner, 1998; Fergusson, Goodwin, & Horwood, 2003; Fergusson, Lynskey,
& Horwood, 1996; Johnson et al., 2007; Munafo, et al., 2008; Sun et al., 2007). Although
there is some support for direct causal relationships between smoking and depression,
other studies suggest indirect relationships (Patton et al., 1998; Rao, 2006; Ritt-Olson et
al., 2005; Tercyak et al., 2002). Specifically, social perceptions and peer influences may
be an important mediator between depression and smoking. Peers play an important
role in establishing expectations and norms about behaviors, and during adolescence,
are especially influential in that respect. It has been well established that social
perceptions and peer influences are primary factors in adolescent smoking (Henry &
Kobus, 2007; Kobus, 2003; Patton et al., 1998; Unger & Rohrbach, 2002; Unger,
Rohrbach, Howard-Pitney, Ritt-Olson, & Mouttapa, 2001). Depression may impact how
the individual processes social information and how they choose to act based on that
information.
For depressed adolescents, socially endorsed risky behaviors may be particularly
appealing. In a study with a mid-adolescent sample, researchers found that those who
overestimated peer involvement in deviant behaviors, including tobacco use, alcohol
use, and serious fighting among others, reported the most depression symptoms and
lowest self-esteem (Jacobs & Johnston, 2005). They also reported higher participation
in such deviant behaviors. Another study showed that perceptions about friends’

20

approval of smoking behaviors mediated the relationship between depression and
smoking (Ritt-Olson et al., 2005). This suggests that depressed students either perceive
more friends who endorse smoking or they may simply congregate based on similar
beliefs or attitudes regarding smoking or depressive views.
According to cognitive theories of depression, depressed individuals are biased
to view their environment negatively (Lakdawalla et al., 2007). Adolescents at risk for
developing depression may be highly attuned to information in the environment that
may convey acceptance or rejection by peers (van Beek & Dubas, 2008). Thus smoking
norms may be perceived by high-risk adolescents as a way to “fit in” or may be more
motivated to comply with what they believed would garner more approval from their
friends or risk rejection. These are complex interactions between peer influences,
depression, and problem behaviors that need to be better understood.
The current study analyzes how depression symptoms, perceptions of pro-
smoking norms, perceptions of smoking prevalence rates among friends, and smoking
behaviors are associated.
Specific Aims and Hypotheses:
1.1 Study 1 will explore the different relationships between social influences,
depression and smoking.
Hyp.1.1.a Social influences and depression will have a positive association
such that those with high levels of depression will have higher

21

positive social norm beliefs about smoking and higher perceived
friend prevalence.
Hyp.1.1.b Social influences and smoking will have a positive association such
that those who have higher positive social norm beliefs about
smoking and higher perceived friend prevalence will also have
higher levels of smoking (or be more likely to have smoked in the
past 30-days than those who have lower perceived social norms).
Hyp.1.1.c Depression will be positively associated with smoking such that
those who have higher levels of depression will be more likely to
have smoked in the past 30-days than those who have lower
levels of depression.
1.2 Study 1 will investigate whether social influences and depression will predict
smoking status in a regression model.
Hyp.1.2.a Higher levels of positive social norm beliefs about smoking, higher
perceived friend prevalence, and higher levels of depression will
predict a higher probability of being a 30-day smoker.
1.3 The association between depression and smoking may be due to differences in
how those at-risk for depression perceive social influences. Thus, Study 1 will
examine if depression moderates the relationship between social influences and
smoking.

22

Hyp.1.3.a The relationship between social influences and smoking will not
vary by level of depression.

METHODS
Sample Selection
The Wuhan Smoking Prevention Trial (WSPT) was a school-based longitudinal,
randomized controlled trial aimed at preventing initiation and escalation of adolescent
smoking behaviors (Chou et al., 2006). Wuhan is a dense urban city and is the capital of
Hubei Province in central China with a population of over 9 million. Middle schools
were randomly selected from each of the seven urban districts within the city of Wuhan
and then additional schools within each district were matched on school size,
teacher/student ratio, and academic ratings to create pairs. One school from each pair
was then randomly assigned to program or control condition (for further details on
study design see Unger et al., 2001).
Seventh grade students were assessed with a 200-item paper-and-pencil
baseline survey prior to program implementation in 1999. The sample for the current
study was limited to Wave 1 (baseline) data. Where the prevalence of male lifetime
smoking and 30-day smoking provided enough variance to test our hypotheses, the
female prevalence rates were minimal (see Table 1). There were also significant gender
differences for depression etiology and smoking in prior studies to warrant analyzing
relationships stratified by gender. Further analyses showed significant gender

23

differences in this sample (see Table 1) thus the mediation and moderation models were
limited to males adolescents (n=1391).
Measures
Smoking. Lifetime smoking was assessed by asking, “Have you ever smoked a
cigarette, even just a few puffs?” (0=no; 1=yes). Thirty-day prevalence was assessed by
asking “Think about the last 30-days. On how many of those days did you smoke
cigarettes?” (0=smoked<1 day in the past month; 1=smoked at least 1 day). The
reference group for 30-day smokers was those individuals who have never smoked, quit
smoking, and ever smoked combined.
Depression. Depression symptoms were measured by asking, “Have you felt
depressed in the last week”, “Have you felt alone in the last week?”, “Have you felt sad
in the last week?”, and “ Have you felt like crying out in the last week?” with four
response choices: (1) Almost never, (2) Seldom, (3) Occasionally, and (4) Often. The
mean of the four-item scale was taken with higher scores indicating higher depression
symptoms (higher risk for depression) with the aggregate baseline mean=1.56
(SD=0.65). The Cronbach’s alpha for the scale was α=0.76, which is considered
acceptable internal reliability. A pilot study conducted with Chengdu, Wuhan, and
Qingdao 10
th
grade adolescents (n=1,388) showed a correlation of .74 between this 4-
item scale with the 20-item Center for Epidemiologic Studies-Depression (CES-D) scale
(Radloff, 1977). Depression scores were also assessed by stratifying by gender and
dichotomously coding for clinically significant symptomatic cut-off scores at top 20%

24

within gender. The cut-off scores were 2.25 for females and 2.5 for males. Females had
higher overall prevalence rates at 22.13% (n=281) compared to 18.21% for males
(n=253). The face validity of these items also suggests that they characterize depressive
mood, particularly somatic symptoms that are prevalent among depressed Chinese
(Parker, Cheah, & Roy, 2001; Parker, Gladstone, & Chee, 2001; Ryder et al., 2008; Simon,
VonKorff, Piccinelli, Fullerton, & Ormel, 1999). The terms “depression” and “depression
symptoms” are used interchangeably in this manuscript to represent the group of
adolescents at greatest risk for developing depression but we do not suggest that these
students should be considered clinically depressed.
Social Influences. Social Influences were assessed by looking at socially accepted
attitudes or beliefs regarding smoking (Social Norms) and perceived number of friends
who smoke (Friend Prevalence). Social norms were determined by asking, “Are smokers
more popular?”, “Does smoking make young people look cooler?”, and “Does smoking
make young people more attractive?” Answer choices ranged from (1) “definitely, no”
to (4) “definitely, yes”. The baseline mean for the social norm scale was 1.82 (SD=.77)
with acceptable internal reliability (Cronbach’s alpha=0. 65). Friend Prevalence was
assessed by asking “How many close male friends do you have?” and “Of your close
male friends, how many do you think smoke?” and repeated for female friends. The
mean of male and female number of friends perceived to smoke was used for the friend
prevalence variable. The range was between 0-12 friends with the mean being .618 and
standard deviation 1.61.

25

Demographic measures. General measures such as age, gender, grade, and
class/school attended were also collected.
Data Analysis
The data were analyzed for descriptive statistics from among demographic,
social influences, depression, and smoking variables. Bivariate correlations were also
examined for relationships between demographic, predictor and outcome variables.
Multi-level logistic regression was conducted with social influence indicators,
depression, and control variables regressed on 30-day smoking.
Mediation Analysis
Multi-level modeling was conducted using proc glimmix (SAS 9.1) due to the
nested structure of the data (individuals within class) and the violation of the
assumption that each observation is independent of each other. Proc mixed (SAS 9.1)
was used to assess the relationship between depression and each mediator variable,
such that the products of the a
x
and b
m
parameters for each of the mediated effects
(social norm and friend prevalence) are summed with cʹxy to specify total, direct, and
indirect effects (see eqs 1-3). Mediation effects and 95% confidence intervals were also
calculated.

Y=i
0
+c
x
X +e
y
(eq.1)
M
1
= i
0
+ a
x
X +e
M

(eq.2a)
M
2
= i
0
+ a
x
X +e
M

(eq.2b)
Y= i
0
+ c ʹ
x
X +b
M
M
1
+ b
M
M
2
+ e
y
(eq.3)

26

Moderation Analysis
A moderator effect occurs when the relationship between two variables, most
often the independent and outcome variable, differ across the levels of a third variable
(MacKinnon, 2008). A moderation effect may also be called an interaction effect and is
represented by equation 4 where c
x
is the coefficient for the relationship between the
dependent variable and the independent variable, C
z
is the coefficient for the
relationship between the moderator variable (Z) and the dependent variable, and C
xz
is
the interaction between the independent variable and moderator variable. The
independent and moderator variables are centered to reduce collinearity and assist with
interpretation of the interaction term. If the interaction term is statistically significant,
then one may say that the relationship between X and Y differ across levels of Z and
further analyses may be conducted to assess how those relationships differ. This study
assessed whether the relationship between social influence indicators and smoking
behaviors varied by depression symptoms.
RESULTS
Baseline sample characteristics of the Wuhan dataset are presented in Table 1.
Males and females were significantly different across smoking, depression symptoms,
and social norm measures.

Y=i
0
+c
x
X +c
z
Z+c
xz
XZ+e
y
(eq.4)

27

Simple correlations between all variables of interest illustrated significant
associations in the expected directions (Table 2). Social norms appeared to have greater
correlation with depression symptoms (r=.2179) than smoking (r=.1641).

Table 1. Baseline Sample Characteristics (n=2661)
Males (n=1391) Females (n=1270) Gender diff
N % n % t-test p-value
Condition
Program 688 50.54 636 50.08 .32 .75
Control 703 49.46 634 49.92
Smoking Outcome
Lifetime 604 43.42 235 18.52 -14.48 <.0001
30-day Smoke 282 20.29 82 6.46 -10.80 <.0001

mean sd mean sd
Depression 1.56 0.65 1.76 .79 -3.77 <.0001
Perc. Friend Smk .78 1.81 .438 1.34 -5.61 <.0001
Social Norm Beliefs 1.81 0.76 1.74 .67 -2.72 .007
Table 2. Simple Correlations with the Wuhan Male Baseline Sample
Depression 30 day Smoke Social Norms Friend
Prevalence
Depression 1
30 Day Smoke .1559*** 1
Social Norms .2179*** .1641*** 1
Friend Prevalence
.1637*** .447*** .2378*** 1
* p<.05; ** p<.01; ***p<.001

28

Among males, positive social normative beliefs about smoking (p=.003),
perceptions of friend smoking (p<.0001), depression symptoms (p.03) and age (p=.0002)
were significantly associated with concurrent smoking in the past month. For mediation
analysis, the first step was to establish a significant path between depression and
smoking (c
xy
). Although the correlation between depression and smoking was
significant, the regression model was not able to converge, thus mediation could not be
assessed using procedures outlined by Baron & Kenny (1986). However, statistical
significance was assessed using the Sobel Z- test to determine if the mediation pathway
was different from zero (Krull & MacKinnon, 1999; MacKinnon & Dwyer, 1993). Table 3
reports the beta estimates for each of the multivariate models used for mediation
analysis and is graphically represented in Figure 1.
One unit change in social norms was associated with a .2502 increase in smoking
and a 1 unit increase for friend smoking was associated with a .4944 increase on
smoking. The adjusted effect of depression on smoking remained significant (c'=.3242,
p=.001) which suggests that this is only a partial mediation. One unit change in
depression was associated with a .3675 effect on smoking through the social influences
mediators. Table 4 reports the specific and total effects as calculated with Sobel Z-test
statistic.

29

Table 3. Wave 1 (Baseline) Models with Wuhan Males
Model 1: 30-day smoking outcome
Independent variables Beta SE p-value
Social Norm (M
1
) .2242 .085 .008
Friend Smoke (M
2
) .4990 .046 <.0001
Depression .3242 .101 .001
Age .4112 .108 .0001

Model 2: Social Norm Outcome (M
1
)
Independent variables Beta SE p-value
Depression .2564 .031 <.0001
Age .0681 .033 .04

Model 3: Friend Smoke Outcome (M
2
)
Independent variables Beta SE p-value
Depression .4688 .072 <.0001
Age .2148 .076 .005

*p<.05; Standard Errors are in parentheses.
Friend
Prev.
M
2

.4688*

.4990*
(.072)
(.046)
Depression
symptoms
X
30-day
Smoking
Y
Social
Norms
M
1

c’=.3242*
.2242*
.2564*

(.031)

(.085)
(.101)
Figure 1: Multiple Mediation Model

30

Depression as a moderator of the relationship between social influences and
smoking were explored. The interaction terms for depression symptoms x social
normative beliefs and depression symptoms x perceived friend prevalence were entered
into the model separately, but were not found to be significant (Table 5).

Table 4. Mediation of Depression symptoms on 30-Day Smoking Through Social Norm
Beliefs and Perceptions of Friend Use.

Point
Estimate

Product of Coefficients

SE Z p-value
Indirect Effects
Social Norms (a
1
b
1
) 0.073

0.0278 2.62 0.0088
Perceived Friend Use
(a
2
b
2
) 0.2945

0.0662 4.454 <.0001
TOTAL Indirect Effect
(a
1
b
1
+a
2
b
2
) 0.3675

0.0709 5.184 <.0001

31

Table 5: Wave 1 (Baseline) Moderation Models with Wuhan Males
Model 4: Social Norm X Depression Moderator
Beta SE p-value
Intercept -1.584 .1805 <.0001
Social Norm .3265 .1015 .001
Friend Smoke .3032 .2438 <.0001
Depression .4523 .1654 .006
Age .3878 .1138 .0007
Social Norm x Depression -.2892 .1943 ns

Model 5: Friend Smoke X Depression Moderator
Beta SE p-value
Intercept -1.506 .1869 <.0001
Social Norm 0.2500 .0840 .003
Friend Smoke .4942 .0509 <.0001
Depression .3538 .1620 .03
Age .3974 .1075 0.0002
Friend Smoke x Depression .0034 .1163 ns

DISCUSSION
The current study provides evidence supporting that depression symptoms,
social influences, and smoking are all significantly associated. Specifically, in support of
hypothesis 1.1 and 1.2, higher levels of depression symptoms were significantly
associated with higher positive social norm beliefs about smoking and perceiving higher
levels of friend smoking prevalence. Depression symptoms and social influence factors
were also significantly associated with concurrent smoking behaviors. The findings from
the current study also support partial mediation effects through social influence factors
between depression symptoms and smoking behaviors. In other words, the relationship
between depression symptoms and smoking may be explained, in part, by the
individual’s smoking social norm beliefs and perceptions of their friends’ smoking

32

behaviors. Cognitive theories of depression suggest that depressed individuals view
their world in negative ways. However, depressed individuals may be biased to perceive
their social world differently rather than simply “negatively.” It is our argument here
that adolescents who are at risk for depression may be biased in perceiving social cues
that confer rejection or acceptance by peers. Smoking related cues may be a
particularly strong vehicle for transmitting rejection or acceptance by peers. A prior
study demonstrated that adolescents may use substance use to gain entry or
acceptance into a social group (Henry & Kobus, 2007). Our findings are also supported
by cognitive theories of depression in which perceptions of the social environment may
be biased. In our case, recognizing smoking cues as normative or as a tool for gaining
positive attention from other peers may be greater among those with depression
symptoms or they may feel more need to comply with their observed norms than others
(Katkin, Sasmor, & Tan, 1966). Another study indicated that depression symptoms were
significantly associated with perceptions of peer approval for smoking behaviors and the
effect between depression symptoms and smoking were partially mediated by
perceptions of peer approval (Ritt-Olson et al., 2005). These studies further support our
hypothesis that adolescents at risk for depression are particularly vulnerable to their
social environment. Our mediation analyses are consistent with these prior studies and
provide more evidence on the specific social influence processes related to smoking
behaviors among those at risk for depression.

33

Moderation analyses did not provide significant results thus we cannot reject the
null hypotheses as expected in hypothesis 1.3. Our conjecture is that the relationship
between social influences and smoking behaviors may not differ across levels of
depression symptoms. It may be that those with higher levels of depression symptoms
are more aware of smoking related cues and thus these higher levels of perception lead
directly to higher levels of smoking behavior, but that the relationship is different for
those with depression symptoms. Another exploratory study in a different adolescent
population also failed to find significant interaction results for depression symptoms x
perceptions of friend approval on smoking behaviors (Ritt-Olson et al., 2005). This gives
confidence in the overall result of the current study that the role of depression on
smoking behavior is mediated through social influence factors rather than depression
symptoms changing the relationship between social influence factors and smoking.
Limitations and Future Directions
The limitations of the current study must be considered in conjunction with our
findings. This study was conducted using cross-sectional data and therefore the findings
do not connote true mediation. Although we are confident that the model was
specified correctly based on literature and theory, only a randomized controlled trial
with multiple data points will be able to demonstrate mediation conclusively.
Interventions that target changes in perceptions of social norms and peer influences
may provide the opportunity to test these pathways further.

34

Prior studies have largely looked at these depression symptoms, social influences
and smoking relationships among United States or European youth. This is the first
study, to our knowledge, that examined these social influence pathways among
Mainland Chinese youth. There may be cultural differences in social normative beliefs
about smoking, especially differences in gender roles for smoking behaviors (Wen et al.,
2007). However, prior cross-cultural studies that compared Chinese youth and U.S.
California youth suggested that risk factors for smoking appeared similar (Unger et al.,
2002). Peer influences remain among the strongest predictors for smoking behaviors in
both U.S. and Chinese Adolescents (Grenard et al., 2006; Unger et al., 2001; Unger et al.,
2002; Zheng et al., 2007). However, gender roles and expectations need to be
considered more carefully due to the rapid social and economic development occurring
in China. Historically, being female was considered a protective factor against smoking
due to the strong social norms and sanctions against a woman smoking. With
globalization and westernization, the gap that once existed between male and female
smoking is diminishing rapidly with the increasing prevalence of cigarette use among
female youth (Chen, 2007; Cheng, 1999). Future studies need to monitor gender effects
on protective and risk factors for the etiology of smoking in China, particularly in
relation to stress and emotional disturbances such as depression. Our current study is
only the beginning in understanding the risk factors and pathways involved in cigarette
use among Chinese males. These findings cannot be generalized to other populations.

35

Future studies will need to examine longitudinally whether depressed
adolescents have biased perceptions which promote smoking or if they select pro-
smoking peers and environments. This distinction will indelibly contribute to prevention
and cessation interventions for smoking and other substances that are generally
promoted through social norms and influence. Specifically, we will be able to
understand why depression symptoms are risk factors for smoking and how
interventions work or don’t work for certain subpopulations. This will lead to new
developments for targeted interventions, more effective tailoring of messages in
universal programming, and insight on the etiology of specific co-morbid conditions
such as smoking and depression.

