University of Southern California Dissertations and Theses

Genetic variants and smoking progression in Chinese adolescents

(USC Thesis Other)

Genetic variants and smoking progression in Chinese adolescents

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content

GENETIC VARIANTS AND SMOKING PROGRESSION IN CHINESE
ADOLESCENTS

by
Wei Sun

______________________________________________________________
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)

August 2007

Copyright 2007 Wei Sun
ii
DEDICATION

To Angel and Andrew, who always help me achieve my dream.

iii
ACKNOWLEDGMENTS

There are many people to thank for their tremendous support over the years
and here only a few of them. I would especially like to thank my advisors, Dr. Andy
Johnson and Dr. David Conti, for their copious amounts of support, advice, and
feedback and suggestions, any time, with encouragement. I would also like to thank
Dr. Jennifer Unger, Dr. Jean Richardson, and Dr. Jean Shih for providing excellence
guidance all along.
Special thanks to all of IPR staff, especially Marny Barovich, Jolanda Lisath,
and Michelle Orsillo. Their friendly faces and acts of kindness have made this journey
worthwhile.
Last, but not least, I would like to thank my parents, my wife, my brother, my
sister in law, my children and nephews, for their countless support, great laughs and
never-ending love.

iv
TABLE OF CONTENTS
DEDICATION ..........................................................................................................................................ii
ACKNOWLEDGMENTS....................................................................................................................... iii
LIST OF TABLES ...................................................................................................................................vi
LIST OF FIGURES...................................................................................................................................x
LIST OF ABBREVIATIONS .................................................................................................................xii
ABSTRACT .......................................................................................................................................... xiii
CHAPTER 1: INTRODUCTION..............................................................................................................1
1.1 TOBACCO USE ..................................................................................................................................2
1.2 GENETIC INFLUENCE ........................................................................................................................4
1.3 ENVIRONMENTAL INFLUENCE ........................................................................................................11
1.4. THE ORIGINAL STUDIES ................................................................................................................21
1.4.1 Wuhan Smoking Prevention Trial ..........................................................................................22
1.4.2 Genetic Variants and Tobacco Use in Chinese Adolescents Study ........................................25
1.4.3 Important Measures ................................................................................................................26
1.5 THEORETICAL MODEL.............................................................................................................31
1.5.1 Major Causal Models in Smoking Prevention........................................................................31
1.5.2 Theoretical Model Developed for this Study..........................................................................38
1.6 SIGNIFICANCE.................................................................................................................................40
1.7 PRIMARY GOALS ............................................................................................................................42
CHAPTER 2: RESEARCH STUDY 1 - ASSOCIATION BETWEEN GENETIC
POLYMORPHISMS OF MAO-A, DISPOSITIONAL ATTRIBUTES AND SMOKING
BEHAVIOR ............................................................................................................................................43
2.1 HYPOTHESIS ...................................................................................................................................44
2.2 STATISTICS METHOD......................................................................................................................47
2.3 RESULTS.........................................................................................................................................48
CHAPTER 3: RESEARCH STUDY 2 - GENETIC AND ENVIRONMENTAL RISK
FACTORS ON SMOKING PROGRESSION.........................................................................................63
3.1 HYPOTHESIS ...................................................................................................................................63
3.2 STATISTICS METHOD......................................................................................................................64
3.3 RESULTS.........................................................................................................................................67
CHAPTER 4: RESEARCH STUDY 3 - MEDIATED MODERATION MODEL FOR
GENETIC AND ENVIRONMENTAL INFLUENCES ON ADOLESCENT SMOKING
PROGRESSION......................................................................................................................................79
4.1 HYPOTHESIS ...................................................................................................................................80
4.2 STATISTICS METHOD......................................................................................................................81
4.3 RESULTS.........................................................................................................................................86
CHAPTER 5: CONCLUSION ................................................................................................................99
5.1 CONCLUSION FROM THREE STUDIES ..............................................................................................99
5.2 STRENGTHS ..................................................................................................................................101
5.3 LIMITATIONS ................................................................................................................................102
v

BIBLIOGRAPHY .................................................................................................................................106
APPENDIX ...........................................................................................................................................120
vi
LIST OF TABLES
Table 1: Demographic Characteristics at Baseline (N=2661) .......................................... 24
Table 2: Percentages of Subjects in Different Smoking Stages Across Waves................ 24
Table 3: Basic Characteristics of Subjects in the Analysis Sample.................................. 49
Table 4: Allele Size by Gender and Program Condition .................................................. 51
Table 5: Descriptive Statistics of Major Variables Related to Smoking Behavior by
Allele Size (Boys) ...................................................................................................... 52
Table 6: Effects of Allele Size, Program, and Their Interaction on Smoking
Behavior (Boys) ......................................................................................................... 53
Table 7: Descriptive Statistics of Major Variables Related to Smoking Behavior by
Genotype (Girls)......................................................................................................... 55
Table 8: Effect of Genotype, Program, and Their Interaction on Smoking Behavior
(Girls) .........................................................................................................................56
Table 9: Descriptive Statistics of Major Variables Related to Dispositional
Attributes by Allele Size (Boys) ................................................................................ 57
Table 10: Descriptive Statistics of Major Variables Related to Dispositional
Attributes by Genotype (Girl) .................................................................................... 58
Table 11: Effect of Genotype, Program, and Their Interaction on Dispositional
Attributes (Girls) ........................................................................................................ 59
Table 12: Descriptive Statistics on the Adjusted Covariates by Gender and Genetic
Factor..........................................................................................................................67
Table 13: Major Results of Multilevel Logistic Regression Model for the Onset of
Cigarette Use by Allele Size (Boys) .......................................................................... 68
Table 14: Major Results of Multilevel Logistic Regression Model for the Onset of
Cigarette Use (Boys) .................................................................................................. 70
Table 15: Major Results of Multilevel Logistic Regression Model for the Onset of
Cigarette Use by Genotype (Girls)............................................................................. 71
Table 16: Major Results of Multilevel Logistic Regression Model for the Onset of
Cigarette Use (Girls) .................................................................................................. 71