36

CHAPTER 3: Evaluating Depressive Symptom Interactions on Smoking Prevention
Program Mediators: A Mediated Moderation Analysis

ABSTRACT
Evidence-based smoking prevention programs that utilize social influences to
enact behavior change have often been successful in their objectives. However, recent
research has shown that program effects are conditional on dispositional and social
environmental circumstances. These programs and their effects offer a unique
opportunity to understand the cognitive processes involved in depression, social
influences, and smoking. The current study examined the meditational pathway design
hypothesized for the Wuhan Smoking Prevention Program to change smoking behaviors
and how depression symptoms might impact those pathways. Among 1391 7
th
-grade
Chinese adolescent boys, those with high depression scores who had experienced
smoking in the past were found to be more likely to change their perceptions of their
friends smoking prevalence resulting in lower smoking rates one-year after program
implementation. As hypothesized, it was the change in perception that accounted for
the successful program effects shown specifically among the depressed group.
Understanding the mechanism responsible for these differential program effects will
contribute to future intervention development.

37

INTRODUCTION
Adolescents learn the rules of social engagement and choose behaviors based on
their assessments of their social environments (Bos, Sandfort, de Bruyn, et al., 2008;
Deb, Mitra, & Mukherjee, 2001; Makri-Botsari, 2005; Nelson, Leibenluft, McClure, et al.,
2005). Smoking during adolescence is a prime example of how peers influence behavior
(Kobus, 2003). Evidence based smoking prevention programs target social influences by
changing perceptions of smoking prevalence rates, changing pro-smoking norms and
beliefs, and increase the perception that smoking is not positively viewed by their peers
(Chou et al., 2006; Graham, Marks, & Hansen, 1991; Hansen & Graham, 1991).
There has been some debate as to whether social influences prevention
programs are generally effective (Botvin, Sussman, & Biglan, 2001; Peterson, Kealey,
Mann, Marek, & Sarason, 2000; Sussman, Hansen, Flay, & Botvin, 2001; Thomas, 2002).
The literature has shown long term effects in the prevention of substance use (Skara &
Sussman, 2003; Thomas & Perera, 2006; Tobler et al., 2000) but more importantly, there
has been great advancement in the understanding of how these programs work and for
whom (Graham, Marks, & Hansen, 1991; Hansen & Graham, 1991; Johnson et al., 2007;
MacKinnon & Luecken, 2008; MacKinnon, Weber, & Pentz, 1988; Sun et al., 2007; Unger
et al., 2004). Yet, while these studies show that programs’ main effects differed across
different subgroups or that the intended targets of the program mediated the program
effects for the average participant, they do not account for the mechanisms that

38

produce program differences between at-risk subgroups such as those who score high
or low on depression.
Depression appears to be a particularly robust risk factor for adolescent smoking
(Audrain-McGovern, Lerman, Wileyto, Wileyto, & Rodriguez, 2004; Fergusson, Goodwin,
& Horwood, 2003; Goodman & Capitman, 2000; Munafo, et al., 2008; Waller et al.,
2006; Wang, et al., 1996). Prevention programs may be particularly powerful among
this high-risk group and understanding the mechanisms behind this relationship might
provide valuable insight for future prevention program development. In a randomized
control trial conducted in the United States, a culturally tailored social influences based
smoking prevention program was delivered to 7
th
grade adolescents and was found to
be successful at 1-year follow-up. Program effects were found to vary by cultural
context (Unger et al., 2004). There was also evidence that the effects of the program
were strongest among those with high scores on depression symptoms and/or hostility
(Johnson et al., 2005). Another randomized controlled trial that implemented a social
influences-based program found the program to be most effective among males who
had higher scores of depression symptoms and who have previously experimented with
cigarette use (Sun et al., 2007). The program effect for these high-risk males was 4.2
times (95% CI 1.45-11.76) larger than the program effect among their low-risk
counterparts. However, it was not clear why the program worked especially well in the
high-risk subsample or why depression and smoking might render one especially
receptive to social influence messages.

39

The current study will examine the relationships between depression symptoms,
social influences, and smoking in the context of a social influences based prevention
program designed to manipulate smoking related cognitions. This analysis will test
hypothesized mechanisms responsible for the stronger program effects observed for
those with higher depression scores (Sun et al., 2007). Study 1 provided evidence to
support the hypothesis that adolescents at risk for depression may be more sensitive to
social information surrounding cigarette use. Specifically, the relationship found
between depression and smoking may, in part, be attributable to social influence factors
such as perceived friend use prevalence rates of smoking and pro-smoking normative
beliefs. Other studies also provide support for this hypothesis. One study
demonstrated that only among those who reported that most of their friends smoked,
depression and anxiety increased the risk of becoming a smoker by almost three times
(Patton et al., 1998). Another cross-sectional study found that adolescents who scored
high on advertising receptivity and depressed moods were most vulnerable to
experimenting with smoking (Tercyak, et al., 2002). Risk was even greater when
exposed to peer smoking. The most interesting finding from this latter study was that
the effects of advertising on smoking behavior were heightened among the depressed
adolescents but exposure to other smokers (e.g. family, peers) by itself did not increase
the risk for being a smoking. This suggests that simply being around others who smoke
may not increase the risk of smoking onset for depressed adolescents. Rather, the
increased risk may be due to how depressed adolescents perceive smoking, such as

40

providing positive attention from peers or as providing more overall interaction with
peers. Normative information from advertising, from social contexts, or from peer
interactions may influence smoking behaviors and among depressed adolescents, it may
strengthen that influence. Cognitive theories of depression (see Lakdawalla et al., 2007)
generally support the hypothesis that perceptions of social normative beliefs (what
attributes smoking will bring to the individual or what attributes the individual believes
others see in smokers) and perceptions of friend use prevalence (how many of their
friends smoke or approve of smoking) may be heightened in a depressed individual
because of the potential for rejection or acceptance by peers. These depressed
individuals may be sensitive to these cues in the environment and may be more
motivated to act in a way that might bring them elevated standing among their peers.
The current study is, to our knowledge, the first to investigate the relationships
among depression symptoms, social influence, and smoking in a longitudinal cohort.
More importantly, this study is the first to use a randomized trial to isolate and test
whether changes in targeted social influence perceptions (via social-influence smoking
prevention program) will change smoking behavior among those with high scores of
depression. If depressed adolescents are more aware of cues in their social
environment, then a program aimed to change or correct perceptions would hold more
weight and result in greater changes in behavior for them than for their non-depressed
counterparts.

41

Specific Aims and Hypotheses
The aims are to:
2.1 assess whether the interaction between program condition and depression
symptoms is related to perceptions about smoking.
Hyp.2.1.a Because the WSPT program was developed to change perceptions
of social normative beliefs and perceived peer prevalence, it is
expected that both social influence indicators would be lower in
the program group compared to the control group.
Hyp.2.1.b Because social influences and depression are hypothesized to be
positively associated, it is expected that the WSPT would reduce
depression symptoms in the program group compared to the
control group but in a smaller magnitude that would not likely be
statistically significant.
Hyp.2.1.c The program effect, or in other words, the relationship between
program and wave 2 smoking behaviors will differ by levels of
depression symptoms. Students who received the program,
versus those who did not, will have lower levels of smoking and
will vary by individual levels of depression.
2.2 evaluate the effects of the interaction between program condition and
depression symptoms on Wave 1 social influence factors in predicting wave 2
smoking behaviors, we propose hypothesis 2.2.a.

42

Hyp.2.2.a Initial levels of depression symptoms may determine levels of
initial social influence factors and whether they were distributed
equally between program and control conditions. It is expected
that positive social norm beliefs about smoking and perceived
prevalence would be higher among depressed adolescents but
equal between program and control groups due to randomization
procedures.
2.3 test a mediated moderation model that assesses whether the effect between
program and social influences vary by levels of depression in its role in predicting
smoking outcomes, we propose hypothesis 2.3.a
Hyp.2.3.a Cognitive theories of depression suggest that individuals with
depression perceive or process social information differently than
those who are not depressed. Therefore, the relationship
between program and social influences (social normative beliefs
and perceived friend prevalence) will be stronger among those
who score high on depression symptoms compared to those with
lower scores. The interaction between program and social
influences was expected to be positively associated with smoking
behaviors. A significant moderation on the first arm of this
mediated pathway would establish the mediated moderator
model as hypothesized.

43

METHODS
Sample Selection
The Wuhan Smoking Prevention Trial (WSPT) was a school-based longitudinal,
randomized controlled trial aimed at preventing initiation and escalation of adolescent
smoking behaviors (Chou et al., 2006). Middle schools were randomly selected from
each of the seven urban districts within the city of Wuhan and then additional schools
within each district were matched on school size, teacher/student ratio, and academic
rankings to create pairs. One school from each pair was then randomly selected to
program or control conditions (see Unger et al.,2001). Seventh grade students were
assessed with a 200-item paper-and-pencil baseline survey prior to program
implementation in 1999 and were re-administered one year later.
The WSPT curriculum was modified from an evidence based social influences
substance use prevention program in the United States: Project SMART (Graham,
Johnson, Hansen, et al., 1990; Graham, Marks, & Hansen, 1991). The program aimed to
establish a social norm that smoking was unacceptable among their peers and that
there were serious social and physical consequences to smoking (Chou et al., 2006). All
study protocols were approved by the Institutional Review Boards of the University of
Southern California and the Wuhan Center for Disease Control and Prevention. All
participants were actively consented.
The data collected for WSPT were used for Study 2. The sample was limited to
males with complete Wave 1 (baseline) and Wave 2 (1-year follow-up) data. A social-

44

influences smoking prevention program was implemented after the administration of
Wave 1 data collection but before Wave 2 data collection.
Curriculum
A 14-session 45 minute classroom based social influences program was
implemented in the program condition arm of the WSPT. Trained health educators
delivered sessions to the program schools, while control schools maintained standard
care or normal academic activities. The WSPT curriculum was based on substance abuse
prevention program named Project SMART which was found to be effective in Southern
California, Kansas/ Missouri, and Indiana (Graham, Johnson, Hansen, et al., 1990;
Hansen, Graham, Wolkenstein, et al., 1988; MacKinnon et al., 1991; Pentz et al., 1989).
Project SMART and the WSPT program aimed to foster an anti-smoking social norm
among adolescents and teach them skills to resist pro-smoking influences. Translation
and tailoring of curriculum components were done to enhance the relevance of the
WSPT program to Chinese adolescents (see Sun et al., 2007). All sessions were delivered
by trained health educators from the Wuhan City Center for Disease Prevention and
Control, with support and guidance from the USC Institute for Health Promotion and
Disease Prevention Research and Project SMART health educators.
Measures
Smoking. Lifetime smoking was assessed by asking, “Have you ever smoked a
cigarette, even just a few puffs?” (0=no; 1=yes). Thirty-day prevalence was assessed by
asking “Think about the last 30-days. On how many of those days did you smoke

45

cigarettes?” (0=smoked<1 day in the past month; 1=smoked at least 1 day). The
reference group for30-day smokers are those individuals who have never smoked, quit
smoking, and ever smokers combined.
Depression. Depression was assessed at wave 1 with four items that asked students
“Have you felt depressed in the last week?”; “Have you felt alone in the last week?”;
“Have you felt sad in the last week?”; and “Have you felt like crying out in the last
week?” Response choices ranged from (1) Never to (4) Often with higher scores
indicating higher symptoms of depression. The depression score was the mean of the
four items with the standardized Cronbach’s alpha=.76 indicated satisfactory internal
consistency. A pilot study conducted with Chengdu, Wuhan, and Qingdao 10
th
grade
adolescents (n=1,388) showed a correlation of .74 between this 4-item scale with the
20-item Center for Epidemiologic Studies-Depression (CES-D) scale (Radloff, 1977,
1991). Depression was dichotomized, where those who scored at the top 20% of
symptoms (score =>2.5) will be coded as being high-risk for depression (1) versus low-
risk (0).
Social Normative Beliefs. Social normative beliefs are defined as socially accepted
attitudes or beliefs regarding smoking. This was assessed by asking students the
following three questions: “Are smokers more popular?”; “Does smoking make young
people look cooler?”; and “Does smoking make young people more attractive?”
Response choices ranged from (1) Definitely, not to (4) Definitely, yes. The mean was

46

taken for the three items (mean=1.82, SD=.77) and the wave 1 standardized Cronbach’s
alpha was .65 which represents acceptable internal reliability.
Perceived Friend Prevalence. Friend prevalence was assessed by first asking students
how many male/female friends they have and then, of those friends, the number of
friends they think smoked in the past. The mean number of male and female friends
that smoke was used for the perceived friend prevalence variable. Friend prevalence
estimates were significantly higher among 30-day smokers (t-value= -10.38, p<.0001)
and among those at highest risk for depression symptoms (t-value=-3.60, p<.0004).
Demographic measures. General measures such as age, weekly allowance,
academic performance, grade, and class/school attended were also collected.
Data Analysis
Multilevel random coefficients models will be used to test the study hypotheses
because the randomization and experimental unit was at the school level and not at the
individual level. Statistical package and procedures, SAS Proc Glimmix and Proc Mixed,
will be used for analyses. This method will set the degrees of freedom for the
independent intervention units and control for observation dependence due to the
nested structure of the data (students within schools).
Attrition Analyses
The propensity score analysis technique (Austin, Grootendorst, & Anderson,
2007; Grunkemeier, Payne, Jin, & Handy, 2002) was used in prior reports on this cohort

47

(Sun et al., 2007) and duplicated in this study in order to control statistically for possible
bias due to unbalanced attrition between the program conditions. Propensity for
attrition scores were estimated for each participant from logistic regression on boys
who participated in the wave 1 survey. Age, number of days smoked in the last 30-days,
academic performance, weekly allowance, hostility, depression, and program condition
were used to predict attrition status (whether they were reassessed 1-year later). Only
academic performance and weekly allowance were found to be significant predictors of
1-year attrition. The propensity for attrition score was calculated by regressing attrition
on program, gender, program*gender, age, depression, 30-day smoking status,
education, academic grade, and weekly allowance. The scores from this model
(ŷ=attrition) were output to create the propensity for attrition score. This propensity for
attrition score was treated as a confounder in the analyses so that the estimated
program effect could be interpreted as if there were balanced attrition rates between
the program and control conditions (Graham & Donaldson, 1993; Hansen, Tobler, &
Graham, 1990).
Mediation and Moderation Analysis
Mediation was assessed using the Causal Steps Approach as outlined by Baron &
Kenny (Baron & Kenny, 1986; MacKinnon, Lockwood, Hoffman, et al., 2002). Equations
1 through 4 represent the mediation steps that will be conducted. The program
condition and the two mediators: Perceived friend prevalence and social normative
beliefs will be regressed on outcome variable, Wave 2 30-day smoking. Propensity for

48

attrition will also be controlled in the model. Because we are testing whether the
program had an effect on the targeted social influence mediators, we will be using wave
2 adjusting for wave 1 variables, capturing change scores. This mediation model will
assess whether the program reduced social influences indicators in the program group
as expected, whether the program reduced smoking as expected, and whether social
influences indicators mediated the relationship between the program and smoking
outcomes.
Moderation analysis will also be assessed (Eq. 4). This study will assess whether
the relationship between program and smoking differed by depression symptoms. We
will also assess whether the relationship between program and Wave 2 social influence
indicators (social normative beliefs and perceived friend prevalence) were moderated
by depression symptoms while adjusting for baseline smoking, age, propensity for
attrition, and baseline social influence levels.

Y=i
0
+c
x
X +e
y
(eq.1)
M
1
= i
0
+ a
x
X +e
M

(eq.2a)
M
2
= i
0
+ a
x
X +e
M

(eq.2b)
Y= i
0
+ c ʹ
x
X +b
M
M
1
+ b
M
M
2
+ e
y
(eq.3)
Y=i
0
+c
x
X +c
z
Z+c
xz
XZ+e
y
(eq.4)

49

Mediated Moderator Analysis
Understanding mediation in the presence of a moderator is complex. A
Mediated moderator effect is present when the effect of an interaction on a dependent
variable is mediated. For example, the effect of a program (X) depends on the level of
depression symptoms (Z) which changes social norms (M) and, in turn, affects smoking
behavior (Y; Morgan-Lopez & MacKinnon, 2006).
We examined the data using the mediated moderator effects model due to
theoretical considerations. Although it is plausible that the mediation effect varies by
depression symptoms (as a moderated mediator effect would demonstrate), it would be
more accurate to consider the interaction between a predisposition for depression
symptoms (Z) and the program (X) to affect how the social influences mediators (M
1
and
M
2
) are perceived. The relationship between social influences and smoking behaviors
was not expected to differ by depression symptom. Rather, the program effect on
perceptions of social influences was expected to vary according to depression
symptoms. Cognitive theories of depression provided the theoretical and empirical
support that suggested that perceptions of social information and how they are
M= i
0
+ a
x
X +e
M

Y = i
0
+ c ʹ
x
X +b
M
M + e
y
=i
0
+ c ʹ
x
X +b
M
(i
0
+ a
x
X +e
M
) + e
y
= i
0
+ c ʹ
x
X +b
M
i
0
+ b
M
a
x
X + b
M
e
M
+ e
y

= i
05
+ b
M
i
0
+ c ʹ
x
X +b
M
a
x
X + b
M
e
M
+ e
y
= i
0
+ b
M
i
0
+ (c ʹ
x
+b
M
a
x
)X + b
M
e
M
+ e
y
(eq.5)

50

processed differ according to depression symptoms. Therefore, it is expected that levels
of depression will interact with the program resulting in differences in social influence
indicators which then affect smoking behavior.
Edward and Lambert’s (2007) proposed Path Analytical framework was used to
reconstruct the logic behind each of the models and while Muller, Judd, and Yzerbyt’s
(2005) proposed framework, a direct extension of the Causal Steps Approach proposed
by Baron and Kenny (1986), was used to assess the presence of a mediated moderator
pathway. Because we are interested in the interaction between Z (predisposition for
depression) and X (program condition) on M (social influence constructs targeted by the
program), this effect was examined using the regression Equation 6, where the
coefficient a
xz
represents the extent to which the path between X and M vary according
to Z. Equation 7 represents the direct effect of X (cʹ
x
) on Y, the relationship between M
and Y (b
m
), and the interaction between X and Z (b
xz
) on Y. While equation 7
demonstrates whether there is an overall moderation between X and Y, it does not yet
show us whether the interaction between X and Z affect M. To do this, first we use our
regression model (eq. 6) that represents our moderation on the path between X and M
and substitute the mediator (M) in a basic Mediation model (Eq.3). The reduced model
(Eq. 8a) now looks like a moderation model in which the coefficients, cʹ
x
+ b
M
a
x
,

represents the direct (cʹ
x
) and indirect (b
M
a
x
) paths between X and Y (Edwards &
Lambert, 2007). Muller’s approach departs from the Edwards by including another
interaction term M x Z to assess whether a change in the mediator effect on Y is

51

detected as the moderator increases, while Edwards uses separate equations to
calculate different models (Edwards & Lambert, 2007; Muller, et al., 2005). For our
purposes, we will use Muller’s model as represented in Eq. 8b.