vii
Table 17: Major Results of Multilevel Random Effects Models for Life-time Ever
Smokers at Baseline by Allele Size (Boys)................................................................ 72
Table 18: Major Results of Multilevel Random Effects Models among Life-time
Ever Smokers at Baseline (Boys)............................................................................... 75
Table 19: Major Results of Multilevel Random Effect Model among Life-time
Ever Smokers (Girls).................................................................................................. 76
Table 20: Descriptive Statistics of Time-varying Variables across Waves (Boys).......... 86
Table 21: Descriptive Statistics of Time-varying Variables across Waves (Girl)............ 87
Table 22: Major Estimated Parameters from the Model with Refusal Self-efficacy
as Mediator by Gender and Genetic Factor................................................................ 89
Table 23: Major Estimated Parameters from the Model with Refusal Self-efficacy
as Mediator by Gender ............................................................................................... 95
Table 24: Results of Multilevel Logistic Regression Model for the Onset of
Cigarette Use by Allele Size (Boys) ........................................................................ 120
Table 25: Results of Multilevel Logistic Regression Model for the Onset of
Cigarette Use (Boys) ................................................................................................ 121
Table 26: Results of Multilevel Logistic Regression Model for the Onset of
Cigarette Use by Genotype (Girls)........................................................................... 122
Table 27: Results of Multilevel Logistic Regression Model for the Onset of
Cigarette Use (Girls) ................................................................................................ 123
Table 28: Results of Random Effects Models for Life-time Ever Smokers at
Baseline by Allele Size (Boys)................................................................................. 124
Table 29: Results of Random Effects Models for Life-time Ever Smokers at
Baseline (Boys) ........................................................................................................ 126
Table 30: Results of Random Effects Models for Life-time Ever Smokers at
Baseline (Girls) ........................................................................................................ 128
Table 31: Estimated Parameters from the Model with Refusal Self-Efficacy as
Mediator by Gender and Genetic Factor.................................................................. 130
Table 32: Estimated Parameters from the Model with Social Norm as Mediator by
Gender and Genetic Factor....................................................................................... 132
Table 33: Estimated Parameters from the Model with Attitude to Cigarette Use as
Mediator by Gender and Genetic Factor.................................................................. 136

viii
Table 34: Estimated Parameters from the Model with Refusal Self-efficacy as
Mediator by Gender ................................................................................................. 139
Table 35: Estimated Parameters from the Model with Social Norm as Mediator by
Gender ...................................................................................................................... 142
Table 36: Estimated Parameters from the Model with Attitude to Cigarette Use as
Mediator by Gender ................................................................................................. 145

ix
LIST OF FIGURES

Figure 1: Bandura's Triadic Model of Reciprocal Determinism ..................................32
Figure 2: Social Influence Theory with Dispositional Attributes as Moderators.........33
Figure 3: Theoretical Stress/Coping Model .................................................................35
Figure 4: Enhanced Theory of Planned Behavior ........................................................36
Figure 5: Conceptual Model for the Analysis of Genetic and Environmental
Factors on Adolescent Smoking Progression ........................................................40
Figure 6: # of Cigarette Use Per Day During Last Month by Prevention Program
and Allele Size (Boys) ...........................................................................................54
Figure 7: Increase of Cigarette Use Per Day During Last Month by Prevention
Program and Allele Size (Boys) ............................................................................54
Figure 8: Interaction Effect Between Genotype and Program Condition among
Females on Depression at Baseline. ......................................................................60
Figure 9: Interaction Effect Between Genotype and Program Condition among
Females on the Change of Hostility/family Conflict.............................................61
Figure 10: # of Cigarette Use per Day During last 30 Days for Boys with 4-
repeats. ...................................................................................................................74
Figure 11: LGCM on Smoking Progression with a Time-Varying Mediator ..............85
Figure 12: Growth Curve of Cigarette Monthly Use across Prevention Program
and Allele Size (Boys) ...........................................................................................88
Figure 13: Growth Curve of Cigarette Monthly Use across Prevention Program
and Allele Size (Girls) ...........................................................................................91
Figure 14: Growth Curve of Refusal Self-efficacy across Waves by Prevention
Program and Genetic Type (Boys with 3 or 4 repeats and Girls with 3/3, 3/4
or 4/4 genotype) .....................................................................................................92
Figure 15: Growth Curve of Social Norm across Waves by Prevention Program
and Genetic Type (Boys with 3 or 4 repeats and Girls with 3/3, 3/4 or 4/4
genotype) ...............................................................................................................93

x
Figure 16: Growth Curve of Attitude to Cigarette Use across Waves by
Prevention Program and Genetic Type (Boys with 3 or 4 repeats and Girls
with 3/3, 3/4 or 4/4 genotype)................................................................................94
Figure 17: Growth Curve of Cigarette Monthly Use across Prevention Program
and Allele Size (Boys) ...........................................................................................97
Figure 18: # of Cigarette Use/Day during Last Month at Baseline by Allele Size
(Male)...................................................................................................................148
Figure 19: # of Cigarette Use/Day during Last Month at Baseline by Allele Size
and Program (Male).............................................................................................148
Figure 20: # of Cigarette Use/Day during Last Month at Year 1 Follow-up by
Allele Size (Male)................................................................................................149
Figure 21: # of Cigarette Use/Day during Last Month at Year 1 Follow-up by
Allele Size and Program (Male) ..........................................................................149
Figure 22: Change of Cigarette Use/Day during Last Month from Baseline to
Year 1 Follow-up by Allele Size (Male) .............................................................150
Figure 23: Change of Cigarette Use/Day during Last Month from Baseline to
Year 1 Follow-up by Allele Size and Program (Male)........................................150
Figure 24: # of Cigarette Use/Day during Last Month at Baseline by Allele Size,
Program, and Depression Score (Male)...............................................................151
Figure 25: # of Cigarette Use/Day during Last Month in Year 1 Follow-up by
Allele Size, Program, and Depression Score (Male) ...........................................151
Figure 26: Change of Cigarette Use/Day during Last Month from Baseline to
Year 1 Follow-up by Allele Size, Program, and Depression Score (Male).........152

xi
LIST OF ABBREVIATIONS

ACU Perceived attitude to Cigarette Use
ANOVA Analysis of variance
CFI Comparative Fit Index
CI Confidence interval
CSCS China Seven City Study
DA Dopamine
DF Degree of freedom
DR Dopamine receptor
GLM Growth Curve Model
LGCM Latent Growth Curve Model
MAO Monoamine Oxidase
MAOA Monoamine Oxidase A
MAOB Monoamine Oxidase B
MDD Major Depressive Disorder
OR Odd Ratio
RSE Perceived refusal self-efficacy
RMSEA Root Mean Square Error of Approximation
RR Relative risk
SD Standard Deviation
SE Standard Error
SES Social-economic status
SEM Structural equation modeling
SN Perceived social norms
SNP Single nucleotide polymorphism
SMK Smoking progression
VNTR Variable number tandem repeat
WSPT Wuhan Smoking Prevention Trial

xii
ABSTRACT

For adolescent smoking progression, understanding the potential risk factors,
including genetic, psychosocial and other environmental factors, is important for
identifying adolescents with high risk, and designing efficient prevention curriculum
for deterring smoking development and nicotine dependence. The three studies of
this dissertation, utilizing the sample from a longitudinal, randomized, school-based
smoking prevention trial with 2661 urban 7
th
grade students at baseline, explore the
patterns and risk factors for progressions of adolescent smoking and the potential
moderating and mediating mechanisms among genetic and environmental risk factors.
Study 1 examined the association between genetic variants and smoking behavior,
depression, and hostility/family conflict. Chinese male adolescents with the 3-repeat
allele of MAOA smoked more numbers of cigarettes per day during last month than
those with 4-repeat allele of MAOA (p=0.019). They had a smaller increase in
cigarette use from baseline to one-year follow-up (p=0.013). We also found that the
females with 3/3 repeat genotype of MAOA had larger increases in hostility scores
than those with 4/4 repeat genotype in the two-year follow-up (p=0.026).
Study 2 investigated the potential risk and protective factors for smoking
progression among adolescents. Regarding the smoking initiation, the boys with 3
repeats is 1.32 times more likely to become a life time ever smokers in the year-one
follow-up. Regarding to the amount of cigarette use, for the boys with 4 repeats, those
in the program group smoked significantly fewer than those in the control group

xiii
(p=0.031). It suggests that a different set of risk/ protective factors exist for those
adolescents with 3- or 4- repeat allele of MAOA.
Study 3 applied Latent growth curve analysis to investigate the mechanism of the
influences of genetic and environmental factors on the smoking progression. In the
model with allele size of MAOA as a risk factor and refusal self-efficacy as a mediator,
the initial status or slope of refusal self-efficacy is significantly associated with initial
status or slope of smoking progression for both boys and girls.
The results from these three studies suggest that repeat polymorphisms of MAOA
is associated with the smoking progression among Chinese adolescent.