Equation 8b represents a simplified version of a single mediator with no control
variables. This study assessed mediated moderation on two mediator variables (M
1
:
Social Normative Beliefs; M
2
: Perceived Friend Prevalence) and controlled for propensity
for attrition. Finally, Muller et al. demonstrated that an equality exists between
parameter estimates (Eq. 9) and was used to investigate mediated moderated
pathways. All analyses were conducted using Proc Glimmix or Proc Mixed (SAS 9.1).
Figure 2 represents graphically the model that was tested.
M =i
0
+a
x
X +a
z
Z+a
xz
XZ+e
m
(eq.6)
Y = i
0
+c ʹ
x
X +c
z
Z+c
xz
XZ+b
m
M+ e
y
(eq.7)
Y = i
0
+ c ʹ
x
X +b
M
M+ e
y

=i
0
+ c ʹ
x
X +b
M
(i
0
+a
x
X +a
z
Z+a
xz
XZ+e
m
)+ e
y
=i
0
+ c ʹ
x
X +b
M
i
0
+ b
M
a
x
X + b
M
a
z
Z+ b
M
a
xz
XZ+ b
M
e
m
+ e
y
=i
0
+b
M
i
0
+ (c ʹ
x
+ b
M
a
x
)X + b
M
a
z
Z+ b
M
a
xz
XZ+ b
M
e
m
+ e
y
(eq.8a)
Y==i
0
+b
x
X +b
z
Z +b
xz
XZ + b
m
M +b
mz
MZ +e
y
(eq.8b)
c
xz
-

b
xz =
b
m
c
xz
+b
mz
cʹ
x
(eq.9)

52

RESULTS
The sample characteristics and attrition analyses are presented in Table 5. There were
1391 males surveyed at wave 1 and 1248 were successfully retained for wave 2
assessment (90% retention rate). Attrition was significantly higher in the program group
(13.1% vs. 7.5%; χ
2
=11.58, p<.0007). Attrition was also significantly higher among
baseline 30-day smokers and perceived friend prevalence (see Table 6). In a logistic
regression model with attrition regressed on experimental condition, 30-day smoking,
depression, friend smoking, and social norm beliefs, the condition x 30-day smoke was
Smoking
Y
Social Norm
Beliefs
M
1

Perceived
Friend
Prevalence
M
2

WSPT Program
X
Depression
symptoms
Z
1

Note: Control variables are age and propensity for attrition
Legend
=a
xz
b
m1

= a
xz
b
m2

Figure 2. Mediated Moderator Model

53

not statistically significant (beta=.9326, p=.08) suggesting that there were no differential
attrition of smokers between conditions.

Among males, a one-year follow up for 30-day current smoking was predicted by
past 30-day smoking (p<.0001), wave 2 (adjusted for wave 1) perceived social norms
(p=.0097), and wave 2 (adjusted for wave 1) perceived friend prevalence (p<.0001).
There was no significant main effect for program on Wave 2 30-day smoking behaviors
(Table 7). A program x depression interaction term was included in the model to test for
moderation on the direct effect between program and wave 2 smoking but the term
was non-significant (b=.0429, SE=.3594).
Table 6. Sample Characteristics
W1 Males (n=1391) W2 Males (n=1248) Attrition difference
n % N % χ
2
p-value
Condition
Program 688 50.54 598 86.9 11.58 <.0007
Control 703 49.46 650 92.5
Smoking Outcome
W1 Lifetime 604 43.42 532 42.6 3.11 .07
W1 30-day Smoke 282 20.29 243 19.5 4.81 .03

mean sd Mean Sd t-value p-value
Depression 1.56 0.65 1.55 .645 -.99 .323
Perc. Friend Smk .78 1.81 .738 1.75 -2.22 .028
Social Norm Beliefs 1.81 0.76 1.81 .756 -1.14 .255

54

Table 7. Wave 2 30-day smoking
beta SE p-value
Intercept -1.918 .2992 <.0001
W1 30dsmk 1.284 .1662 <.0001
Program -.2131 .4106 ns
W2 Social Norms .2441 .0942 .0097
W2 Friend Smoke .2614 .0331 <.0001
W1 Social Norms .1315 .0964 ns
W1 Friend Smoke .0568 .0378 ns
W1 Depression .1608 .1842 ns
Propensity Score 7.366 1.669 <.0001

To begin the mediated moderation analyses, we first examined whether there
were any effects of the program predicting wave 2 social influence indicators. It
appears that the program does not predict either social norms or perceived friend
prevalence (Table 8: models 1 and 3, respectively). We explored whether the program
effect on the social influences indicator differed according to levels of depression
symptoms. It appears that depression did not moderate between program and social
normative beliefs (Table 8: model 2). Although not statistically significant, there may be
evidence that the program’s effect on perceived friend prevalence was moderated by
depression (p=.08; Table 8: Model 4). The program effect on each of the proposed
mediators was not robust enough to test the full mediated moderation model further.

55

Table 8: Wave 2 Social Influence Models
Model 1: Wave 2 Social Norms outcome
Independent Variables Beta SE p-value
W1 Social Norm .3325 .0279 <.0001
Program .0131 .0415 ns
Depression 0489 .055 ns
Propensity score 1.59 .4848 .001

Model 2: Wave 2 Social Norms Outcome with
program X depression
Independent Variables Beta SE p-value
W1 Social Norm .3341 .0458 <.0001
Program -.003 .0458 ns
Depression .0008 .0790 ns
Propensity score 1.56 .486 .001
Program x Depression .092 .109 ns

Model 3: Wave 2 Friend Prevalence Outcome
Independent Variables Beta SE p-value
W1 Friend Prevalence .2636 .0299 <.0001
Program .0238 .1001 ns
Depression .0429 .1318 ns
Propensity score 9.209 1.196 <.0001

Model 4: Wave 2 Friend Prevalence Outcome with
program X depression
Independent Variables Beta SE p-value
W1 Friend Prevalence ..2636 .0299 <.0001
Program .1044 .1106 ns
Depression .2706 .1872 Ns
Propensity score 9.340 1.197 <.0001
Program x Depression -.4469 .2613 .087

Post hoc analyses were conducted to assess whether the program was only
relevant to depressed adolescents who have tried smoking in the past. These analyses
are justifiable since the strongest program effects were for secondary prevention (Chou

56

et al., 2006) and were found among a special group with pre-existing conditions
(smoking and depression) prior to the intervention (Sun et al., 2007). This co-morbidity
between smoking and depression symptoms may place an individual at increased risk
for further disease development above and beyond any single morbidity (Sun et al.,
2007). Thus, the post hoc analyses considered the co-morbidity (CoM: where
depression=1 and wave 1 30-day smoke=1 vs. non-CoM where depression=0 and wave 1
30-daysmoke=0) in place of depression symptoms in each model.
Similar to initial findings, prior wave 1 smoking (p<.0001), wave 2 social norms
(adjusted for wave 1 social norms; p=.02), wave 2 friend prevalence (adjusted for wave
1 friend prevalence; p<.0001), and program (p=.01) significantly predicted wave 2 30-
day smoking but the CoM term did not. We examined whether there were any effects
of the program predicting wave 2 social influence indicators with CoM replacing
depression symptoms in each model. It appears that the program does not predict
either social norms or perceived friend prevalence (Table 8: models 1 and 3,
respectively). We explored whether the program effect on the social influences
indicator differed according to levels of CoM. It appears that CoM did not moderate
between program and social normative beliefs (Table 8: model 2). However, there was
a significant interaction effect for perceived friend prevalence as moderated by
program x CoM (p=.02; Table 9: Model 4). The social norm mediator was dropped from
further mediated moderation analyses.

57

Table 9: Wave 2 Social Influence Models
Model 1: Wave 2 Social Norms outcome
Beta SE p-value
Intercept -.0149 .073 ns
W1 Social Norm .3313 .0280 <.0001
Program .0207 .1268 ns
CoM .2097 .1031 .04
Propensity score 1.357 .5057 .007

Model 2: Wave 2 Social Norms Outcome with
program X CoM
Beta SE p-value
Intercept -.0129 .0735 ns
W1 Social Norm .3326 .0281 <.0001
Program .0156 .1272 ns
CoM .1403 .1659 ns
Propensity score 1.346 .5062 .008
Program x CoM .1095 .2050 ns

Model 3: Wave 2 Friend Prevalence Outcome
Beta SE p-value
Intercept -.0176 .1777 ns
W1 Friend Prevalence .2486 .0305 <.0001
Program -.0527 .3057 ns
CoM .6860 .2496 .006
Propensity score 8.444 1.231 <.0001

Model 4: Wave 2 Friend Prevalence Outcome with
program X CoM
Beta SE p-value
Intercept -.0444 .1778 ns
W1 Friend Prevalence .2423 .0305 <.0001
Program .0141 .3065 ns
CoM 1.384 .3913 .0004
Propensity score 8.595 1.231 <.0001
Program x CoM -1.116 .4825 .02

*All models were controlled for wave 1 30-day smoking
** All models used Proc logistic/glm and school ID was added to the
model to help control for school level effects.

58

Mediated moderation analysis was conducted with the perceived friend
prevalence as the mediator and CoM as the moderator. Table 10 reports the results of
the regression model and Table 11 reports the equations used to establish mediated
moderation.
Table 10: Mediation and Moderation on Wave 2 30-day Smoking
Model 5: Perceived Friend Prevalence (mediator) and CoM
(moderator)
Beta SE p-value
Intercept -2.09 .1159 <.0001
W1 30-day smoking 1.463 .2284 <.0001
Program -.9886 .3932 .01
CoM .2076 .3903 Ns
Program x CoM .6430 .3484 .07
Perc. Friend Prev. .2928 .0433 <.0001
Perc. Friend Prev. x CoM -.0640 .1380 Ns

*Model adjusted for wave 1 perceived friend prevalence,
propensity score, and school.

Without using mixed models, the overall treatment effect was significant (Table
10; Eq.1; p=.05) which fulfilled the first condition for mediated moderation analyses.
The program did not significantly predict wave 2 Perceived Friend Prevalence
(Mediator). However, the interaction term Program x CoM was significant (a
xz
=-1.116,
p=.02) which established the program effect on the smoking outcome as a function of Z
(see Table 10, eq.6). The mediator’s effect on the outcome (Table 10: Eq 8b. b
m
=.2928,
p<.0001) was also significant which established a significant mediated moderator path
as hypothesized (Muller et al., 2005). CoM was also tested as a moderator on the path
between perceived friend prevalence and smoking but was not significant. The absence

59

of a significant interaction on the path between perceived friend prevalence and
smoking suggested that the effect of social norms on smoking was similar across CoM
groups. It also suggested that the mediated path between program and smoking
behavior was not contingent on CoM status.

60

Table 11: Mediated Moderator Equations for Wave 2 30-day Smoking with Perceived
Friend prevalence(mediator) and CoM (moderator)

Estimate

SE p-value
Eq1: Overall Effect:
Y=i + c
x
X + e
y

Program (c
x
) -.6906

.3573 .05
Eq2: Mediator
M=i + a
x
X +e
m

Program (a
x
) -.0527

.3057 Ns
Eq3: Indirect Effect and Direct
Effect
Y=i +c’
x
X + b
m
M+e
y

Perceived friend
smoke (b
m
)
.2850

.0413 <.0001
Indirect Effect
(a
x
b
m
)
-.0150

Direct Effect (c’
x
) -.8594 .3763 .022
Eq4:Moderation Effect
Y= i + c
x
X + c
z
Z + c
xz
XZ + e
y

Program x CoM
(c
xz
)
.6972 .3313 .04
Eq5:Total Effect
(calculated from Eq.3)
Total Effect (c’+
a
x
b
m
)
-.8744
Eq6:Moderation Effect on
Mediator
M= i + a
x
X +a
z
Z +a
xz
XZ +e
m

Program (a
x
) .0141 .3065 Ns
CoM (a
z
) 1.3839 .3913 .0004
Program x CoM
(a
xz
)
-1.116 .4825 .02
Eq7:Full Moderation on
Outcome with Mediator
Y= i + C’
x
X +b
z
Z +b
xz
XZ + b
m
M +e
y

Program (c’
x
) -.9830 .3927 .01
CoM (b
z
) .1681 .3829 Ns
Program x CoM
(b
xz
)
.6426 .3526 .07
Friend Prevalence
(b
m
)
.2873 .0416 <.0001
Eq8b: Mediated Moderation
Model controlling for the
mediator
Y= i + C’
x
X +b
z
Z +b
xz
XZ + b
m
M
+b
mz
MZ+ e
y

Program (C’
x
) -.9886 .3932 .01
CoM (b
z
) .2076 .3903 Ns
Program x CoM
(b
xz
)
.6430 .3483 .07
Friend Prev (b
m
) .2928 .0433 <.0001
Friend Prev x CoM
(b
mz
)
-.0643 .1380 ns

61

Table 12: Mediated Moderated Slope Summary
Beta p-value Interpretation
Eq. 4: Overall Moderation effect (c
xz
) .6972 .04
Change in overall treatment
effect on Y as Z increases
Eq. 6: Moderated effect on Mediator
(a
xz
)
-1.116 .02
Change in treatment effect on M
as Z increases
Eq.8b: Mediator on Outcome (b
m
) .2928 <.0001
Mediator effect on Y on average
within the two treatment levels
and at the average level of Z
*All slope parameters must be significant for a mediated moderation effect to exist.

DISCUSSION
This study was the first to examine how depression symptoms affected smoking
prevention program effects with specific regard to social influence processes. We did
not find that the overall program effect on wave 2 30-day smoking varied by level of
depression symptoms as we stated in hypothesis 2.1. However, when we explored the
co-morbidity of depression symptoms and prior smoking experience as a program
moderator, we found it to be significant. This finding was reported previously (Sun et
al., 2007), however, in this context it may provide further information as to how an
individual might perceive information regarding pro-smoking norms. We may speculate
that the program, which aims to change pro-smoking cognitions, provides an alternative
source of social information about smoking to students. Those students who are
sensitive to how they might fit in with their peers, and as we hypothesize this group to
be those with high levels of depression symptoms, may more readily adopt the social
information provided to them by the program. Students with depression- smoking co-

62

morbidity may be more attuned to the corrections presented by the curriculum. Unlike
their depressed counterparts who have never experimented with smoking, they may
have the experience to know that smoking does not necessarily make one popular or
cool or bring more positive attention from peers, thus they would be able to confirm the
messages presented in the curriculum and change their behavior accordingly to avoid
further possible peer rejection.
We explored whether depression moderated the effect the program had in
changing targeted social influence mediators: positive social normative beliefs and/or
perceptions of friend 30-day smoking prevalence. We found that high depression scores
interacted with the program in predicting only perceived prevalence, but the effects
were not strong enough to test the full mediated moderation model. When we
duplicated the model using CoM as the moderator, we found the effect of the program
on wave 2 perceived prevalence to be particularly strong among those with COM
present. This partially supports hypothesis 2.3 and provides evidence for our earlier
speculation that changes in social influences made by the program will be different
depending on students’ depressive status.
The most important and novel contribution to prevention literature is the
mediated moderated findings as hypothesized in hypothesis 2.3. Evidence suggests that
the program changed perceptions of perceived peer prevalence as intended but only for
those with a co-morbidity of high depression symptoms and prior smoking. The
magnitude of the overall treatment effect on smoking behaviors depends on the

63

presence of the CoM, and the mediating process that is responsible for that moderation
lies in the social influence variable being manipulated by the program. In simpler words,
the program was able to change perceptions among those with CoM, and those changes
were responsible for changes in smoking behavior. This finding is especially important
in future curriculum development in which smoking or other socially driven behaviors
are the main outcomes of interest. Gene x environment (social, physical, etc.)
interactions need to be considered when developing hypotheses to test and in designing
interventions since those interactions may drive the success or failure of the program if
the mediators and pathways are not specified well (Johnson et al., 2007).
Depressed adolescents may be easily influenced and conform to the values and
beliefs of those they identify as friends or closest peers (Katkin, Blum Sasmor, & Tan,
1966). If their referent group endorses deviant behaviors then depressed adolescents
may be more likely to engage in those behaviors for the added benefit of (perceived)
acceptance (Besic & Kerr, 2009). These analyses suggest that depressed individuals can
also be influenced by prevention program messages that manipulated perceptions of
smoking prevalence. Studies that assess conformity and decision-making under peer
influence among depressed adolescents will contribute greatly to understanding the
social and cognitive mechanisms involved in behavior choice.
Limitations and Future Directions
There are several limitations to the findings of this study that warrant discussion.
Despite randomized assignment at the school level, the conditions were unbalanced.

64

Initial analyses controlled for the unbalanced conditions by using a mixed model and
indicated school intercepts as a random effect. However, models would not converge.
Thus, we reverted to logistic and linear regression models and controlled school as a
covariate. By doing so, we run the risk of a Type I error so the results presented here
should be interpreted with great caution. Future studies may need to oversample for
high risk adolescence to test specific hypothesized pathways and great care should be
taken to confirm balanced randomization to treatment conditions.
Differential attrition was also present in this study. Although we controlled for
attrition by producing a propensity score and treating it as covariate in all our modeling,
we made assumptions as to which variables may have contributed to drop out. There
may be another unmeasured factor that would bias drop-out rate and thereby biasing
our study results thus, caution is necessary in interpreting and generalizing these study
findings. Future studies will need to focus on reducing drop-out or where warranted,
imputation should be conducted.
The results indicated that perceptions surrounding friend use prevalence were
changed among the high risk co-morbid group. What we could not test in this study was
whether their perceptions of their friends use changed or whether they changed who
they selected as friends. Overestimations of peer smoking behaviors have been shown
to be a significant predictor of adolescent smoking and contributed to the overall norm
that smoking was acceptable by peers (Graham et al., 1991; Gunther, Bolt, Borzekowski,
Liebhart, & Dillard, 2006; Lai, Ho, & Lam, 2004; Reid, Manske, & Leatherdale, 2008).

65

Social influences smoking prevention programs aimed to change these overestimations
and perceptions surrounding peer acceptance of smoking. Whether the WSPT program
changed the depressed students’ estimations or if it worked in a way to encourage
depressed students to choose friends who did not smoke is left to be determined by
future studies. Given the support in the literature which suggests depressed
adolescents have a difficult time making friends and are often at the periphery of the
social group, we are confident that the WSPT program changed cognitions rather than
being able to affect change in friendships within a year. (Aseltine et al., 1994; Bosacki,
Dane, & Marini, 2007; Henrich, Blatt, Kuperminc, Zohar, & Leadbeater, 2001;
Oldehinkel, Rosmalen, Veenstra, Dijkstra, & Ormel, 2007; Pedersen, Vitaro, Barker, &
Borge, 2007). Furthermore, both interpersonal and cognitive theories of depression
suggest that difficulty with peer interactions, whether they are accurate with their
perceptions or if they are biased in their perceptions, would agree that changing
friendship groups would be unlikely (Joiner & Coyne, 1999; Lakdawalla et al., 2007;
Marcus & Nardone, 1992).
Finally, these analyses were conducted only among males in mainland China,
thus these results may not be generalizable to any other population. Future studies will
need to test the cross-cultural validity of the proposed pathways being tested here. It
may be that Chinese adolescents are more adept at recognizing social cues since their
collectivistic culture encourages social harmony (Bond & Smith, 1996; Triandis, 1995).
Furthermore, depression in China may manifest differently or may have a different

66

meaning from other western populations (Parker, Gladstone, & Chee, 2001; Ryder et al.,
2008; Song et al., 2008). Future cross-cultural studies will need to assess differences in
depressive symptomatology and social influences.
This study is the first to test pathways and cognitive processes of a social-
influences based prevention program for high-risk youth. Yet, future studies need to
examine whether adolescents at risk for depression perceive their social environment
differently or if they interpret their environment differently. For example, using
smoking as an examples, future studies should look at whether a depressed adolescent
over-estimates their friends’ smoking behaviors or if they simply have more friends who
smoke. Studies can also assess whether over-estimation (i.e. cognition) or direct peer
influence (i.e. actual social environment providing opportunities to smoke) is
responsible for heightened risk for smoking often found among depressed adolescents.
These are all interesting hypotheses that will give great insight to the specific
mechanisms to target when developing future prevention programs.