1
CHAPTER 1: INTRODUCTION

Tobacco use remains the leading preventable cause of death and disease in the
United States and most of the developed and developing world, including China
(CDC, 2002; Kandel, Chen, Warner, Kessler, & Grant, 1997; Yang et al., 1999). The
majority of people who smoke in the U.S. begin and escalate during adolescence
(Botvin & Botvin, 1992; CDC, 2005; Kessler et al., 1997). Smoking initiation is
occurring at earlier ages with economic development in China, but the potential for
initiation continues into middle and late adulthood, unlike in the U.S.(Chou et al.,
2006; W. Sun et al., 2006; Yang et al., 1999) Twin, adoption, and familial studies
have shown that both genetic and environmental factors contribute for adolescent
smoking progression (Lessov, Swan, Ring, Khroyan, & Lerman, 2004; Li, 2003;
Madden, Pedersen, Kaprio, Koskenvuo, & Martin, 2004; Niu et al., 2000; Osler,
Holst, Prescott, & Sorensen, 2001; Slomkowski, Rende, Novak, Lloyd-Richardson, &
Niaura, 2005; Swan, 1999).
Understanding the related risk factors and different patterns of adolescent
smoking progression is important for identifying adolescents with high risk, and
designing efficient prevention curriculum for preventing smoking progressive
development and nicotine dependence. However, the potential risk factors, including
genetic, psychosocial and cultural factors, for adolescent smoking progressions are
poorly understood, especially their combinatorial effects.
Utilizing a longitudinal sample, this dissertation study conducted extended
secondary data analyses to explore the patterns and risk factors for progressions of

2
adolescent smoking and the potential moderating and mediating mechanisms among
genetic and environmental risk factors. It is hypothesized that genetic factors
responsible for dispositional attributes, such as hostility and depression, act to both
influence an individual's tobacco use progression and to moderate the effectiveness of
interventions of smoking prevention and some other psychosocial risk factors. The
study employed data from the Wuhan Smoking Prevention Trial (WSPT) and its
complementary genetic epidemiologic study. WSPT is a longitudinal, randomized,
school-based smoking prevention trial with 1,337 (50.2%) and 1,324 (49.8%) urban
students in the control and program groups, respectively. The average age at baseline
was 12.5 years. The longitudinal design of this project provides the best way to assess
the intervention program and the related risk factors for adolescent smoking
progression. The original specific aims of this project were primarily to test the effects
of policy interventions, changed social norms, and social variables on reducing
adolescent tobacco use across cultural settings. WSPT was complemented with the
investigation of polymorphisms within several candidate genes using single nucleotide
polymorphisms (SNPs) selected from public SNP databases. The original WSPT and
its complementary project have obtained abundant genetic and environmental data.
They are perfect resource for the analyses of patterns and genetic and environmental
risk factors of smoking progression among adolescents.
1.1 Tobacco Use
Cigarette smoking is the most prevalent form of substance use (Kandel, Chen,
Warner, Kessler, & Grant, 1997) and the leading preventable cause of disease and

3
death in the United States (CDC, 2002). As a serious public health problem itself,
smoking is an important risk factor for many types of diseases, such as pulmonary
disease (Mannino, 2003), lung and other cancers (Murray, Varnell, & Blitstein, 2004),
coronary heart disease, stroke (Critchley & Capewell, 2004; McGinnis & Foege,
1993), and oral diseases (Patel & Homnick, 2000).
Despite increased public knowledge of the negative health effects of smoking,
many adolescents still experiment with tobacco use, and some progress to regular
cigarette use. Studies have shown that early initiation of smoking, even occasionally,
can lead to a rapid escalation to regular smoking levels (Botvin & Botvin, 1992;
Colder et al., 2001; Everett et al., 1999; Fergusson & Horwood, 1995). Most adult
smokers start smoking before 18 years of age (Kessler et al., 1997) , with the critical
period for the development of experimentation with smoking and subsequent regular
smoking spanning from early to late adolescence (Botvin & Botvin, 1992). Each day,
nearly 3,900 young people between the ages of 12 and 17 years initiate cigarette
smoking in the United States (CDC, 2005). In this age group, each day estimated
1,500 young people progress to daily cigarette smokers in this country (SAMHSA,
2005). Nationwide, the prevalence rates of tobacco use among middle and high school
students are 8% and 22% respectively, with estimates slightly higher for females than
males (CDC, 2005). Tobacco use among White, Hispanic, and Asian adolescents is
increasing more rapidly than that among Black adolescents (Pierce et al., 1998;
USDHHS, 1994).
According to the results from a Chinese population-based national survey
administered in 1996 in 30 provinces of China, 63% of men and 3.8% of women were

4
current smokers, and 34.1% of Chinese, aged 15–59 years, smoked at least one
cigarette per day; this rate had increased by 3.4 percentage points since 1984 when the
first national survey was administered (Yang et al., 1999). The prevalence rates of
experimenting were 47.8% for boys and 12.8% for girls. China Seven City Study
(CSCS) reported that the overall prevalence rates of ever smokers and past-30-day
smokers in middle and high schools were 24.3% and 9.0%. And smoking was much
more prevalent in boys than in girls (Weiss, Spruijt-Metz, Palmer, Chou, & Johnson,
2006). About 20% of boy and 10% of girl never smokers at baseline became life-time
ever smokers at year-1 follow-up, respectively (W. Sun et al., 2006). Therefore,
smoking progression in China adolescents is a significant concern, and the
identification of factors that influence progression is critical for smoking prevention
and intervention efforts. Identifying different patterns of smoking progression is
important because such heterogeneity has important implications for both prevention
and research. Additionally, various trajectories of smoking may be linked to different
motivational and etiological pathways.
1.2 Genetic Influence
Studies have showed that progression from casual use to nicotine addiction
typically occurs during adolescence (Choi, Pierce, Gilpin, Farkas, & Berry, 1997;
Heishman et al., 1997; Pierce, Choi, Gilpin, Farkas, & Merritt, 1996). However, this
prior research has largely focused on identification of environmental factors,
especially psychosocial factors, of simple transitions without including genetics (B.
Flay, 1993; Leventhal & Cleary, 1980; Patil et al., 2001; USDHHS, 1994).