67

CHAPTER 4: Exploring Social Competence, Depression Symptoms, and Smoking Related
Social Influence Factors

ABSTRACT
The role of the social environment appears to be particularly important in
determining how depression symptoms are associated with smoking during
adolescence. Sociability, defined as the ability and preference for social interaction, and
social competence deficits are also important in understanding the social environment.
The current study investigated the relationship between social competence, depression
symptoms, and psychosocial smoking risk factors among 10
th
grade Chinese adolescents.
Results show that depression symptoms were negatively associated with refusal skill
efficacy. Sociability was positively associated with refusal skill efficacy, after controlling
for depression, age, class, and smoking status. Adolescents who were more sociable
were more confident in refusing cigarette offers while depressed adolescents were less
confident. Adolescents who experienced depression symptoms in the past month
perceived higher friend use prevalence than their non-depressed counterparts. Social
norm perceptions were the same for depressed and non-depressed adolescents but
were lowest among those with both depression symptoms and social deficits. Social
competence may need to be studied together with depression to truly understand the
cognitive processes involved in perceiving modifiable social influence factors.

68

INTRODUCTION
. Interpersonal theories of depression generally posit that poor social interactions
contribute to the development and persistence of depressive symptoms (Joiner &
Coyne, 1999; Lara & Klein, 1999). Peer rejection during adolescence may be an
especially potent event that contributes to depressive symptoms (Lev-Wiesel, Nuttman-
Shwartz, & Sternberg, 2006). Rejection by same sex peers have been associated with
increases in depressive symptoms (Brendgen, Wanner, Morin, & Vitaro, 2005) and poor
family and peer relationships have also been prospectively linked to increased
depression risk (Allen et al., 2006; Allen, Porter, McFarland, McElhaney, & Marsh, 2007;
Gutman & Sameroff, 2004). Inherent social competence abilities may be at the root of
these poor social relationships.
Social competence may be related to depression development in adolescents
(Lewinsohn, 1974; Lewinsohn & Gotlib, 1995). Social competence is necessary for an
individual to perceive social cues accurately in their environment and to choose
appropriate responses while social deficits have been associated with the experience of
negative affect (Marcus & Nardone, 1992; Reijntjes, Stegge, Terwogt, Kamphuis, &
Telch, 2006), peer rejection (Reijntjes, Stegge, & Terwogt, 2006), and ultimately
depression (Uhrlass, Schofield, Coles, & Gibb, 2009). Negative beliefs of oneself and of
others were found to be associated with dysfunctional social behavior and lower
positive peer status (Rudolph, Hammen, & Burge, 1995, 1997). Those at risk for
depression had fewer friends, were less popular, and had lower quality friendships than

69

their lower risk counterparts (Field, Diego, & Sanders, 2001; Kistner, 2006; Kistner,
David-Ferdon, Repper, & Joiner, 2006). Other studies suggest that the relationship
between social competence and depression is mediated by an individual’s perception of
connectedness with others (Williams & Galliher, 2006) or lack of social support (Rockhill,
Vander Stoep, McCauley, & Katon, 2009; Whitton, Larson, & Hauser, 2008). Thus, it
appears that social acceptance and belonging may be important for individuals at risk
for depression. If social acceptance and belonging is being sought by depressed
adolescents, then motivation to comply with perceived peer norms for risk behaviors,
such as smoking, would be especially elevated among this group.
Adolescents who are at the periphery of social networks appear to be highly
influenced by those central to their network (Pearson & Michell, 2000; Pearson et al.,
2006) and may be more affected by perceptions of their friends’ smoking and
perceptions of overall peer smoking prevalence (Aloise-Young, Graham, & Hansen,
1994). Depression symptoms may increase their perceptions of smoking related norms
and behaviors by inaccurately estimating behaviors or inaccurately attributing
motivation behind cigarette offers. Alternatively, social competence may be responsible
for inaccurate perceptions of the social environment while depression may simply be
the consequence of the perception errors.
This study aimed to explore social competence and depression on the perceptions of
smoking related social influences. Social competence may affect how an individual
perceives social rules and normative expectations of those around them. Thus, social

70

competence may lead to better friendship quality and popularity (that is, sociability) and
may be associated with pro-smoking social influences. Social deficits may not be
associated at all with pro-smoking social norm beliefs because perceptions of any norms
would be lacking. Those with depression symptoms may be biased toward higher
perceptions of peer use and acceptance because of the positive qualities often
associated with smoking by adolescents. Where social competence relies on accuracy of
perceptions to influence behavior, those with depression symptoms may be biased to
perceive social information that might signal acceptance or rejection by peers.
Investigating how social competence and depression affects perceptions of social
information will benefit future prevention intervention design by detailing divergent
pathways to target. Furthermore, examining how social competence and depression
affects refusal skill efficacy, or an individual’s perception of their own ability to refuse a
cigarette offer, will provide insight on the cognitive processes involved in evaluating
their own social abilities. The current study investigated the roles of social competence
and depression on perceptions of smoking related social norms, perceived friend use
prevalence, and refusal skill efficacy.
Specific Aims and Hypotheses
The aims are to
3.1 assess the relationship between smoking, depression, social competence
(sociability and social deficit), and smoking specific social influences.

71

Hyp. 3.1.a It was hypothesized that depression and social deficits would be
positively correlated and depression and sociability would be
negatively correlated.
Hyp. 3.1.b It was hypothesized that depression and social norms would be
positively correlated, depression and perceived friend prevalence
would be positively correlated, and depression and friend refusal
would be negatively correlated;
Hyp. 3.1.c It was hypothesized that sociability and social norms would be
positively correlated, sociability and perceived friend prevalence
would not be significantly correlated, and sociability and friend
refusal would be positively correlated (greater sociability would
be associated with greater refusal ability). Social deficits and
social norms were expected to be negatively correlated, social
deficits and perceived friend prevalence were expected to be
negatively correlated, and social deficits and friend refusal skill
were expected to be negatively correlated.
3.2 investigate how social competence would impact the relationship between
depression and pro-smoking social influence indicators.
Hyp.3.2.a It was hypothesized that depression would be positively
associated with sociability (but not social deficits) and depression
would be negatively associated with perceived social norms.

72

Hyp.3.2.b It was hypothesized that depression and social deficits (and not
sociability) would be negatively associated with perceived friend
refusal skills.
Hyp.3.2.c It was hypothesized that depression but not sociability or social
deficits would be significantly associated with perceived friend
prevalence.
3.3 investigate the effect of depression on social influences across levels of social
competence.
Hyp. 3.3.a It was hypothesized that social deficits would moderate the
relationship between depression and social normative beliefs,
after controlling for smoking, age and class.
Hyp 3.3.b It was hypothesized that social deficits would moderate the
relationship between depression and refusal skills such that those
with higher deficits would have a stronger negative relationship
between depression and refusal skills.
METHODS
Sample Selection
The Chengdu Pilot Study was a paper based survey conducted during the Spring
of 2006 with no intervention implemented. Chengdu is a dense urban city and is the
capital of Sichuan Province in southwest China with a population over 10 million. The
purpose of the pilot was to assess the cognitions of 10
th
grade Chinese adolescents in a

73

range of psychosocial, personality, and tobacco/alcohol use constructs. Four classes,
two from an academic school and two from a vocational school, were selected through
a convenience sample by the Chengdu Center for Disease Control and Prevention.
Academic schools are college preparatory schools, while vocational schools prepare
students for a specific trade. Schools are also ranked in terms of academic status and
rigor within school type. To increase variability of the sample, one high academic status
and one moderate/low academic status school was selected within each school type.
One 10
th
grade class from each school was randomly selected to participate in the pilot
study. A total of 234 students were consented and surveyed. The entire study was
approved by the Institutional Review Boards of the University of Southern California and
the Chengdu Center for Disease Control. All participants had active parental consent
and student assent.
Measures
Social Competence. Social Competence is measured using inventories derived from
the Autism Spectrum Quotient (AQ; Baron-Cohen, Wheelwright, Skinner, Martin &
Clubley, 2001). The AQ is a 50-item self-report inventory which assesses 5 components
of autism: Sociability, Communication, Attention Switching, Imagination, and Attention
to Detail. The AQ was translated into Chinese and was tested in Fall 2005 with 529
young adult workers (mean age=28.1 years, SD =7.34) in Chengdu, China (Gallaher,
unpublished manuscript). Two interpretable factors emerged which were named

74

Sociability and Social Deficits. Correlations between the shortened scales and the full
AQ were .77 for Sociability and .88 for Social Deficits.
Sociability was assessed by asking students whether they agreed with the following
statements: “I enjoy social chit-chat”; “I find it easy to ‘read between the lines’ when
someone is talking to me”; “I know how to tell if someone listening to me is getting
bored”; and “I am good at social chit-chat.” Response options ranged from (1) Definitely
agree to (4) Definitely disagree. Items were reverse coded so that higher scores
reflected higher skills and the mean was taken of the items. The standardized
Cronbach’s alpha was .65 which represented acceptable internal consistency.
Social Deficits (AQ-Def) were assessed by asking students whether they agreed with
the following statements: “Other people frequently tell me that what I’ve said is
impolite, even though I think it is polite”; “I frequently find that I don’t know how to
keep a conversation going”; “When I talk on the phone, I’m not sure when it’s my turn
to speak”; “I am often the last to understand the point of a joke”; and “People often tell
me that I keep going on and on about the same thing.” Response options ranged from
(1) Definitely agree to (4) Definitely, disagree. Items were reverse coded so that higher
scores reflected higher social deficits and the mean was taken of the items. The
standardized Cronbach’s alpha coefficient was acceptable for Social Deficits a (α=.63).
Sociability and social deficits appeared to be orthogonal constructs in the overall
sample but were found to be significantly correlated among the female sample. Thus,
we chose to keep the sociability and social deficit constructs separate to better assess

75

the independent relationships in the social influence models. Furthermore, the face
validity of the scales appeared different from each other. The sociability scale appeared
to measure preference for social interactions while the social deficits scale appeared to
measure inability to recognize appropriate behaviors or read social cues.
Depression. Depression was assessed using three separate items: “Have you ever in
your whole life had a period of 2 weeks or more when you felt down, depressed, or
hopeless OR had little interest or pleasure in doing things?”; “During the past month,
have you often been bothered by feeling down, depressed, or hopeless?”; and “During
the past month, have you often been bothered by little interest or pleasure in doing
things?” Students were to answer (1) Yes or (2) No to each of the questions. These
questions were shown to have 96% sensitivity and 57% specificity (Whooley, Avins,
Miranda, & Browner, 1997). Questions were treated as individual items for exploratory
analyses. The item that assessed past month feelings of depression or hopelessness had
high correlations with the other two depression items (r-square= .46 to .55) and
correlated well with other constructs of interest. Thus, this item was used to represent
the depression symptom construct.
Social Norms (normative beliefs about smoking). Social norms were assessed by
asking students what they believe or expect to happen as a result of smoking cigarettes.
The following statements were provided: “Smoking helps a person look better or more
attractive”; “People look up to those who smoke”; “Smoking is important to make or
keep friends”; “Smoking makes a person more friendly or outgoing”; “Smoking makes

76

people look tough or cool”; “Most popular people smoke cigarettes”; “Smoking makes a
person feel more comfortable around others.” Response choices ranged from (1) Never
to (5) Always. Items were summed and the mean was taken so that higher scores
reflected positive perceptions of smoking.
Friend Prevalence. Perceived friend use prevalence was assessed by asking students
“how many of your good friends smoke cigarettes at least once a month?” Response
items ranged from (1) none to (4) all. Higher scores indicate higher perceived friend
use.
Friend Refusal Skills. Friend refusal skill was measured with one item which asked
students “If one of your best friends offered you a cigarette, would you smoke it?”
Response items for this question ranged from (1) Definitely, yes to (4) Definitely not. A
higher score would indicate higher refusal skill ability.
Smoking. Lifetime smoking was assessed by asking “have you ever tried cigarette
smoking, even a few puffs?” Answer choices were (1) No and (2) Yes. Past 30-day
smoking was assessed by asking “During the past 30 days, on the days you smoked, how
many cigarettes did you smoke per day?” Response choices were: (1) I did not smoke
cigarettes during the past 30 days; (2) Less than 1 cigarette per day; (3) 1 cigarette per
day; (4) 2 to 5 cigarettes per day; (5) 6 to 10 cigarettes per day; (6) 11 to 20 cigarettes
per day; (7) more than 20 cigarettes per day. Because the data was skewed to the right,
30-day prevalence was assessed using a dichotomous variable where (1) Smoked at
least 1 cigarette per day were considered smokers and (0) those who smoked less than 1

77

cigarette per day or did not smoke cigarettes during the past 30 days were considered
non-smokers.
Data Analysis
The data were analyzed for descriptive statistics among demographic, social
influences, depression, and smoking variables (see Table 12). Bivariate correlations
were also examined for relationships between demographic, predictor and outcome
variables. The data were stratified by gender and general linear models were separately
run with social norm beliefs, refusal friend efficacy, and perceived prevalence as
dependant variables of interest. Age, class, and smoking status were controlled in each
model. Depression symptoms were run as an independent variable predicting
(separately) social norms, refusal skill efficacy, and perceived prevalence. Since smoking
status was determined by past 30-day use, the single indicator for past 30-day
depression symptoms was used for all regression analyses. It also appeared to have
high correlations with the other depression questions indicating a good representation
of symptoms as well more stable of a measure than ever depressed. Models were run
with sociability predicting (separately) social norms, refusal skill efficacy, and perceived
prevalence and repeated again but with social deficit as the independent variable.
Next, with depressive symptoms in the model as the independent variable, sociability or
social deficit was added to the model to assess the change in the variance explained and
the change in the beta estimate of depression symptoms. Finally, depression x

78

sociability or depression x social deficit was also added to the models to assess any
moderation effect on these social influence outcomes.
RESULTS
Sample Characteristics
The sample characteristics were reported in Table 13. A total of 234 students
participated in the survey (50% female, mean age=16.2 years). There were significant
Table 13. Sample Characteristics
All (n=234) Male (n=107) Female (n=117) t-test p-value
Mean SD Mean SD Mean SD
Age 16.2 0.58 16.33 0.61 16.09 0.52 -3.2 0.0016
Grades 2.55 1.06 2.54 1.06 2.59 1.07 0.32 ns
Allowance 5.37 2.35 5.07 2.39 5.65 2.30 1.83 .07

N % n % n % Chisq P-value
School type
Academic 114 50.89 54 24.11 60 26.79 0.0148 ns
Vocational 110 49.11 53 23.66 57 25.45
Smoking
Ever Smoke 92 41 54 24.1 38 17 7.47 0.0068
30-day Smoke 25 11.4 21 9.55 4 1.82 14.49 <.0001
Depression
Ever Depressed 99 44.59 46 20.72 53 23.89 0.1179 ns
30-day depressed 66 29.6 29 13 37 16.59 0.6138 ns
30-day low interest 89 39.91 42 18.83 47 21.08 0.0371 ns

Mean SD Mean SD Mean SD t-test p-value
Social Influence
Refusal Skill 3.46 .890 3.31 1 3.60 .76 2.46 .01
Social Norm 1.50 .628 1.63 .71 1.39 .524 -2.87 .004
Friend Prevalence 1.69 .6677 1.84 .76 1.56 .55 -3.08 .002
Social Competence
Sociability 2.94 .567 2.92 .602 2.95 .529 .41 Ns
Social Deficits 2.02 .533 2.07 .534 1.97 .481 -1.29 Ns

79

gender differences among age and smoking behaviors with males being older and
having significantly higher smoking rates. There were no significant gender differences
in depression symptoms as expected and may be attributable to single questions not
providing enough sensitivity to detect levels of depression. Smoking related social
norms, perceived prevalence and refusal skills were significantly different for males and
females but sociability and social deficits were not.
Simple Correlations
Simple correlations showed sociability and social deficits to be separate but
correlated constructs. Social deficits were significantly correlated with all three
indicators for depression (Table 14). Correlations between depression and perceived
social influence indicators were also significant while correlations between depression
and smoking were not. The sample was stratified by gender and the correlations by
gender are presented in Table 15. The correlation patterns appeared to be different for
females and males. For females, the social norms factor was positively correlated with
social competence (both sociability and social deficit) and depression (ever and past
month depression). No other clear pattern between smoking perceptions, social
competence, or depression was present in this female sample. For males, the social
norms factor was only correlated with past month depression and smoking related
factors but not sociability or social deficits. It appeared that smoking behaviors and
perceptions were all significantly correlated, perceptions were only correlated with past

80

month depression, and depression (all three indicators) were significantly correlated
with social deficits.
Depression Only Models
Past month depression, past 30-day smoking, and control variables (age and
class) were regressed on Social Norm Beliefs (see Table 16). The study sample was
stratified by gender and models were run separately. Model fit statistics indicated that
the specified model was not adequate for the female sample (F-value=1.72, p-value
>.15, r
2
=.09) but was acceptable for the male sample (F-value=4.23, p-value=.0008,
r
2
=.21). For males, past month depression was not significantly associated with
perceptions of social norms (beta=.236, p>.15).
Past month depression, past 30-day smoking, and control variables (age and
class) were regressed on perceived refusal skills (see Table 16). The study sample was
stratified by gender and models were run separately. Model fit statistics indicated that
the specified model was not adequate for the female sample (F-value=1.06, p>.15,
r
2
=.06) but was acceptable for the male sample (F-value=5.94, p<.0001, r
2
=26). For
males, past month depression was moderately associated with refusal skills (beta=-.328,
p=.10) suggesting that an individual who experienced depression symptoms in the past
month were less likely to refuse a cigarette from a friend.
Past month depression, past 30-day smoking, and control variables (age and
class) were regressed on perceived friend prevalence (see Table 16). The study sample
was stratified by gender and models were run separately. Model fit statistics indicated

81

that the specified model was not adequate for the female sample (F-value=1.78, p>.15,
r
2
=.09) but was acceptable for the male sample (F-value=6.89, p<.0001, r
2
=.30). For
males, past month depression was moderately associated with perceived friend use
prevalence (beta=.285, p=.06) indicating that an individual who experienced depression
symptoms in the past month were more likely to perceive greater smoking prevalence
rates among their friends, even after controlling for current smoking, age, and class.
Sociability Only Models
Sociability, past 30-day smoking, and control variables (age and class) were
regressed on social norms (see Table 17). The study sample was stratified by gender
and models were run separately. Model fit statistics indicated that the specified model
was not adequate for the female sample (F-value=1.77, p>.15, r
2
=.09). Although the
model fit was acceptable for the male sample (F-value=3.8, p=.002, r
2
=.2), sociability
was not significantly associated with social norm beliefs (beta=.0488, p>.15).
Sociability, past 30-day smoking, and control variables (age and class) were
regressed on friend refusal skills (see Table 17). Model fit statistics indicated that the
specified model was not adequate for the female sample (F-value=.68,p>.15, r
2
=.04) but
was acceptable for the male sample (F-value=6.71, p<.0001, r
2
=.29). For males,
sociability was significantly associated with refusal skills (beta=-.368, p=.01). The inverse
relationship suggested that those with better sociability would find it more difficult to
refuse a cigarette from a friend.

82

Sociability, past 30-day smoking, and control variables (age and class) were
regressed on perceived friend use prevalence (see Table 17). The model fit was not
significant for females and despite good model fit for males, sociability was not
associated with perceived friend use.
Social Deficit Only Models
Social deficits were not found to be significantly associated with social norms,
refusal skills, or perceive friend prevalence for females or males (see Table 17).
Sociability and Depression Models
All sociability and depression models were conducted among the male
subsample only due to the lack of appropriate model fit among females. Sociability and
past month depression were regressed on social norms. After controlling for 30-day
smoking, age, and class, sociability and depression were not significantly associated with
social norms.
Past month depression was moderately associated with refusal skill efficacy
(beta=-.328, p=.10). When sociability was added to the model (std. beta=.382, p=.009),
an additional 3% of the variance for refusal skill was explained and adjusted the
depression estimate (std. beta=-.1597, p=.07) by 51%. However, the sociability beta
estimate changed direction from the sociability only model (std. beta=-.368, p=.01) to
the sociability and depression model (std. beta=.382, p=.009). This suggested that
depression may have been a distorter variable. A distorter variable is one that changes
the relationship between two variables by either accentuating the relationship or

83

changing the direction altogether (MacKinnon, 2008). It was also possible that the
relationship between sociability and refusal skill efficacy was confounded by depression
symptoms however, simple correlations do not show a relationship between depression
and sociability among males. A confounder, by definition, would have to have a
relationship with both the independent and dependant variable, which was not the case
here. Thus, the true relationship may be interpreted as greater sociability is positively
associated with greater refusal skill, after controlling for depression, smoking, age and
class.
Sociability and past month depression were both entered in the same model and
regressed on perceived friend prevalence; neither was significant. Sociability x
depression interaction was also explored but was not found to be significant.
Social Deficit and Depression Models
Correlations between social deficits and depression were significant among
males thus, the relationship was explored further. Males who reported depression
symptoms in the past month had higher mean scores of social deficits than their non-
depressed counterparts (p=.02). After controlling for age and class, social deficits were
significantly associated with past month depression (beta=.7895, p=.05), suggesting that
deficits and depression may be inter-related. Social deficits may interfere with how an
individual interprets social interactions and if combined with depression symptoms,
produce vulnerability for pro-smoking messages. Social deficit x depression interaction
term was included in the social norm regression model. The interaction term was

84

significant (std beta=-.131, p=.03) which indicated that the relationship between
depression and social norms were strongest for those with the greatest social deficits
(Figure 3). The negative association suggested that those who experienced depression
in the past month and had the highest social deficit scores had lower perceived pro-
smoking social norm beliefs. For those with the lowest social deficit scores, perceived
pro-smoking social norms beliefs did not differ for those who experienced depression in
the past month compared to those who had not.
Social deficits and past month depression were regressed on perceived refusal
skills in the same model. Both social deficits and depression were not significantly
associated with refusal skills. The social deficit x depression interaction term was also
not significant. Similarly, social deficit and depression were not significantly associated
with perceived friend prevalence and the interaction term social deficit x depression
was also not significant.

85

86

87

88

89

DISCUSSION
This exploratory study provided insight on the relationship between social
competencies, depression, and perceived smoking related social influences. There
appeared to be a complex relationship between social competence and depression for
each of the smoking related social influence indicators. Social competence, as
measured by two separate scales for sociability and social deficits, proved to be
especially complex. Sociability appeared to measure how much an individual enjoyed
social interaction and whether they felt they were competent in these general social
exchanges. Social deficits were not the opposite of sociability. Rather, social deficits
appeared to measure the inability to recognize social cues or recognize when they
would make social mistakes. The distinction between sociability and social deficits

90

provided further detail on the pathways involved in perceptions of social influences.
This was the first study, to our knowledge, that explored sociability, social deficits, and
depression with smoking related perceptions.
Depression symptomatology was significantly associated with perceived friend
smoking prevalence, even after controlling for smoking behavior. This finding suggests
that adolescents who are at risk for depression may be biased in overestimating their
friends’ smoking behaviors. These analyses replicated the conclusions from Studies 1
and 2 but in an older cohort of male adolescents in central China. Furthermore, the
relationship between depression and greater perceptions of peer approval for smoking
was also found among multi-ethnic adolescents in the United States (Ritt-Olson et al.,
2005). We speculate that this consistent finding may be associated with depressed
adolescents having an increased sensitivity to proximal peer influence. In other words,
depressed adolescents may pay greater attention to their closest friends and are more
motivated to fit in with them rather than acting on more generalized perceptions of
smokers (i.e. social norms). These results may alternatively suggest that depressed
adolescents are selecting friends who smoke. Although disentangling the perception
versus selection of smoking friends is beyond the scope of this study, this study does
suggest that depressed adolescents may be especially influenced by their friends’
smoking beliefs and/or behaviors.
Sociability appeared to only have a significant inverse relationship with refusal
skill efficacy among males. At first, this was interpreted to mean that the more sociable

91

or socially skilled an individual the more difficult they found refusing a cigarette offer
from a friend. Another way to interpret this would be to consider the social
environment of an average male adolescent in China. It may be that the most socially
acceptable behavior would be to accept a cigarette when offered by a friend. However,
when the relationship between sociability and refusal skills was examined in the context
of depressive symptoms, depression symptoms appeared to distort the true
relationship. After controlling for depressive symptoms, sociability was positively
associated with refusal skills suggesting that the more sociable an individual the more
likely they would refuse a cigarette offer from a friend. The sociability x depression
interaction term was not significant, implying that the relationship between sociability
and refusal skill efficacy did no differ across level of depression symptom. Furthermore,
the direct effects between depression and refusal skill and between sociability and
refusal skill remained significant indicating that they were independently associated
with refusal skill efficacy. Adolescents who are more sociable may have the skill sets to
effectively refuse a cigarette from a friend or they may feel more confident in their
ability to do so, regardless of whether they choose to refuse an offer or not. Thus,
implications of this finding suggested that smoking prevention programs aimed
primarily at strengthening refusal skills and communication skills may need to address
more than confidence in abilities (Graham, Marks, & Hansen, 1991; Hansen & Graham,
1991). Future programs also need to identify motivating factors that contribute to
acceptance or refusal of cigarette offers. For example, increasing anti-smoking

92

normative information, or more specifically, increasing the belief that smoking will not
garner acceptance by peers, may be a more effective message by giving individuals a
reason to refuse cigarette offers.
One of the most interesting contributions of this study is the interaction term
deficits x depression as it relates to pro-smoking social normative beliefs. It appeared
that the relationship between depression and social norms was stronger among those
with higher social deficits. Although the relationship was stronger in magnitude the
relationship was paradoxically negative. In other words, those with high social deficits
and high depression symptoms perceived less pro-smoking social norms. The
normative messages portrayed by cigarette companies, media, celebrities, and peer
culture is that cigarettes make a person appear cool, mature, appealing (Chen et al.,
2006; Gunther et al., 2006; Kobus, 2003). These are subtle normative influences that
may require a certain level of social competence to recognize the positive smoking
attributes and ascribe to them. A socially competent adolescent would realize what a
cigarette offer might mean in terms of positive social interaction rather than simply
whether or not they want a cigarette (Charlton, Minagawa, & While, 1999).
Adolescents with high levels of social deficits and high depression symptoms may
be a unique group in which depression symptoms are consequences of repeated poor
social interactions. This supports interpersonal theories of depression (Joiner & Coyne,
1999; Joiner, Metalsky, Lew, & Klocek, 1999; Kistner, 2006; Sacco & Vaughan, 2006).
The measure used to assess social deficits was adapted from a clinical assessment for

93

autism and this study may be detecting those with clinically significant indicators for
true social deficits. Depression and social competence are multi-faceted issues in which
no single definition exists for these dispositions. These analyses may have identified a
specific type of adolescent with interpersonal deficits although their deficit may not
result directly in risk for smoking through social norm perceptions.
This study was exploratory in nature so several limitations warrant discussion.
The study utilized a small convenience sample which limits the findings from being
generalized to other populations. Furthermore, no causal inferences can be made
based on the cross-sectional data presented in this study. Depression and social
competence have a complex relationship that cannot be easily examined without
longitudinal data. Adolescence is a time of rapid social, emotional, psychological, and
physical growth and the interactions between these domains will undoubtedly affect
how personality and peer relationships interact to produce risk behavior. Although we
specified our models based on theory and built our hypotheses on prior studies, future
studies will need to examine these relationships in detail across this developmental
period. Furthermore, the current study aimed to understand the effects of depression
and social competence on psychosocial risk factors for smoking, specifically perceptions
of social norms, perceived friend prevalence, and refusal skill efficacy. Patterns for
smoking risk through these social influence factors must be examined in future studies
to establish meditational pathways.

94

The measure for depression symptoms may not have been sensitive enough to
detect a true risk for developing depression so the findings of this study may
characterize a group of students who may have experienced a hardship within a month
of survey administration. Due to a potential unmeasured, non-random event, these
findings may not be valid for determining depressive risk and the associations with our
social influence variables. Future studies should use a more complete scale such as the
Centers for Epidemiologic Studies-Depression scale (CES-D; Field, Diego, & Sanders,
2001; Radloff, 1977; Song et al., 2008) and measured over several time periods.
Finally, a priory hypotheses need to be developed and applied to longitudinal
data based on the exploratory results presented in this study. Future directions should
include a careful study of the relationship between social competence and depression.
Since these analyses are cross-sectional, the relationships between social competence,
depression, and smoking related norms cannot be temporally inferred. Future studies
should also consider experimental manipulation in order to detect accurate cognitive
processes for depressed, socially incompetent, and normal adolescents. Understanding
these processes will give interventionists areas in which to target for manipulation and
may potentially provide new therapies for those at risk for depression as well as
smoking.

95

CHAPTER 5
CONCLUSIONS
This dissertation investigated the role of depression symptoms on social
information processing and smoking among male adolescents in central China. It was
hypothesized that depressed adolescents process social information in a way that may
manifest as sensitivity to social influences. The three studies presented evidence to
support how depression is indirectly related to smoking behavior through social
influences, how manipulating social perceptions would be most effective in reducing
smoking risk among depressed adolescent subgroups, and how depression and social
competence interact in perceiving the social environment.
Study 1 investigated how perceptions of social influences, depression symptoms,
and concurrent past month smoking were related. Results supported that perception of
pro-smoking social norms and perceived friend use partially mediated the effect of
depression on past month smoking behavior. Those who had higher depression scores
also had higher perceptions of smoking related social influences suggesting that those
with depression symptoms may have biased perceptions or biased processing of social
information. This indirect pathway between depression and smoking was also found in
a similar study conducted with 9
th
graders in the western United States. Perceived
friend approval of smoking partially mediated between depression symptoms and
smoking behavior (Ritt-Olson et al., 2005). These studies were conducted with
adolescents from two cultures (China and United States) and used different social

96

influence indicators, but the similarity in results collectively support that depressed
adolescents are sensitive to social influences and may choose smoking behaviors based
on their perceptions.
Study 2 examined whether changes in perceived social influences through a
social influences based smoking prevention program resulted in greater changes in
smoking behavior among depressed adolescents. Results showed that the program was
effective in changing perceptions of friend use prevalence but only among those with a
co-morbidity of high depressive symptoms and smoking at baseline. Co-morbidity did
not moderate the relationship between perceptions and smoking behaviors one-year
after program implementation. Taken together, the moderation results suggested that
program messages were particularly salient to the co-morbid group in changing
perceptions of friend use but overall risk factor of perceived friend prevalence similarly
predicted smoking behaviors between the co-morbid and non-comorbid groups. Study
2 partially supported the hypothesis that depressed adolescents perceived their
proximal social environment differently than their non-depressed counterparts and this
difference in perception, resulted in different smoking behaviors.
Although the survey instrument assessed perception, we cannot preclude other
interpretations. Depressed adolescents may be selecting friends who smoke because of
their smoking status or other attributes that are often associated with smoking, like
popularity or attractiveness (Arnett, 2007; Killeya-Jones, Nakajima, & Costanzo, 2007;
Wang, Eddy, & Fitzhugh, 2000). These individuals who smoke in early adolescence may

97

also be influential to the larger peer group, central to the peer network, and maintain
elevated peer status regardless of smoking status in late adolescence (Killeya-Jones et
al., 2007). It is also possible that depressed adolescents may be identifying individuals
they consider to be friends but who do not mutually consider the depressed adolescent
a friend. Thus the positive attributes and pro-smoking norms set by the popular
individual may continue to be influential to the depressed adolescent regardless of true
friendship reciprocity.
Conflicting studies show that smokers were not central to peer networks but
were actually on the periphery of these networks. One study using social network
analysis found that most adolescent smokers were those who did not belong to a peer
group (i.e. isolates; (Ennett & Bauman, 1994). Authors eliminated these isolates from
peer influence analyses because they stipulated that influence can only happen directly
between group members. Nonsmokers in smoking networks were more likely to smoke
a year later compared to those nonsmokers in nonsmoking networks which suggested
that smoking behaviors could be attributed to peer influence. Smokers were less likely
to change social position than nonsmokers and authors interpreted this finding to
suggest that a selection process existed which favored nonsmoking over smoking
behaviors, thus smoking behaviors could be attributed to peer selection for this group of
adolescents. Finally, nonsmokers were dropped from or chose to leave smoking peer
groups more than nonsmokers in nonsmoking peer groups which represented
deselection from smoking peer networks. Together, the results suggested that both

98

peer influence and selection contributed similarly to smoking behavior, but selection
may be stronger for nonsmokers because of the fluidity of the peer network structure.
Similarly, another study with adolescents showed that smokers were greater among
dyads (reciprocal friendships) and isolates but lower among integrated peer networks
(Pearson et al., 2006). It is possible that peer networks can also reject individuals from
network inclusion, thereby creating isolates. Smoking could be the result of peer
rejection and isolates may choose to smoke to cope with negative status or possibly to
gain attention. Thus, depressed adolescents could be isolated from the greater social
network, either through self-selecting out of social integration or through peer
rejection, and be exposed to more deviant behaviors conducted by other isolates, such
as smoking.
Finally, another social network analysis study found that liaisons, or those
individuals who would link between two or more peer groups, were more likely to be
smokers than either peer group members or isolates (Henry & Kobus, 2007). Authors
argue that these isolates are exposed to more smoking peers and are therefore more
likely to smoke than others. However, peer influence and selection may be more
complex in predicting smoking behavior than network status (Mercken, Candel, Willems,
& de Vries, 2007). In early adolescence, friendship selection may be a stronger predictor
of smoking initiation but overtime, influence becomes a stronger predictor during mid
adolescence (Mercken et al., 2009).

99

The results are mixed whether peer influence or selection is responsible for
smoking behaviors in adolescence. Peer influence and peer selection are intricately
linked to perceptions of smoking norms and smoking prevalence rates. Thus, no
conclusions can be made regarding whether depressed adolescents perceive social
information differently or if they are selecting friends who smoke. Study 2 of this
dissertation utilized a social influences based prevention program to correct
overestimations of peer use, build a social environment that conveys peer disapproval
of smoking, and build social skills to resist peer influence. Depressed adolescents who
had smoked prior to receiving the program either perceived less friends who smoke or
selected friends who did not smoke one year after implementation. Although we have
interpreted this finding as a change in perception, the social skills taught in the program
may have provided the skills needed for these adolescents to refuse smoking and or
choose a new group of friends. Furthermore, the Wuhan Smoking Prevention Trial
program also used novel group activities that may have increased exposure to other
students and development of new relationships. Future studies will need to examine in
detail whether depression is associated with biased perceptions of social cues related to
rejection or if the program was successful in changing friendships. Future studies will
also be able to assess these hypotheses using social network data. No study to our
knowledge has assessed depression symptoms and smoking behavior using social
network analysis techniques. Such a study would be able to make observations on how
depression symptoms and discrepancies between perceived and actual friend smoking

100

behaviors predict social network status (i.e. isolate, liaison, central) or smoking
behaviors. Multiple observations will also show any overlap between individual
disposition and smoking behaviors, as well as transmission of influence or selection
processes that may occur due to disposition or behavior. Understanding the complex
relationships between individual disposition, friend attributes, and smoking risk would
undoubtedly inform future prevention programming.
Study 3 explored how social competence and depression related to perceived
social influences. We hypothesized that depressed adolescents would be more sensitive
to social information that might confer peer acceptance or rejection. However, social
incompetence may better explain misperceptions of the social environment or may be
the source of peer acceptance or rejection. Results showed that perceptions of own
refusal skill efficacy were related to sociability and depression but a large overlap in the
variables existed. Sociability, once controlled for depression, was associated with
greater perceived refusal skill efficacy. The ability to connect with others and enjoy
social interaction appears to be a protective factor against smoking. One study, which
supported our results, found that perceived ease of making friends protected against
smoking (Page, Ihasz, Simonek, Klarova, & Hantiu, 2006). It may be that these
individuals have the social skills needed to communicate with peers or have the self-
esteem to refuse offers. Intervention programs often target refusal skills by teaching
communication strategies and building confidence in performing those strategies. This
finding suggests that this is an appropriate strategy for the average adolescent.

101

However, a significant inverse relationship between depression symptoms and refusal
skill efficacy remained and may point to vulnerable individuals who do not have the skill
set or confidence needed to refuse an offer. Researchers may need to closely consider
whether building communication skills and confidence is an attainable goal with these
adolescents given the budget and time allotted for their intervention program.
Depression may also be associated with biased perceptions of own abilities therefore,
cognitive manipulations may be needed in addition to communication skill development
in order to build complete self-efficacy when faced with a smoking offer. Further study
is needed to assess whether there is a perceptual difference in refusing a cigarette for
depressed and non-depressed adolescents. For depressed adolescents, refusal may
confer heightened risk for personal rejection while for non-depressed adolescents, it
may not mean anything more than not wanting to smoke. In a study that assessed
social perceptions, depression symptoms was prospectively associated with decreasing
accuracy in peer acceptance and inaccurate perceptions predicted increases in
depression symptoms (Kistner, 2006; Kistner, David-Ferdon, Repper, & Joiner, 2006).
Depressed individuals may use cigarette offers as a way to gauge acceptance and may
misinterpret smoking the cigarette when offered as peer acceptance while others may
simply view it as a polite gesture or not meaningful enough to change a friendship or
relationship between peers.
Deficits in social competence may play an important role in depression and social
influences. Study 3 found that the relationship between depression and perceived pro-

102

smoking social norms was modified by social deficit scores. Those who were depressed
and had greater social deficits had lower pro-smoking social norm beliefs compared to
their non-depressed counterparts. These individuals may represent a type or group of
adolescents who are not capable of perceiving distal or generalized social normative
messages in the environment. Assuming that pro-smoking norms are associated with
greater risk for smoking (Eisenberg & Forster, 2003; Page et al., 2006; Unger, Rohrbach,
Howard-Pitney, Ritt-Olson, & Mouttapa, 2001), then these adolescents with both
depression and social deficits may be at lower risk for smoking. However, this finding
does not eliminate the influence a close person may have on someone with social
deficits and depression. Future studies should assess social competence, depression,
and peer influence/selection together in order to get a clearer picture on the
interactions which may produce vulnerabilities.
LIMITATIONS
There are several limitations to these studies that warrant discussion. The
expression of depressive symptoms may differ between western individualistic (i.e.
United States) and eastern collectivistic (i.e. Chinese) societies. Somatic symptoms of
depression were found to be more common among non-western societies (Parker,
Gladstone, & Chee, 2001; Ryder et al., 2008). A recent study that compared somatic
and cognitive symptoms between a sample of U.S. and Hong Kong Chinese adolescents
found that somatic symptoms were indeed higher among the Chinese sample but
cognitive symptoms were equal between the two groups (Stewart et al., 2002). In a

103

cross-cultural comparison study between Mainland Chinese and Ontario-Canadian
adults, more psychological symptoms were endorsed across three methods of
assessment by the North American participants than the Chinese participants, while
somatic symptom endorsement were mixed (Ryder et al., 2008). The authors surmised
that the difference in psychological symptom endorsement levels was due to the
Chinese sample’s tendency to not value cognitive emotional states and reduce the
centrality of those emotions to their lives. Thus, the inconsistencies between these
studies on somatic versus cognitive symptom endorsements between Chinese and
North American (U.S. and Canadian) samples suggest that this is an area that needs
further research in order to generalize any findings related to depression across
cultures. However, other studies offer optimism that our findings will contribute to the
depression literature, despite these limitations. Self-report assessments of depression
symptoms were conducted with Mainland Chinese children (Chen, Rubin, & Li, 1995). In
this study, the western-based Children’s Depression Inventory (CDI; Kovacs, 1985) was
used to assess depressive symptomatology and the prevalence of children scoring
greater than 20 on the CDI (indicating high levels of depressive symptoms) was identical
to validity studies performed among western children (Kovacs, 1985). Another study
using a western self-report assessment for depression symptoms also found support for
potential cross-cultural validity. The factor structure of the Center for Epidemiological
Studies Depression Scale (CES-D; Radloff, 1977; Radloff, 1991) was examined using a
Hong Kong Chinese youth sample (Lee et al., 2008). The results of the CES-D study