5
Recent studies of brain development during puberty have documented a
continuous increase in the volume of global and local white matter, as well as
asynchronous age-related decreases in the volume of gray matter in different cortical
regions (Paus, 2005). Additional research found the frontal cortical regions are the
ones most frequently associated with drug addiction (Goldstein & Volkow, 2002).
Findings suggest that addiction is in fact a product of cortically regulated cognitive
and emotional processes, which lead to overvaluing of reinforcers such as nicotine
(Goldstein & Volkow, 2002).
Research has indicated that smoking behavior are at least partially under genetic
control (Lessov, Swan, Ring, Khroyan, & Lerman, 2004; Li, 2003; Swan, 1999).
Many twin studies have found higher concordance rates for smoking and for quitting
smoking among monozygotic compared to dizygotic twins, suggesting that a genetic
predisposition may be involved (Batra, Patkar, Berrettini, Weinstein, & Leone, 2003;
Carmelli, Swan, Robinette, & Fabsitz, 1992; Madden, Pedersen, Kaprio, Koskenvuo,
& Martin, 2004). A substantial heritability of smoking initiation (60% in men, 51% in
women) was found in three large adult twin samples from different countries (Heath &
Martin, 1993). An adoption study further supports a role for a genetic component in
smoking behaviors (Osler, Holst, Prescott, & Sorensen, 2001). It found that adoptees'
status as ever, current, heavy, or former smokers was related to their biologic relatives'
smoking status, which supports the finding in twin studies of a genetic influence on
smoking within the same generation. Moreover, familial studies in addition showed
that risk of nicotine dependence is statistically significantly greater among second
siblings if the first sibling is nicotine dependent, but these studies didn't really separate

6
out genes vs. environment (Bierut et al., 1998; Niu et al., 2000). And a twin-sibling
study showed the main effects of both shared environment and genetics found on
adolescent smoking frequency (Slomkowski, Rende, Novak, Lloyd-Richardson, &
Niaura, 2005).
It is generally accepted that genetic factors associated with personality traits
influence people’s behavior, and personality traits have been linked to smoking
progression. To achieve high levels of novelty and variety, people with the sensation-
seeking temperament tend to take physical and social risks (Gerra et al., 1999). That
is why people who have a higher score on sensation-seeking indices are more likely to
try smoking than those with lower scores (Etter, Pelissolo, Pomerleau, & De Saint-
Hilaire, 2003). Furthermore, it is indicated that dispositional attributes, such as
depression, hostility, novelty-seeking phenotype (part of sensation-seeking), etc., have
a genetic basis (Audrain-McGovern, Lerman, Wileyto, Rodriguez, & Shields, 2004;
Cloninger, 1988; Tsai, Hong, Yu, & Chen, 2004). Decreased MAOA activity has
been linked to enhanced reward and the sensation-seeking temperament (Hallman,
Sakurai, & Oreland, 1990; Johansson, Von Knorring, & Oreland, 1983). MAO
(Monoamine oxidase), an mitochondrial enzyme consisting of two isoforms (MAOA
and MAOB), is found in neuronal and non-neuronal cells (Abell, 1987; Shih &
Thompson, 1999). Because of its main function, the breakdown of neurotransmitters,
it is a key enzyme to regulate the serotonin and dopamine levels in the brain. With
regard to functional polymorphisms of MAOA gene, a variable number of tandem
repeat (VNTR), located at 1200 base pairs up-stream of the transcription site, involves
transcriptional activity in transfected cells (W. K. Nilsson, 2006). Males are

7
hemizygous carriers of one MAOA allele, whereas women carry two alleles.
Therefore, only women can be heterozygote.
Like other substance use, after initiation, smoking invokes reward pathways,
which promote further consumption of nicotine (Chao & Nestler, 2004; Koob &
Nestler, 1997). Specifically, nicotine triggers the reward pathway through activation
of the acetylcholine receptors on dopaminergic neurons in the ventral tegmental region
of the brain (Comings, 1998). The release of dopamine (DA) reinforces the behavior
and contributes to nicotine addiction. This has been observed in both animal studies
(Corrigall, Franklin, Coen, & Clarke, 1992) and human studies (Laviolette & van der
Kooy, 2003). It is suggested that defects in various combinations of the genes for
these neurotransmitters may lead the individual to search out “unnatural rewards” such
as smoking, alcohol, etc. (Comings & Blum, 2000).
All three stages of tobacco addiction (initiation, persistence and quitting) appear to
be under genetic control (True et al., 1997) with the different stages being controlled
by different genetic mechanisms, although there may be some overlaps (Kendler et al.,
1999). Heritability estimates for smoking initiation range from 46-84%, while those
for continuation range from 53-70%. For example, Koopmans’ research showed that
smoking initiation was influenced by genetic factors (39%) and shared environmental
influences (54%). Once smoking is initiated genetic factors determine to a large extent
(86%) the smoking quantity (Koopmans, Slutske, Heath, Neale, & Boomsma, 1999).
Numerous population-based association studies have been performed to examine
the effects of a number of candidate genes, such as dopamine receptor (DR) and
transporter genes (Batra, Patkar, Berrettini, Weinstein, & Leone, 2003; Duggirala,