104

found that Hong Kong youth are able to distinguish between somatic and affective
symptoms of depression but experience sadness as a combination of the two types of
symptoms. Self-report cognitive symptoms of depression may be successfully assessed
among Chinese samples. Since our investigations utilize a combination of cognitive and
somatic symptom measures of depression, we are confident in our ability to make
reasonable conclusions relevant to depressive symptoms in our study population.
Individualistic societies reward independence. The absence of social inhibition
and withdrawal are indicators of healthy self-esteem and social adjustment (Markus &
Kitayama, 1991; Singelis, Triandis, Bhawuk, & Gelfand, 1995; Triandis, 1995). In
contrast, collectivistic societies reward dependence and encourage caution, social
inhibition, and shyness (Markus & Kitayama, 1991). Adolescent identity is therefore
culturally determined and may influence the behavioral choices one would make within
a social group. For example, pro-smoking attributions that represent freedom,
independence, and autonomy may be more appealing to adolescents within
individualistic environments. For those who maintain collectivist ideals, pro-smoking
messages that value belongingness, conformity, and harmony may be more powerful in
influencing smoking behavior. Therefore, the reasons for smoking are highly dependent
on the cultural norm and social environment of the individual and may not be
generalizable across cultures. Several studies that examine risk factors for smoking
among Chinese adolescents indicate great similarity with their Western counterparts
(Grenard et al., 2006; Zhu, et al., 1996). Peer influences remain one of the strongest

105

indicators for smoking risk for both U.S. and Chinese adolescents (Grenard et al., 2006;
Peters, Hedley, Lam, Betson, & Wong, 1997; Unger & Rohrbach, 2002; Zhang, Wang,
Zhao, & Vartiainen, 2000). In a cross-cultural comparison between Mainland Chinese
and U.S. adolescents, a study found that friend smoking prevalence was strongly
associated with smoking and the strength of the relationship did not differ between the
two samples (Unger et al., 2002). A mediation study among Mainland Chinese
adolescents found that perceived best friend smoking, perceived peer estimates of
smoking, and having pro-smoking attitudes increased the likelihood of smoking six-
months later (Chen et al., 2006). In the same study, researchers found evidence that
perceived peer prevalence partially mediated the effect of pro-smoking attitudes on
adolescent smoking (Chen et al., 2006). These findings are entirely consistent with
social influences and social comparison theories used to develop western smoking
prevention programs. Such programs have shown to produce successful prevention
effects among Chinese cohorts (Chen, Fang, Li, Stanton, & Lin, 2006; Chou et al., 2006;
Sun et al., 2007). Thus, we are confident that the relationship between social influences
and smoking behaviors may be generalizable across cultures but perhaps in our case,
only among males. Furthermore, support for cross-cultural application of theoretical
models in the current studies is encouraging. Given the similarities between China and
U.S. samples on the constituent parts of the model (depression, social influences,
smoking), we have reason to be optimistic that our findings may be generalizable at

106

least between Mainland China, Hong Kong-China, and U.S. adolescent samples. Further
validation studies are warranted.
In studies 1 and 2, Social Normative Beliefs were defined as socially accepted
attitudes or beliefs regarding smoking. This was assessed by asking students if they
thought smokers were more popular, make young people look cooler, and make young
people more attractive. China, as a collectivistic society, values conformity and
harmony. They may define popularity by an individuals’ capacity to get along with
everyone. In contrast, popularity in the U.S. may be defined by an individuals’ capacity
to get noticed by engaging in novel or high risk behaviors thus rewarding independence
or separation from the group. This example highlights how the same item may be
interpreted differently based on socially accepted definitions. Analyses in these studies
will not be able to distinguish the qualitative differences that may be present in the
definitions of “popular” and “cool” or other social norms. However, in a focus group
conducted with a subsample from the Study 3 cohort, looking cool or mature and other
commonly held attributes of smokers were among the most common reasons for why
one would smoke but would only be applicable for males. Reasons for girls smoking and
what types of attributes are associated with a female smoker remained mixed. This
might explain why the results for Study 3 were also mixed for females and it may be due
to the changing culture and growing influence of westernization in China.
Receptivity to social influence program messages by depressed adolescents may not
be supported in China. Since the Wuhan Smoking Prevention Trial (WSPT) was largely

107

derived from a Western social influences based substance abuse prevention program
(Project SMART; Graham, Johnson, Hansen, Flay, & Gee, 1990; Pentz et al., 1989), the
program mediators (i.e. modifiable risk factors associated with smoking onset) may not
have been culturally attuned for Chinese adolescents. We have previously discussed
how definitions of norms (e.g. popular, cool) may be different between Chinese and U.S.
groups and how qualitative interpretation of norms may be important in determining
the true relationships leading to smoking behavioral choices. However, this discussion is
most relevant in developing future interventions by understanding the mechanism of
change. Study 2 was the first, to our knowledge, to examine these mediation pathways
longitudinally in a Chinese cohort. However, since the program condition did not
predict a change in social norms it may mean that the program was unsuccessful in
operationalizing social norms constructs or the way that the material was presented was
not acceptable to this population. For example, the WSPT program asks students to
discuss among themselves if they believe smoking is cool. One of the discussion points
stipulate that if students say they don’t think it’s cool but others kids do, then ask them,
“Do you agree with them? If they said everybody thinks it’s cool to smoke, what would
you say?” In terms of operationalizing, this example may not have been effective in
changing students’ beliefs regarding “smoking is cool.” In terms of cultural
appropriateness, the question was posed: “what if they said everyone else thinks it’s
cool, what would you say?” In a collectivistic society, it would be expected that students
would sway toward agreement with the statement that smoking is cool if they do

108

perceive everyone else believed it as well. The biggest challenge for future
interventions in China would be the operationalizing of an equivalent refusal skills
lesson. Refusing a cigarette offer is considered rude and goes against the cultural norm
in such a way that a program may lose credibility with Chinese adolescents if refusal skill
training is not thoroughly reconsidered and specifically tailored for the Chinese culture.
Again, our focus group data confirmed that refusing a cigarette is too difficult and in
direct opposition of cultural norms. Further study into culturally appropriate constructs
that would produce healthy behavior change among Chinese adolescents is needed.
FUTURE DIRECTIONS
Depression and smoking during adolescents may be linked through social
perceptions and interactions. Although this dissertation framed social influence as a risk
for smoking, friends can have positive and protective effects against smoking (Ennett &
Bauman, 1994; Maxwell, 2002). Belonging to a predominantly non-smoking peer
network (Ennett & Bauman, 1994; Mercken, Candel, Willems, & de Vries, 2009) or
identifying with multiple groups with opposing smoking norms appear to protect against
smoking (Verkooijen, de Vries, & Nielsen, 2007). Through social comparison and
vicarious learning, adolescents learn about appropriate and acceptable behaviors along
with the consequences of those behaviors. Smoking is largely a socially driven behavior
in which the rituals of smoking, social benefits of engaging with others who smoke and
the rewards associated with smoking have all contributed to smoking risk. Studies have
shown how social influences, whether through direct interactions with or observations

109

of friends and peers,(Ennett & Karl E. Bauman, 1994; Hoffman, Sussman, Unger, &
Valente, 2006; Landrine, Richardson, Klonoff, & Flay, 1994; Simons-Morton, 2007) or
through popular media (Audrain-McGovern et al., 2003; Tercyak, Goldman, Smith, &
Audrain, 2002; Weiss, Spruijt-Metz, Palmer, Chou, & Johnson, 2006; Xiao & Kohrman,
2008) will continue to affect risk behaviors like smoking into the future. Interventions
should utilize positive peer relationships to protect against smoking behaviors by
fostering peer interactions, communication skills, and tolerance for different levels of
social competence. Increasing peer interactions through group activities will force
individuals, who would otherwise not have met or initiated contact, to build
communication skills and work together in a positive environment. If this strategy is
used in conjunction with smoking prevention messages, then adoption of anti-smoking
norms would be more salient for the group because of the confirmation of those beliefs
among group members.
Noticing smoking in the environment was found to be a greater predictor of
smoking behavior than perceived peer prevalence (Eisenberg & Forster, 2003).
Depressed adolescents may be highly sensitive in detecting smoking behaviors because
of the novelty of the behavior or the attributes smoking may convey. Depressed
adolescents have shown to overestimate other problem behaviors among their peers
(Jacobs & Johnston, 2005). Current social-influences based prevention programs utilize
influence by correcting overestimates of smoking behaviors and beliefs and by fostering
an anti-smoking norm environment. Future interventions should go one step further

110

and include activities that allow confirmation of negative social norms of smoking by
peers. These activities need to be interactive and relegated to peer groups for
discussions rather than dictated from a teacher or other authority figure. Students will
need to feel that the conclusions that are drawn (i.e. not as many kids smoke as I
thought, most other students think smoking is not cool) are authentic and come from
other peers. The moment of learning is not necessarily when true rates of smoking are
revealed to students but when students are able to comprehend that everyone thinks
that everyone else thinks smoking is cool but in private close to no one actually believes
smoking is cool. The more uncertain an individual is about the accuracy of their
judgment, the more susceptible they become to social influence (Deutsch & Gerard,
1955; Kaplan & Miller, 1987; Katkin, Blum Sasmor, & Tan, 1966). The discrepancy
between what one believes vs. what they think others believe, leaves students exposed
to the messages of the program which can dictate an anti-smoking norm. It allows
students to question their observations and realize that they could perceive the world
incorrectly and should modify their perceptions to fit in with the norm—a strategy often
used in cognitive behavioral therapy for clinically depressed individuals. This strategy
may be especially strong for depressed individuals due to heighted motivation to fit in
with perceived norms and may lead them to either re-evaluate the smoking cues in their
environment or accept the messages of the prevention program at face value once
confirmed by other peers.

111

Adolescents who associated with very popular students (those receiving the
most nominations) received some protective benefit against depression symptoms
while popular students were not any more or less likely to have depression symptoms
than the average student (Ueno, 2005). It may be that sense of belonging or the
positive affect associated with social integration protects against depression (Ueno,
2005). Therefore, smoking prevention interventions that have social skill building and
encourage group work may increase belonging and thereby protect against depression
symptoms too. Depressed individuals are also more prone to conformity under social
pressure compared to non-depressed counterparts (Katkin et al., 1966). Anti-smoking
norms could be magnified in close proximity with other peers and depressed
adolescents may experience heightened or more intense pressure to conform to those
norms. Protecting depressed adolescents from negative influences could be achieved
by creating activities that focus on everyone’s ability to be a positive influence to those
around them, to model pro-health behaviors and social interactions, and developing
personal strengths to build self-esteem and social integration.
Interventions should also consider targeting the prevention of depression
symptoms. Social competence and depression symptoms have complex interactions
that need further study but it is clear that positive social interactions are important in
the development of competencies and prevention of depression. Individuals with social
deficits may have been ostracized or excluded from their peer networks due to those
social deficits, exacerbating or triggering depressive symptoms. These rejected

112

individuals may congregate with other deviant peers to cope or find acceptance (Besic &
Kerr, 2009). Communication is not likely to improve within this group if social deficits
are high and more importantly, exposure or influence for unhealthy behaviors places
these individuals at greater risk for negative outcomes.
Interventions that specifically teach social skills, communication skills, and
problem solving have been found to increase well-being and reduce depressive
symptoms (Domitrovich, Cortes, & Greenberg, 2007; Riggs, Greenberg, Kusche, & Pentz,
2006). Smoking prevention programs also target these skills and have found program
effects to be especially positive among depressed and hostile adolescents (Johnson et
al., 2005; Sun et al., 2007; Unger et al., 2004). Furthermore, a smoking prevention study
found that emotional intelligence, or the ability to read and interpret emotions in self
and others, was positively associated with recognition of social consequences of
smoking (Trinidad, Unger, Chou, & Johnson, 2005). However, teaching these skills may
not be enough to protect against further development of depression symptoms. These
skills will need to be practiced among a wide range of individuals and successful
episodes need to be experienced for those at greatest risk for depression.
Communication strategies may be taught to adolescents with social deficits but normal
adolescents will also need to be taught strategies to interact kindly and with tolerance
for atypical social encounters. As discussed earlier, interventions that focuses on
developing and recognizing individual strengths as a communication tool, increased
peer interaction across different networks, and manipulating peer influence to

113

disseminate pro-health or positive behaviors (as opposed to pro-smoking or other
negative health behavior) will greatly contribute to the prevention of unhealthy
outcomes.

114

BIBLIOGRAPHY
Allen, J. P., Insabella, G., Porter, M. R., Smith, F. D., Land, D., & Phillips, N. (2006). A Social-
Interactional Model of the Development of Depressive Symptoms in Adolescence.
Journal of Consulting and Clinical Psychology, 74(1), 55-65.

Allen, J. P., Porter, M., McFarland, C., McElhaney, K. B., & Marsh, P. (2007). The relation of
attachment security to adolescents' paternal and peer relationships, depression, and
externalizing behavior. Child Development, 78(4), 1222-1239.

Allen, J. P., Porter, M. R., & McFarland, F. C. (2006). Leaders and followers in adolescent close
friendships: Susceptibility to peer influence as a predictor of risky behavior, friendship
instability, and depression. Development and Psychopathology, 18(1), 155-172.

Aloise-Young, P. A., Graham, J. W., & Hansen, W. B. (1994). Peer influence on smoking initiation
during early adolescence: A comparison of group members and group outsiders. Journal
of Applied Psychology, 79(2), 281-287.

Andersen, S. L., & Teicher, M. H. (2008). Stress, sensitive periods and maturational events in
adolescent depression. Trends in Neurosciences, 31(4), 183-191.

Arnett, J. J. (2007). The myth of peer influence in adolescent smoking initiation. Health Educ
Behav, 34(4), 594-607.

Aseltine, R. H., Gore, S., & Colten, M. E. (1994). Depression and the social developmental
context of adolescence. Journal of Personality and Social Psychology, 67(2), 252-263.

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, 161(7), 1224-1230.

Audrain-McGovern, J., Tercyak, K. P., Shields, A. E., Bush, A., Espinel, C. F., & Lerman, C. (2003).
Which Adolescents are Most Receptive to Tobacco Industry Marketing? Implications for
Counter-Advertising Campaigns. Health Communication, 15(4), 499-513.

Austin, P. C., Grootendorst, P., & Anderson, G. M. (2007). A comparison of the ability of different
propensity score models to balance measured variables between treated and untreated
subjects: a Monte Carlo study. Stat Med, 26(4), 734-753.

Bandura, A. (1977). Social Learning Theory. New York: General Learning Press.

Baron, R. M., & Kenny, D. A. (1986). The moderator-mediator variable distinction in social
psychological research: Conceptual, strategic, and statistical considerations. Journal of
Personality and Social Psychology, 51(6), 1173-1182.

115

Beam, M. R., Gil-Rivas, V., Greenberger, E., & Chen, C. (2002). Adolescent problem behavior and
depressed mood: Risk and protection within and across social contexts. Journal of Youth
and Adolescence, 31(5), 343-357.

Beck, A. T. (1987). Cognitive models of depression. Journal of Cognitive Psychotherapy, 1(1), 5-
37.

Beck, A. T., Brown, G., Steer, R. A., Eidelson, J. I., & Riskind, J. H. (1987). Differentiating anxiety
and depression: A test of the cognitive content-specificity hypothesis. Journal of
Abnormal Psychology, 96(3), 179-183.

Beck, A. T., & Clark, D. A. (1991). Anxiety and depression: An information processing perspective.
Schwarzer, Ralf.

Besic, N., & Kerr, M. (2009). Punks, Goths, and Other Eye-Catching Peer Crowds: Do They Fulfill a
Function for Shy Youths? Journal of Research on Adolescence, 19(1), 113-121.

Birmaher, B., Bridge, J. A., Williamson, D. E., Brent, D. A., Dahl, R. E., Axelson, D. A., et al. (2004).
Psychosocial functioning in youths at high risk to develop major depressive disorder.
Journal of the American Academy of Child & Adolescent Psychiatry, 43(7), 839-846.

Birmaher, B., Ryan, N. D., Williamson, D. E., & Brent, D. A. (1996). Childhood and adolescent
depression: A review of the past 10 years, Part II. Journal of the American Academy of
Child & Adolescent Psychiatry, 35(12), 1575-1583.

Bond, R., & Smith, P. B. (1996). Culture and conformity: A meta-analysis of studies using Asch's
(1952b, 1956) line judgment task. Psychological Bulletin, 119(1), 111-137.

Bos, H. M. W., Sandfort, T. G. M., de Bruyn, E. H., & Hakvoort, E. M. (2008). Same-sex attraction,
social relationships, psychosocial functioning, and school performance in early
adolescence. Developmental Psychology, 44(1), 59-68.

Bosacki, S., Dane, A., & Marini, Z. (2007). Peer relationships and internalizing problems in
adolescents: Mediating role of self-esteem. Emotional & Behavioural Difficulties, 12(4),
261-282.

Botvin, G. J., Sussman, S., & Biglan, A. (2001). The Hutchinson Smoking Prevention Project: A
lesson on inaccurate media coverage and the importance of prevention advocacy.
Prevention Science, 2(2), 67-70.

Brendgen, M., Wanner, B., Morin, A. J. S., & Vitaro, F. (2005). Relations with Parents and with
Peers, Temperament, and Trajectories of Depressed Mood During Early Adolescence.
Journal of Abnormal Child Psychology, 33(5), 579-594.

116

Brook, J. S., Schuster, E., & Zhang, C. (2004). Cigarette smoking and depressive symptoms: A
longitudinal study of adolescents and young adults. Psychological Reports, 95(1), 159-
166.

Campbell, J. D., & Fairey, P. J. (1989). Informational and normative routes to conformity: The
effect of faction size as a function of norm extremity and attention to the stimulus.
Journal of Personality and Social Psychology, 57(3), 457-468.

Charlton, A., Minagawa, K.-e., & While, D. (1999). Saying "no" to cigarettes: A reappraisal of
adolescent refusal skills. Journal of Adolescence, 22(5), 695-707.

Chen, S.C. (2007). Smoking in China: Still Lighting Up [Electronic Version]. Forbes.com from
http://www.forbes.com/2007/09/26/china-smoking-trends-markets-equity-
cx_jc_0925markets03.html.

Chen, X., Fang, X., Li, X., Stanton, B., & Lin, D. (2006). Stay Away from Tobacco: A pilot trial of a
school-based adolescent smoking prevention program in Beijing, China. Nicotine &
Tobacco Research, 8(2), 227-237.

Chen, X., Rubin, K. H., & Li, B.-s. (1995). Depressed mood in Chinese children: Relations with
school performance and family environment. Journal of Consulting and Clinical
Psychology, 63(6), 938-947.

Chen, X., Stanton, B., Fang, X., Li, X., Lin, D., Zhang, J., et al. (2006). Perceived smoking norms,
socioenvironmental factors, personal attitudes and adolescent smoking in China: A
mediation analysis with longitudinal data. Journal of Adolescent Health, 38(4), 359-368.

Cheng, T. O. (1999). Teenage smoking in China. Journal of Adolescence, 22(5), 607-620.

Chinet, L. o., Plancherel, B., Bolognini, M., Bernard, M., Laget, J., Daniele, G., et al. (2006).
Substance use and depression. Comparative course in adolescents. European Child &
Adolescent Psychiatry, 15(3), 149-155.

Chou, C. P., Li, Y., Unger, J. B., Xia, J., Sun, P., Guo, Q., et al. (2006). A randomized intervention of
smoking for adolescents in urban Wuhan, China. Prev Med, 42(4), 280-285.

Cole, D. A., Maxwell, S. E., Martin, J. M., Peeke, L. G., Seroczynski, A. D., Tram, J. M., et al. (2001).
The development of multiple domains of child and adolescent self-concept: A cohort
sequential longitudinal design. Child Development, 72(6), 1723-1746.

Covey, L. S., Glassman, A. H., & Stetner, F. (1998). Cigarette smoking and major depression.
Journal of Addictive Diseases, 17(1), 35-46.

117

Crum, R. M., Storr, C. L., Ialongo, N., & Anthony, J. C. (2008). Is depressed mood in childhood
associated with an increased risk for initiation of alcohol use during early adolescence?
Addictive Behaviors, 33(1), 24-40.

Dahl, R. E., & Spear, L. P. (2004). Adolescent brain development: Vulnerabilities and
opportunities. Annals of the New York Academy of Sciences, Vol, 472.

Davey, C. G., Yucel, M., & Allen, N. B. (2008). The emergence of depression in adolescence:
Development of the prefrontal cortex and the representation of reward. Neuroscience &
Biobehavioral Reviews, 32(1), 1-19.

Deb, S., Mitra, K., & Mukherjee, I. (2001). Adolescents' Perceived Social Norms and Values: A
Calcutta-based Study. Journal of the Indian Academy of Applied Psychology, 27(1-2),
127-135.

Deutsch, M., & Gerard, H. B. (1955). A study of normative and informational social influences
upon individual judgment. The Journal of Abnormal and Social Psychology. Vol., 51(3),
629-636.

Dierker, L. C., Vesel, F., Sledjeski, E. M., Costello, D., & Perrine, N. (2007). Testing the dual
pathway hypothesis to substance use in adolescence and young adulthood. Drug and
Alcohol Dependence, 87(1), 83-93.

Domitrovich, C. E., Cortes, R. C., & Greenberg, M. T. (2007). Improving young children's social
and emotional competence: a randomized trial of the preschool "PATHS" curriculum. J
Prim Prev, 28(2), 67-91.

Edwards, J. R., & Lambert, L. S. (2007). Methods for integrating moderation and mediation: A
general analytical framework using moderated path analysis. Psychological Methods,
12(1), 1-22.

Eisenberg, M. E., & Forster, J. L. (2003). Adolescent Smoking Behavior: Measures of Social
Norms. American Journal of Preventive Medicine, 25(2), 122-128.

Ennett, S. T., & Bauman, K. E. (1994). The contribution of influence and selection to adolescent
peer group homogeneity: The case of adolescent cigarette smoking. Journal of
Personality and Social Psychology, 67(4), 653-663.

Ennett, S. T., & Bauman, K. E. (1994). The contribution of influence and selection to adolescent
peer group homogeneity: the case of adolescent cigarette smoking. J Pers Soc Psychol,
67(4), 653-663.

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(8), 1357-1367.

118

Fergusson, D. M., Lynskey, M. T., & Horwood, L. J. (1996). Comorbidity between depressive
disorders and nicotine dependence in a cohort of 16-year-olds. Arch Gen Psychiatry,
53(11), 1043-1047.

Festinger, L. (1954). A theory of social comparison processes. Human Relations, 7, 117-140.

Field, T., Diego, M., & Sanders, C. (2001). Adolescent depression and risk factors. Adolescence,
36(143), 491-498.

Gerard, H. B., & Rotter, G. S. (1961). Time perspective, consistency of attitude, and social
influence. The Journal of Abnormal and Social Psychology, 62(3), 565-572.

Glassman, A. H., & Covey, L. S. (1996). Smoking and affective disorder. American Journal of
Health Behavior, 20(5), 279-285.

Goodman, E., & Capitman, J. (2000). Depressive symptoms and cigarette smoking among teens.
Pediatrics, 106(4), 748-755.

Graham, J. W., & Donaldson, S. I. (1993). Evaluating interventions with differential attrition: The
importance of nonresponse mechanisms and use of follow-up data. Journal of Applied
Psychology, 78(1), 119-128.

Graham, J. W., Johnson, C. A., Hansen, W. B., Flay, B. R., & Gee, M. (1990). Drug use prevention
programs, gender, and ethnicity: evaluation of three seventh-grade Project SMART
cohorts. Prev Med, 19(3), 305-313.

Graham, J. W., Marks, G., & Hansen, W. B. (1991). Social influence processes affecting
adolescent substance use. Journal of Applied Psychology, 76(2), 291-298.

Gregory, A. M., Rijsdijk, F. h. V., Lau, J. Y. F., Napolitano, M., McGuffin, P., & Eley, T. C. (2007).
Genetic and environmental influences on interpersonal cognitions and associations with
depressive symptoms in 8-year-old twins. Journal of Abnormal Psychology, 116(4), 762-
775.

Grenard, J. L., Guo, Q., Jasuja, G. K., Unger, J. B., Chou, C.-P., Gallaher, P. E., et al. (2006).
Influences affecting adolescent smoking behavior in China. Nicotine & Tobacco
Research, 8(2), 245-255.

Grunkemeier, G. L., Payne, N., Jin, R., & Handy, J. R. (2002). Propensity score analysis of stroke
after off-pump coronary artery bypass grafting. The Annals of Thoracic Surgery, 74(2),
301-305.

Gunther, A. C., Bolt, D., Borzekowski, D. L. G., Liebhart, J. L., & Dillard, J. P. (2006). Presumed
Influence on Peer Norms: How Mass Media Indirectly Affect Adolescent Smoking.
Journal of Communication, 56(1), 52-68.

119

Gutman, L. M., & Sameroff, A. J. (2004). Continuities in depression from adolescence to young
adulthood: Contrasting ecological influences. Development and Psychopathology.
Special Issue: Transition from Adolescence to Adulthood, 16(4), 967-984.

Hankin, B. L., Mermelstein, R., & Roesch, L. (2007). Sex Differences in Adolescent Depression:
Stress Exposure and Reactivity Models. Child Development, 78(1), 279-295.

Hansen, W. B., & Graham, J. W. (1991). Preventing alcohol, marijuana, and cigarette use among
adolescents: peer pressure resistance training versus establishing conservative norms.
Prev Med, 20(3), 414-430.

Hansen, W. B., Graham, J. W., Wolkenstein, B. H., Lundy, B. Z., & et al. (1988). Differential
impact of three alcohol prevention curricula on hypothesized mediating variables.
Journal of Drug Education, 18(2), 143-153.

Hansen, W. B., Tobler, N. S., & Graham, J. W. (1990). Attrition in substance abuse prevention
research: A meta-analysis of 85 longitudinally followed cohorts. Evaluation Review,
14(6), 677-685.

Henrich, C. C., Blatt, S. J., Kuperminc, G. P., Zohar, A., & Leadbeater, B. J. (2001). Levels of
interpersonal concerns and social functioning in early adolescent boys and girls. Journal
of Personality Assessment, 76(1), 48-67.

Henry, D. B., & Kobus, K. (2007). Early adolescent social networks and substance use. Journal of
Early Adolescence, 27(3), 346-362.

Hermans, D., Vandromme, H., Debeer, E., Raes, F., Demyttenaere, K., Brunfaut, E., et al. (2008).
Overgeneral autobiographical memory predicts diagnostic status in depression.
Behaviour Research and Therapy, 46(5), 668-677.

Hoffman, B. R., Sussman, S., Unger, J. B., & Valente, T. W. (2006). Peer influences on adolescent
cigarette smoking: A theoretical review of the literature. Substance Use & Misuse, 41(1),
103-155.

Hoffman, K. B., Cole, D. A., Martin, J. M., Tram, J., & Seroczynski, A. D. (2000). Are the
discrepancies between self- and others' appraisals of competence predictive or
reflective of depressive symptoms in children and adolescents: A longitudinal study, part
II. Journal of Abnormal Psychology, 109(4), 651-662.

Jacobs, J. E., & Johnston, K. E. (2005). "Everyone Else Is Doing It": Relations Between Bias in
Base-Rate Estimates and Involvement in Deviant Behaviors. Jacobs, Janis E.

Jellinek, M. S., & Snyder, J. B. (1998). Depression and suicide in children and adolescents. Pediatr
Rev, 19(8), 255-264.

120

Johnson, C. A., Cen, S., Gallaher, P., Palmer, P. H., Xiao, L., Ritt-Olson, A., et al. (2007). Why
smoking prevention programs sometimes fail. Does effectiveness depend on
sociocultural context and individual characteristics? Cancer Epidemiol Biomarkers Prev,
16(6), 1043-1049.

Johnson, C. A., Unger, J. B., Ritt-Olson, A., Palmer, P. H., Cen, S. Y., Gallaher, P., et al. (2005).
Smoking prevention for ethnically diverse adolescents: 2-year outcomes of a
multicultural, school-based smoking prevention curriculum in Southern California. Prev
Med, 40(6), 842-852.

Joiner, T., & Coyne, J. C. (1999). The interactional nature of depression: Advances in
interpersonal approaches. Washington, DC, US: American Psychological Association,
423.

Joiner, T. E., Jr., Kistner, J. A., Stellrecht, N. E., & Merrill, K. A. (2006). On seeing clearly and
thriving: Interpersonal perspicacity as adaptive (not depressive) realism (or where three
theories meet). Journal of Social & Clinical Psychology, 25(5), 542-564.

Joiner, T. E., Jr., Metalsky, G. I., Lew, A., & Klocek, J. (1999). Testing the causal mediation
component of Beck's theory of depression: Evidence for specific mediation. Cognitive
Therapy and Research, 23(4), 401-412.

Joormann, J., & Gotlib, I. H. (2007). Selective attention to emotional faces following recovery
from depression. Journal of Abnormal Psychology, 116(1), 80-85.

Kaplan, M. F., & Miller, C. E. (1987). Group decision making and normative versus informational
influence: Effects of type of issue and assigned decision rule. Journal of Personality and
Social Psychology, 53(2), 306-313.

Katkin, E. S., Blum Sasmor, D., & Tan, R. (1966). Conformity and achievement-related
characteristics of depressed patients. Journal of Abnormal Psychology, 71(6), 407-412.

Keenan-Miller, D., Hammen, C. L., & Brennan, P. A. (2007). Health outcomes related to early
adolescent depression. Journal of Adolescent Health, 41(3), 256-262.

Kessler, R. C., Avenevoli, S., & Ries Merikangas, K. (2001). Mood disorders in children and
adolescents: an epidemiologic perspective. Biological Psychiatry, 49(12), 1002-1014.

Killeya-Jones, L. A., Nakajima, R., & Costanzo, P. R. (2007). Peer standing and substance use in
early-adolescent grade-level networks: A short-term longitudinal study. Prevention
Science, 8(1), 11-23.

Kistner, J. (2006). Children's peer acceptance, perceived acceptance, and risk for depression. In
T. E. Joiner, J. S. Brown & J. Kistner (Eds.), The interpersonal, cognitive, and social nature
of depression. (pp. 1-21). Mahwah, NJ, US: Lawrence Erlbaum Associates Publishers.

121

Kistner, J. A., David-Ferdon, C. F., Repper, K. K., & Joiner, T. E., Jr. (2006). Bias and Accuracy of
Children's Perceptions of Peer Acceptance: Prospective Associations with Depressive
Symptoms. Journal of Abnormal Child Psychology, 34(3), 349-361.

Kobus, K. (2003). Peers and adolescent smoking. Addiction, 98 Suppl 1, 37-55.

Kovacs, M. (1985). The Children's Depression Inventory. Psychopharmacology Bulletin, 21, 995-
998.

Krull, J. L., & MacKinnon, D. P. (1999). Multilevel mediation modeling in group-based
intervention studies. Evaluation Review, 23(4), 418-444.

Ladouceur, C. D., Dahl, R. E., Williamson, D. E., Birmaher, B., Axelson, D. A., Ryan, N. D., et al.
(2006). Processing emotional facial expressions influences performance on a Go/NoGo
task in pediatric anxiety and depression. Journal of Child Psychology and Psychiatry,
47(11), 1107-1115.

Ladouceur, C. D., Dahl, R. E., Williamson, D. E., Birmaher, B., Ryan, N. D., & Casey, B. J. (2005).
Altered Emotional Processing in Pediatric Anxiety, Depression, and Comorbid Anxiety-
Depression. Journal of Abnormal Child Psychology, 33(2), 165-177.

Lai, M. K., Ho, S. Y., & Lam, T. H. (2004). Perceived peer smoking prevalence and its association
with smoking behaviours and intentions in Hong Kong Chinese adolescents. Addiction,
99(9), 1195-1205.

Lakdawalla, Z., Hankin, B. L., & Mermelstein, R. (2007). Cognitive theories of depression in
children and adolescents: A conceptual and quantitative review. Clinical Child and
Family Psychology Review, 10(1), 1-24.

Landrine, H., Richardson, J. L., Klonoff, E. A., & Flay, B. R. (1994). Cultural diversity in the
predictors of adolescent cigarette smoking: The relative influence of peers. Journal of
Behavioral Medicine, 17(3), 331-346.

Lara, M. E., & Klein, D. N. (1999). Psychosocial processes underlying the maintenance and
persistence of depression: implications for understanding chronic depression. Clin
Psychol Rev, 19(5), 553-570.

Lee, S. W., Stewart, S. M., Byrne, B. M., Wong, J. P. S., Ho, S. Y., Lee, P. W. H., et al. (2008). Factor
structure of the Center for Epidemiological Studies Depression Scale in Hong Kong
adolescents. Journal of Personality Assessment. Special Issue: Empirical correlates of the
MMPI-2 restructured clinical (RC) scales in mental health, forensic, and nonclinical
settings, 90(2), 175-184.

122

Lev-Wiesel, R., Nuttman-Shwartz, O., & Sternberg, R. (2006). Peer rejection during adolescence:
Psychological long-term effects--A brief report. Journal of Loss & Trauma, 11(2), 131-
142.

Lewinsohn, P. M. (1974). A behavioral approach to depression. In R. Friedman & M.M. Katz
(Eds.), The psychology of depression: Contemporary theory and research (pp 318).
Oxford, England: John Wiley & Sons.

Lewinsohn, P. M., & Gotlib, I. H. (1995). Behavioral theory and treatment of depression. In E. E.
Beckham, & W. R. Leber (Eds.), Handbook of depression (2nd ed.). (pp. 352-375). New
York, NY, US: Guilford Press.

MacKinnon, D. P. (2008). Introduction to statistical mediation analysis. New York, NY: Taylor &
Francis Group/Lawrence Erlbaum Associates.

MacKinnon, D. P., & Dwyer, J. H. (1993). Estimating mediated effects in prevention studies.
Evaluation Review, 17(2), 144-158.

MacKinnon, D. P., Johnson, C. A., Pentz, M. A., Dwyer, J. H., Hansen, W. B., Flay, B. R., et al.
(1991). Mediating mechanisms in a school-based drug prevention program: first-year
effects of the Midwestern Prevention Project. Health Psychol, 10(3), 164-172.

MacKinnon, D. P., Lockwood, C. M., Hoffman, J. M., West, S. G., & Sheets, V. (2002). A
comparison of methods to test mediation and other intervening variable effects.
Psychological Methods, 7(1), 83-104.

MacKinnon, D. P., & Luecken, L. J. (2008). How and for whom? Mediation and moderation in
health psychology. Health Psychology, 27(2, Suppl), S99-S100.

MacKinnon, D. P., Weber, M. D., & Pentz, M. A. (1988). How do school-based drug prevention
programs work and for whom? Drugs & Society, 3(1-2), 125-143.

MacPhee, A. R., & Andrews, J. J. W. (2006). Risk factors for depression in early adolescence.
Adolescence, 41(163), 435-466.

Makri-Botsari, E. (2005). Risk/protective effects on adolescent depression: Role of individual,
family and peer factors. Psychological Studies, 50(1), 50-61.

Marcus, D. K., & Nardone, M. E. (1992). Depression and interpersonal rejection. Clinical
Psychology Review, 12(4), 433-449.

Markus, H. R., & Kitayama, S. (1991). Culture and the self: Implications for cognition, emotion,
and motivation. Psychological Review, 98(2), 224-253.

123

Maxwell, K. A. (2002). Friends: The role of peer influence across adolescent risk behaviors.
Journal of Youth and Adolescence, 31(4), 267-277.

McCaffery, J. M., Papandonatos, G. D., Stanton, C., Lloyd-Richardson, E. E., & Niaura, R. (2008).
Depressive symptoms and cigarette smoking in twins from the National Longitudinal
Study of Adolescent Health. Health Psychology. Special Issue: Tobacco and health
psychology, 27(3, Suppl), S207-S215.

Mercken, L., Candel, M., Willems, P., & de Vries, H. (2007). Disentangling social selection and
social influence effects on adolescent smoking: the importance of reciprocity in
friendships. Addiction, 102(9), 1483-1492.

Mercken, L., Candel, M., Willems, P., & de Vries, H. (2009). Social influence and selection effects
in the context of smoking behavior: changes during early and mid adolescence. Health
Psychol, 28(1), 73-82.

Montague, M., Enders, C., Dietz, S., Dixon, J., & Cavendish, W. M. (2008). A longitudinal study of
depressive symptomology and self-concept in adolescents. The Journal of Special
Education, 42(2), 67-78.

Morgan-Lopez, A. A., & MacKinnon, D. P. (2006). Demonstration and evaluation of a method for
assessing mediated moderation. Behavior Research Methods, 38(1), 77-87.

Muller, D., Judd, C. M., & Yzerbyt, V. Y. (2005). When moderation is mediated and mediation is
moderated. Journal of Personality and Social Psychology, 89(6), 852-863.

Munafo, M. R., Hitsman, B., Rende, R., Metcalfe, C., & Niaura, R. (2008). Effects of progression to
cigarette smoking on depressed mood in adolescents: Evidence from the National
Longitudinal Study of Adolescent Health. Addiction, 103(1), 162-171.

Needham, B. L. (2007). Gender differences in trajectories of depressive symptomatology and
substance use during the transition from adolescence to young adulthood. Social
Science & Medicine, 65(6), 1166-1179.

Nelson, E. E., Leibenluft, E., McClure, E., & Pine, D. S. (2005). The social re-orientation of
adolescence: A neuroscience perspective on the process and its relation to
psychopathology. Psychological Medicine, 35(2), 163-174.

Neshat-Doost, H. T., Moradi, A. R., Taghavi, M. R., Yule, W., & Dalgleish, T. (2000). Lack of
attentional bias for emotional information in clinically depressed children and
adolescents on the dot probe task. Journal of Child Psychology and Psychiatry, 41(3),
363-368.
Newman, B. M., Newman, P. R., Griffen, S., O'Connor, K., & Spas, J. (2007). The relationship of
social support to depressive symptoms during the transition to high school.
Adolescence, 42(167), 441-459.

124

Nrugham, L., Larsson, B., & Sund, A. M. (2008). Specific depressive symptoms and disorders as
associates and predictors of suicidal acts across adolescence. J Affect Disord, 111(1), 83-
93.

Oldehinkel, A. J., Rosmalen, J. G. M., Veenstra, R., Dijkstra, J. K., & Ormel, J. (2007). Being
admired or being liked: Classroom social status and depressive problems in early
adolescent girls and boys. Journal of Abnormal Child Psychology, 35(3), 417-427.