8
Almasy, & Blangero, 1999; Erblich, Lerman, Self, Diaz, & Bovbjerg, 2004), serotonin
transporter and nicotinic acetylcholine receptor (Ishikawa et al., 1999; Lerman et al.,
1998), cytochrome P450 (Batra, Patkar, Berrettini, Weinstein, & Leone, 2003; Walton,
Johnstone, Munafo, Neville, & Griffiths, 2001), on smoking behavior. It showed that
multiple genes might be associated with a specific smoking stage, and different
smoking stages might be associated with the same candidate genes.
Marcus reviewed the relationship between several genes (DRD2, DAT, 5HTT, and
CYP2A6 genes) and smoking behavior (Munafo, Clark, Johnstone, Murphy, &
Walton, 2004). Evidence indicated effects of the DRD2 Taq1A polymorphism and
smoking initiation, the 5HTT LPR and CYP2A6 reduced-activity polymorphisms and
smoking cessation, and the DRD2 Taq1A and CYP2A6 reduced-activity
polymorphisms and cigarette consumption. This study suggests implications for the
design of future studies, such as the development of more specific phenotypes to
increase the genetic signal in candidate gene studies. Tyndale presented that inhibiting
CYP2A6 in vivo decreased nicotine metabolism and smoking behavior (Tyndale &
Sellers, 2001).
Among the candidate genes, dopaminergic genes have been studied for heritable
influences on cigarette smoking in many studies. Findings demonstrate associations
between allele 9 of a dopamine transporter gene polymorphism (SLC6A3-9) and lack
of smoking, late initiation of smoking, and length of quitting attempts (Lerman et al.,
1999). A significant association between SLC6A3-9 and smoking status was
confirmed and was due to an effect on cessation rather than initiation (Sabol et al.,
1999). The SLC6A3-9 polymorphism was also associated with low scores for novelty

9
seeking, which was the most significant personality correlate of smoking cessation.
Thus, it is reasonable to hypothesize that individuals carrying the SLC6A3-9
polymorphism have altered dopamine transmission, which reduces their need for
novelty and reward by external stimuli, including cigarettes.
Of the candidate dopaminergic genes, the D2 dopamine receptor (DRD2) gene
located on chromosome 11 has been extensively studied. DRD2 gene is reported as
one of a multi-factorial set of risk factors associated with smoking behavior (Comings
et al., 1996). For the adolescents with a previous smoking experience, it was reported
that the likelihood of progressing to a higher level of smoking by the 11th grade
increased almost twofold with each additional DRD2 A1 allele. The likelihood of
smoking progression with each additional A1 allele was more pronounced among
adolescents with substantial depression symptoms (Audrain-McGovern, Lerman,
Wileyto, Rodriguez, & Shields, 2004).
While a general genetic influence has been suggested with previous studies, it is
still not clear exactly how specific gene polymorphisms relate to dispositional
attributes or nicotine metabolism, especially in connection to smoking progression.
Research on single nucleotide polymorphisms (SNPs) in candidate genes within
suspected etiologic pathways are thus needed. To test the genetic association with
smoking progression, two approaches can be applied: 1) to type all SNPs in a
candidate gene; 2) to type the whole genome. The first approach identifies all SNPs in
the gene of interest according to a public database (i.e. Hapmap), pick several SNPs
spaced across the entire gene of interest, and type these SNPs and evaluate the
association of each SNPs and constructed haplotypes with smoking progression. The

10
second approach spans the whole genome. No hypothesis about which genes might be
involved is necessary.
Based on our hypothesis, MAOA is responsible for dispositional attributes, and
acts to both influence an individual's tobacco use progression and to moderate the
effectiveness of interventions of smoking prevention and some other psychosocial risk
factors. MAOA, an enzyme that degrades amine neurotransmitters, such as dopamine,
norepinephrine, and serotonin, provides the major enzymatic clearing step for
serotonin and norepinephrine (Meyer-Lindenberg et al., 2006). The VNTR
polymorphism in the MAOA promoter has been reported to have multiple alleles.
With a large sample of healthy volunteers, a study found that the polymorphism of
MAOA consists of a 30-bp repeated sequence present in 2, 3, 3.5, 4, or 5 copies
(Meyer-Lindenberg et al., 2006). Only the variants with 3 or 4 repeats are common in
different ethnic populations (Sabol, Hu, & Hamer, 1998). Research demonstrated that
the deficiency in MAOA is associated with aggressive behavior in both animal and
human studies (Cases et al., 1995). Platelet MAO activity is associated with certain
personality traits, with low activity linked to traits such as impulsiveness, sensation-
seeking and avoidance of monotony, all possible expressions of low central
serotonergic activity (Hallman, Sakurai, & Oreland, 1990; Johansson, Von Knorring,
& Oreland, 1983; Oreland, Hallman, & Damberg, 2004). I conducted the analyses
based on the available environmental data and the genetic data regarding to repeated
polymorphisms of MAOA gene.
A potentially promising area of research regarding adolescent smoking prevention
is the investigation of the underlying mechanism of smoking by examining the