Page, R. M., Ihasz, F., Simonek, J., Klarova, R., & Hantiu, I. (2006). Cigarette Smoking, Friendship
Factors, and Social Norm Perceptions Among Central and Eastern European High School
Students. Journal of Drug Education, 36(3), 213-231.

Parker, G., Cheah, Y. C., & Roy, K. (2001). Do the Chinese somatize depression? A cross-cultural
study. Soc Psychiatry Psychiatr Epidemiol, 36(6), 287-293.

Parker, G., Gladstone, G., & Tsee Chee, K. (2001). Depression in the planet's largest ethnic
group: The Chinese. American Journal of Psychiatry, 158(6), 857-864.

Patton, G. C., Carlin, J. B., Coffey, C., Wolfe, R., Hibbert, M., & Bowes, G. (1998). Depression,
anxiety, and smoking initiation: A prospective study over 3 years. American Journal of
Public Health, 88(10), 1518-1522.

Pearson, M., & Michell, L. (2000). Smoke rings: Social network analysis of friendship groups,
smoking and drug-taking. Drugs: Education, Prevention & Policy, 7(1), 21-37.

Pearson, M., Sweeting, H., West, P., Young, R., Gordon, J., & Turner, K. (2006). Adolescent
substance use in different social and peer contexts: A social network analysis. Drugs:
Education, Prevention & Policy, 13(6), 519-536.

Pedersen, S., Vitaro, F., Barker, E. D., & Borge, A. I. H. (2007). The timing of middle-childhood
peer rejection and friendship: Linking early behavior to early-adolescent adjustment.
Child Development, 78(4), 1037-1051.

Pentz, M. A., Dwyer, J. H., MacKinnon, D. P., Flay, B. R., Hansen, W. B., Wang, E. Y., et al. (1989).
A multicommunity trial for primary prevention of adolescent drug abuse. Effects on drug
use prevalence. Jama, 261(22), 3259-3266.

Peters, J., Hedley, A. J., Lam, T.-H., Betson, C. L., & Wong, C.-M. (1997). A comprehensive study
of smoking in primary school children in Hong Kong: Implications for prevention. Journal
of Epidemiology & Community Health, 51(3), 239-245.

Peterson, A. V., Jr., Kealey, K. A., Mann, S. L., Marek, P. M., & Sarason, I. G. (2000). Hutchinson
Smoking Prevention Project: long-term randomized trial in school-based tobacco use
prevention--results on smoking. J Natl Cancer Inst, 92(24), 1979-1991.

125

Radloff, L. S. (1977). The CES-D Scale: A self-report depression scale for research in the general
population. Applied Psychological Measurement, 1(3), 385-401.

Radloff, L. S. (1991). The use of the Center for Epidemiologic Studies Depression Scale in
adolescents and young adults. Journal of Youth and Adolescence. Special Issue: The
emergence of depressive symptoms during adolescence, 20(2), 149-166.

Rao, U. (2006). Links Between Depression and Substance Abuse in Adolescents: Neurobiological
Mechanisms. American Journal of Preventive Medicine, 31(6, Supplement 1), 161-174.

Reid, J. L., Manske, S. R., & Leatherdale, S. T. (2008). Factors related to adolescents' estimation
of peer smoking prevalence. Health Education Research, 23(1), 81-93.

Reijntjes, A., Dekovic, M., Vermande, M., & Telch, M. J. (2007). Children's feedback preferences
in response to an experimentally manipulated peer evaluation outcome: The role of
depressive symptoms. Journal of Abnormal Child Psychology, 35(3), 497-507.

Reijntjes, A., Stegge, H., & Terwogt, M. M. (2006). Children's Coping with Peer Rejection: The
Role of Depressive Symptoms, Social Competence, and Gender. Infant and Child
Development, 15(1), 89-107.

Reijntjes, A., Stegge, H., Terwogt, M. M., Kamphuis, J. H., & Telch, M. J. (2006). Children's Coping
with In Vivo Peer Rejection: An Experimental Investigation. Journal of Abnormal Child
Psychology, 34(6), 877-889.

Riggs, N. R., Greenberg, M. T., Kusche, C. A., & Pentz, M. A. (2006). The mediational role of
neurocognition in the behavioral outcomes of a social-emotional prevention program in
elementary school students: effects of the PATHS Curriculum. Prev Sci, 7(1), 91-102.

Ritt-Olson, A., Unger, J., Valente, T., Nezami, E., Chou, C.-P., Trinidad, D., et al. (2005). Exploring
Peers as a Mediator of the Association Between Depression and Smoking in Young
Adolescents. Substance Use & Misuse, 40(1), 77-98.

Rockhill, C. M., Vander Stoep, A., McCauley, E., & Katon, W. J. (2009). Social competence and
social support as mediators between comorbid depressive and conduct problems and
functional outcomes in middle school children. J Adolesc, 32(3), 535-553.

Rodriguez, D., Moss, H. B., & Audrain-McGovern, J. (2005). Developmental Heterogeneity in
Adolescent Depressive Symptoms: Associations With Smoking Behavior. Psychosomatic
Medicine, 67(2), 200-210.

Rudolph, K. D. (2008). Developmental influences on interpersonal stress generation in
depressed youth. Journal of Abnormal Psychology, 117(3), 673-679.

126

Rudolph, K. D., & Conley, C. S. (2005). The Socioemotional Costs and Benefits of Social-
Evaluative Concerns: Do Girls Care Too Much? Journal of Personality, 73(1), 115-137.

Rudolph, K. D., Hammen, C., & Burge, D. (1995). Cognitive representations of self, family, and
peers in school-age children: links with social competence and sociometric status. Child
Dev, 66(5), 1385-1402.

Rudolph, K. D., Hammen, C., & Burge, D. (1997). A cognitive-interpersonal approach to
depressive symptoms in preadolescent children. J Abnorm Child Psychol, 25(1), 33-45.

Rudolph, K. D., Ladd, G., & Dinella, L. (2007). Gender differences in the interpersonal
consequences of early-onset depressive symptoms. Merrill-Palmer Quarterly. Special
Issue: Gender and Friendships, 53(3), 461-488.

Ryder, A. G., Yang, J., Zhu, X., Yao, S., Yi, J., Heine, S. J., et al. (2008). The cultural shaping of
depression: somatic symptoms in China, psychological symptoms in North America? J
Abnorm Psychol, 117(2), 300-313.

Sacco, W. P., & Vaughan, C. A. (2006). Depression and the Response of Others: A Social-
Cognitive Interpersonal Process Model. Joiner, Thomas E.

Simon, G. E., VonKorff, M., Piccinelli, M., Fullerton, C., & Ormel, J. (1999). An international study
of the relation between somatic symptoms and depression. N Engl J Med, 341(18),
1329-1335.

Simons-Morton, B. (2007). Social influences on adolescent substance use. American Journal of
Health Behavior, 31(6), 672-684.

Singelis, T. M., Triandis, H. C., Bhawuk, D., & Gelfand, M. J. (1995). Horizontal and vertical
dimensions of individualism and collectivism: A theoretical and measurement
refinement. Cross-Cultural Research: The Journal of Comparative Social Science, 29(3),
240-275.

Skara, S., & Sussman, S. (2003). A review of 25 long-term adolescent tobacco and other drug use
prevention program evaluations. Prev Med, 37(5), 451-474.

Song, Y., Huang, Y., Liu, D., Kwan, J. S., Zhang, F., Sham, P. C., et al. (2008). Depression in college:
depressive symptoms and personality factors in Beijing and Hong Kong college
freshmen. Compr Psychiatry, 49(5), 496-502.

Stewart, S. M., Lewinsohn, P. M., Lee, P. W. H., Ho, L. M., Kennard, B., Hughes, C. W., et al.
(2002). Symptom patterns in depression and "subthreshold" depression among
adolescents in Hong Kong and the United States. Journal of Cross-Cultural Psychology,
33(6), 559-576.

127

Sun, P., Unger, J. B., Guo, Q., Gong, J., Ma, H., Palmer, P. H., et al. (2007). Comorbidity between
depression and smoking moderates the effect of a smoking prevention program among
boys in China. Nicotine Tob Res, 9 Suppl 4, S599-609.

Sussman, S., Hansen, W. B., Flay, B. R., & Botvin, G. J. (2001). Re: Hutchinson Smoking
Prevention Project: long-term randomized trial in school-based tobacco use prevention--
results on smoking. J Natl Cancer Inst, 93(16), 1267; author reply 1269-1271.

Taghavi, M. R., Neshat-Doost, H. T., Moradi, A. R., Yule, W., & Dalgleish, T. (1999). Biases in
visual attention in children and adolescents with clinical anxiety and mixed anxiety-
depression. Journal of Abnormal Child Psychology, 27(3), 215-223.

Tercyak, K. P., Goldman, P., Smith, A., & Audrain, J. (2002). Interacting effects of depression and
tobacco advertising receptivity on adolescent smoking. Journal of Pediatric Psychology,
27(2), 145-154.

Thomas, R. (2002). School-based programmes for preventing smoking. Cochrane Database Syst
Rev(4), CD001293.

Thomas, R., & Perera, R. (2006). School-based programmes for preventing smoking. Cochrane
Database Syst Rev, 3, CD001293.

Timbremont, B., & Braet, C. (2005). Selective information-processing in depressed children and
adolescents: Is there a difference in processing of self-referent and other-referent
information? Behaviour Change, 22(3), 143-155.

Tobler, N. S., Roona, M. R., Ochshorn, P., Marshall, D. G., Streke, A. V., & Stackpole, K. M. (2000).
School-based adolescent drug prevention programs: 1998 meta-analysis. Journal of
Primary Prevention, 20(4), 275-336.

Triandis, H. C. (1995). Individualism & collectivism. New directions in social psychology. Boulder,
CO, US: Westview Press, 259.

Trinidad, D. R., Unger, J. B., Chou, C. P., & Johnson, C. A. (2005). Emotional intelligence and
acculturation to the United States: interactions on the perceived social consequences of
smoking in early adolescents. Subst Use Misuse, 40(11), 1697-1706.

Ueno, K. (2005). The effects of friendship networks on adolescent depressive symptoms. Social
Science Research, 34(3), 484-510.

Uhrlass, D. J., Schofield, C. A., Coles, M. E., & Gibb, B. E. (2009). Self-perceived competence and
prospective changes in symptoms of depression and social anxiety. J Behav Ther Exp
Psychiatry, 40(2), 329-337.

128

Unger, J. B., Chou, C. P., Palmer, P. H., Ritt-Olson, A., Gallaher, P., Cen, S., et al. (2004). Project
FLAVOR: 1-Year Outcomes of a Multicultural, School-Based Smoking Prevention
Curriculum for Adolescents. Am J Public Health, 94(2), 263-265.

Unger, J. B., & Rohrbach, L. A. (2002). Why do adolescents overestimate their peers' smoking
prevalence? Correlates of prevalence estimates among California 8th-grade students.
Journal of Youth and Adolescence, 31(2), 147-153.

Unger, J. B., Rohrbach, L. A., Cruz, T. B., Baezconde-Garbanati, L., Howard, K. A., Palmer, P. H., et
al. (2001). Ethnic variation in peer influences on adolescent smoking. Nicotine Tob Res,
3(2), 167-176.

Unger, J. B., Rohrbach, L. A., Howard-Pitney, B., Ritt-Olson, A., & Mouttapa, M. (2001). Peer
influences and susceptibility to smoking among Californian adolescents. Substance Use
& Misuse, 36(5), 551-571.

Unger, J. B., Yan, L., Chen, X., Jiang, X., Azen, S., Qian, G., et al. (2001). Adolescent smoking in
Wuhan, China: Baseline data from the Wuhan smoking prevention trial. American
Journal of Preventive Medicine, 21(3), 162-169.

Unger, J. B., Yan, L., Shakib, S., Rohrbach, L. A., Chen, X., Qian, G., et al. (2002). Peer influences
and access to cigarettes as correlates of adolescent smoking: a cross-cultural
comparison of Wuhan, China, and California. Prev Med, 34(4), 476-484.

van Beek, Y., & Dubas, J. S. (2008). Decoding basic and non-basic facial expressions and
depressive symptoms in late childhood and adolescence. Journal of Nonverbal Behavior,
32(1), 53-64.

Van Voorhees, B. W., Paunesku, D., Kuwabara, S. A., Basu, A., Gollan, J., Hankin, B. L., et al.
(2008). Protective and vulnerability factors predicting new-onset depressive episode in a
representative of U.S. adolescents. Journal of Adolescent Health, 42(6), 605-616.

Verkooijen, K. T., de Vries, N. K., & Nielsen, G. A. (2007). Youth Crowds and Substance Use: The
Impact of Perceived Group Norm and Multiple Group Identification. Psychology of
Addictive Behaviors, 21(1), 55-61.

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.
Archives of Women's Mental Health, 9(3), 139-150.

Wang, M. Q., Eddy, J. M., & Fitzhugh, E. C. (2000). Smoking acquisition: peer influence and self-
selection. Psychol Rep, 86(3 Pt 2), 1241-1246.

129

Wang, M. Q., Fitzhugh, E. C., Turner, L., Fu, Q., & et al. (1996). Association of depressive
symptoms and school adolescents' smoking: A cross-lagged analysis. Psychological
Reports, 79(1), 127-130.

Weiss, J. W., Spruijt-Metz, D., Palmer, P. H., Chou, C. P., & Johnson, C. A. (2006). Smoking among
adolescents in China: an analysis based upon the meanings of smoking theory. Am J
Health Promot, 20(3), 171-178.

Wen, X., Chen, W., Muscat, J. E., Qian, Z., Lu, C., Zhang, C., et al. (2007). Modifiable family and
school environmental factors associated with smoking status among adolescents in
Guangzhou, China. Prev Med, 45(2-3), 189-197.

Whitton, S. W., Larson, J. J., & Hauser, S. T. (2008). Depressive symptoms and bias in perceived
social competence among young adults. Journal of Clinical Psychology, 64(7), 791-805.

Whooley, M. A., Avins, A. L., Miranda, J., & Browner, W. S. (1997). Case-finding instruments for
depression. Two questions are as good as many.[see comment]. Journal of General
Internal Medicine, 12(7), 439-445.

Williams, K. L., & Galliher, R. V. (2006). Predicting depression and self-esteem from social
connectedness, support, and competence. Journal of Social & Clinical Psychology, 25(8),
855-874.

Williamson, D. E., Forbes, E. E., Dahl, R. E., & Ryan, N. D. (2005). A Genetic Epidemiologic
Perspective on Comorbidity of Depression and Anxiety. Child and Adolescent Psychiatric
Clinics of North America, 14(4), 707-726.

Xiao, S., & Kohrman, M. (2008). Anthropology in China's health promotion and tobacco. Lancet,
372(9650), 1617-1618.

Zhang, L., Wang, W., Zhao, Q., & Vartiainen, E. (2000). Psychosocial predictors of smoking
among secondary school students in Henan, China. Health Educ. Res., 15(4), 415-422.

Zheng, P., Guo, F., Chen, Y., Fu, Y., Ye, T., & Fu, H. (2007). A randomized controlled trial of group
intervention based on social cognitive theory for smoking cessation in China. J
Epidemiol, 17(5), 147-155.

Zhu, B.-P., Liu, M., Shelton, D., Liu, S., & et al. (1996). Cigarette smoking and its risk factors
among elementary school students in Beijing. American Journal of Public Health, 86(3),
368-375.

Abstract (if available)

Abstract The current studies provide evidence of the psychosocial processes involved in generating risk for smoking behaviors among those who exhibit high levels of depression symptoms. Study 1 examined the relationship between depressive symptoms, smoking social influences, and smoking behaviors among a sample of adolescents. Study 1 results supported the hypothesis that the relationship between depression symptoms and smoking behaviors were at least partially mediated by both pro-smoking social norm beliefs and perceived friend smoking prevalence. There was no evidence that depression symptoms moderated the relationship between social influences and smoking. Study 2 examined whether changes in social influence cognitions (via a social influences based smoking prevention program) would affect smoking behaviors of students with high levels of depression symptoms more so than those with low or no symptoms. Study 2 results provided evidence that the smoking prevention program changed perceptions of friend smoking prevalence rates among adolescents who had high scores of depression and who have previously experimented with smoking. It was this change in perception that was responsible for the observed reduction in 30-day smoking one year after program implementation. While perceptions of the social environment might differ due to underlying cognitive processes between depressed adolescents and non-depressed adolescents, an important question to ask is whether social competencies may be a better predictor of social influence factors related to risk behaviors like smoking. Study 3 explored the relationships between depression symptoms, social competence (sociability and/or deficits), and smoking-related psychosocial perceptions. Results of study 3 were consistent with the findings of studies 1 and 2 in that depression symptoms were associated with higher perceived friend prevalence.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Social network influences on depressive symptoms among Chinese adolescents

PDF

Cultural risk and protective factors for tobacco use behaviors and depressive symptoms among American Indian adolescents in California

PDF

A network analysis of online and offline social influence processes in relation to adolescent smoking and alcohol use

PDF

Adolescent social networks, smoking, and loneliness

PDF

Anxiety symptoms and nicotine use among adolescents and young adults

PDF

Using observed peer discussions to understand adolescent depressive symptoms and interpersonal interactions

PDF

The link between maternal depression and adolescent daughters' risk behavior: the mediating and moderating role of family

PDF

Social self-control and adolescent substance use

PDF

Factors and correlates of sexual behaviors among young adults from continuation high schools

PDF

Genetic variants and smoking progression in Chinese adolescents

PDF

Depression severity, self-care behaviors, and self-reported diabetes symptoms and daily functioning among low-income patients receiving depression care

PDF

Untangling the developmental relations between depression and externalizing behavior among maltreated adolescents

PDF

Heterogeneity among adolescents with elevated depressive symptoms: distinct patterns of social and academic attributes

PDF

Cognitive attributions for smoking and their roles on subsequent smoking progression and regression

PDF

Sociocultural stress, coping and substance use among Hispanic/Latino adolescents

PDF

The effects of mindfulness on adolescent cigarette smoking: Measurement, mechanisms, and theory

PDF

Contextualizing social network influences on substance use among high risk adolescents

PDF

Risk factors associated with smoking initiation among Chinese adolescents: a matched case-control study

PDF

The effect of an educational fotonovela on the prevention of secondhand smoke

PDF

The influence of contextual factors on the processes of adoption and implementation of evidence-based substance use prevention and tobacco cessation programs in schools

Asset Metadata

Creator Sakuma, Kari-Lyn Kobayakawa (author)

Core Title The role of depression symptoms on social information processing and tobacco use among adolescents

School Keck School of Medicine

Degree Doctor of Philosophy

Degree Program Preventive Medicine (Health Behavior)

Publication Date 08/04/2009

Defense Date 06/04/2009

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag Adolescent,Depression,intervention,OAI-PMH Harvest,perception,smoking,social norms

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Azen, Stanley Paul (committee chair), Johnson, Carl Anderson (committee chair), Palinkas, Lawrence A. (committee member), Stacy, W. Alan (committee member), Unger, Jennifer B. (committee member)

Creator Email karilyn.sakuma@gmail.com,ksakuma@psu.edu

Permanent Link (DOI) https://doi.org/10.25549/usctheses-m2468

Unique identifier UC1485277

Identifier etd-Sakuma-3113 (filename),usctheses-m40 (legacy collection record id),usctheses-c127-173745 (legacy record id),usctheses-m2468 (legacy record id)

Legacy Identifier etd-Sakuma-3113.pdf

Dmrecord 173745

Document Type Dissertation

Rights Sakuma, Kari-Lyn Kobayakawa

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Repository Name Libraries, University of Southern California

Repository Location Los Angeles, California

Repository Email cisadmin@lib.usc.edu