11
relationship between smoking behavior and gene polymorphisms from MAOA.
Specifically, polymorphisms within the serotonin and dopamine systems may impact
the brain development among adolescents and may, in turn, alter various dispositional
attributes, such as anxiety, aggressiveness, and depression. This is important because
these dispositional attributes have been found to be associated with smoking initiation,
progression, and response to certain prevention strategies (Dierker, Avenevoli, Stolar,
& Merikangas, 2002; Glassman et al., 1990; Johnson et al., 2000; Patton et al., 1998;
Ritt-Olson et al., 2005).

1.3 Environmental Influence
Environmental factors related to smoking progression are knowledge, intentions,
attitudes, health-related behavior, personality characteristics and school-related
variables, smoking behavior of parents, siblings, peers and significant adults, family
characteristics, social support, acculturation, and socio-economic status, etc.
(Hawkins, Catalano, & Miller, 1992; Tyas & Pederson, 1998; USDHHS, 1994) The
risk and protective factors for adolescent smoking in individual level basically can be
separated into two sub-groups: 1) social-demographics factors, which are related to
individual's characteristics or dispositions; and 2) the individual psychosocial factors,
which are individual psychosocial correlates of smoking, including the personal and
behavioral factors. The contextual risk and protective factors for adolescent smoking
include the factors from broader environment. They are: 1) social contextual factors
(parental smoking, parental attitudes, sibling smoking, family environment, peer
smoking, peer attitudes and norms, attachment to family and friends, etc.); 2)

12
environmental contextual factors (availability of tobacco, advertising and tobacco-
related policies, television and movies, etc.); and 3) cultural contextual factors
(advertising and tobacco-related policies, television and movies, etc.). Research
should not focus on contextual factors only without considering the influences from
the risk/protective factors in the individual level. And the cultural contextual factors,
in general, function as moderating factors that influence the smoking progression for
adolescents.
Initiation and prevalence of smoking among adolescents typically rise with
increasing age (Turner, Mermelstein, & Flay, 2004). For example, the age of smoking
initiation was significantly related to current frequent smoking, daily smoking, and
whether students had ever smoked daily, and a younger age of smoking initiation was
associated with smoking more cigarettes per day than was initiating at an older age
(Everett et al., 1999). Grade reflects the development of adolescents as well. Higher
school grade was found to be more likely to differentiate regular smoking from earlier
smoking stages (Lloyd-Richardson, Papandonatos, Kazura, Stanton, & Niaura, 2002).
Historically, the prevalence of smoking was higher among men than women in
most cultures. However, there is geographical and cultural variation in the pattern of
gender differences in smoking. Motives for smoking tend to differ by gender. Among
adolescent girls, body image and eating issues are predictors of smoking initiation
(Stice & Shaw, 2003), but among boys aggression and conduct disorders appear to be
fairly consistent predictors of smoking (McMahon, 1999). Even more complexities
arise when other causal links, such as the relationship between depression,

13
delinquency, and smoking present differently for adolescent females than for
adolescent males (Whalen, Jamner, Henker, & Delfino, 2001).
The risk of smoking progression for adolescents with different ethnicity
background is different, but the reasons for this difference are not clear. Research
showed that Adolescents from different minority groups (African American, Hispanic
and Asian) had decreased odds of transitioning to a higher smoking stage than White
students. (Harrell, Bangdiwala, Deng, Webb, & Bradley, 1998; Lloyd-Richardson,
Papandonatos, Kazura, Stanton, & Niaura, 2002). Among Asian American
adolescents, the risk of early smoking initiation is about a third of that of Caucasians.
However, the risk among Asian Americans continues to increase throughout
adolescence, while the same risk among Caucasians and African Americans plateaus
around 14-15 years of age. Significant differences in the levels and patterns of
smoking initiation among Asian American subgroups were observed as well (Chen &
Unger, 1999).
Some other social-demographics factors are associated with smoking behavior for
adolescents as well, such as the parental socioeconomic status (SES) and family
structure. Research showed that higher levels of parental socioeconomic variables are
often inversely related to smoking status in adolescents. SES probably overlaps with
parental educational level, pocket money, or residence status, etc. Family structure
has been found to be associated with adolescent smoking in many studies over the
previous decade (Richardson, Radziszewska, Dent, & Flay, 1993; Shakib et al., 2003;
Wei Sun et al., In Press; Tyas & Pederson, 1998). Research evidence leads to the

14
conclusion that intact, two-parent families are protective against smoking for
adolescents.
Adolescents’ school performance has been found to be consistently related to
smoking status. Those students who do well in school and/or are committed to school
are less likely to smoke than those who have poor educational performance (Choi,
Pierce, Gilpin, Farkas, & Berry, 1997; Wei Sun et al., In Press; Yorulmaz, Akturk,
Dagdeviren, & Dalkilic, 2002).
Lifestyle behaviors tend to occur together in adolescents like in adults (Hawkins,
Catalano, & Miller, 1992; Tyas & Pederson, 1998; USDHHS, 1994). Research
showed that alcohol and other substance use increased the risk of smoking among
adolescents, whereas participation in sport activities or other physical exercise
consistently protected against smoking (Vink, Willemsen, Engels, & Boomsma, 2003)
. Not following a healthy lifestyle can be considered a type of risk taking. Research
reported that propensity toward rebelliousness and risk taking in childhood predicts
adolescent smoking (Burt, Dinh, Peterson, & Sarason, 2000). Risk-taking is related to
novelty-seeking (a type of personality trait), which is associated with smoking
behavior (Gerra et al., 1999; Josendal, Aaro, & Bergh, 1998). Those adolescents who
demonstrate impulsive and risk-taking behavior and an increased need for stimulation
are more receptive to tobacco advertising and are at high risk for smoking initiation
(Audrain-McGovern et al., 2003).
Attitudes to smoking/smokers and knowledge of health effects of smoking are
associated with smoking progression as well (Tyas & Pederson, 1998; Unger et al.,
1999), although they may not be as important as the other factors. Pederson and her

15
colleagues reported that knowledge and attitudes towards smoking were statistically
significantly related to smoking status, with smokers being less negative than the other
groups (Pederson, Koval, McGrady, & Tyas, 1998). But, Charlton and Blair only
found the relationship between positive attitudes to smoking and initiation of smoking
to be significant for females.
The measure of susceptibility to smoking could be an effective tool for identifying
adolescents at the increased risk of experimenting with cigarettes (Unger, Johnson,
Stoddard, Nezami, & Chou, 1997). It was reported that, compared with regular
smokers, never smokers were 15% more likely to report greater non-acceptability of
smoking (Fagan, Eisenberg, Stoddard, Frazier, & Sorensen, 2001). Never smokers
perceived more negative social norms and less pressure regarding smoking as well as
more disadvantages, long-term physical consequences (Ausems, Mesters, van
Breukelen, & De Vries, 2003).
The relationship between perceived health status and smoking status was
significant (Prokhorov et al., 2003; Rius, Fernandez, Schiaffino, Borras, & Rodriguez-
Artalejo, 2004). Current smokers considered themselves less healthy than did
experimental smokers and those who had never smoked. Dissatisfaction with weight
was also related to smoking; individuals who believe that they were too heavy were
more likely to be involved with smoking.
Vulnerability to smoking for adolescents may be due to psychiatric comorbidities
(i.e., anxiety, depression, alcohol or substance use). Adolescents with clinically
significant depressive symptoms were more likely to smoke than those without these
symptoms (Tercyak, Goldman, Smith, & Audrain, 2002). Starting to smoke was more

16
than nine times as likely among students drinking alcohol at least twice a month than it
was among abstinent students (Lloyd-Richardson, Papandonatos, Kazura, Stanton, &
Niaura, 2002). But alcohol use was most influential on earlier smoking (Lloyd-
Richardson, Papandonatos, Kazura, Stanton, & Niaura, 2002). Research showed that
depressive symptoms were significantly associated with intention to smoke, and the
association between depressive symptoms and adolescent smoking can generalize
across diverse ethnic groups (Nezami et al., 2005).
A deficiency in competence (poor decision-making skills and low personal
efficacy) has been linked to acquiring beliefs in the perceived benefits of smoking and
these perceived benefits are then related to subsequent smoking (Epstein, Griffin, &
Botvin, 2000). Students who were more likely to initiate smoking were more
rebellious (Choi, Harris, Okuyemi, & Ahluwalia, 2003).
Smoking behavior of friends and siblings was strongly associated with smoking
(Grenard et al., 2006; Kobus, 2003; Wei Sun et al., In Press; Unger et al., 2002). Flay
found friends' smoking has a stronger effect on adolescents' smoking behavior,
particularly on initiation, but the pathways of friends' influences among different
ethnic groups might be different (B. R. Flay et al., 1994). Students who were more
likely to progress from experimentation to a higher level of smoking thought their
peers approved of smoking (Choi, Harris, Okuyemi, & Ahluwalia, 2003). Pederson
reported that the progress from never smoker to experimental stage was predicted by
peer modeling, and the progress from experimental stage to regular use/nicotine
dependence was predicted by modeling of parental smoking (Pedersen & Lavik,
1991). Using national data from a sample of 20,747 adolescents, Lloyd-Richardson

17
and her colleagues found that peer smoking was by far the strongest predictor of
smoking progression (Lloyd-Richardson, Papandonatos, Kazura, Stanton, & Niaura,
2002). The students who had at least two friends who smoked were more than six
times as likely to transition from experimental smokers to intermittent smokers, and
the adolescents with peer smoking were almost ten times as likely to transition from
intermittent smokers to regular/established smokers (those that smoked daily for the
past 30 days). Another study confirmed the results as well that childhood close friends
who smoke influence not only initiation but also escalation of adolescents' smoking
(Bricker et al., 2005).
Parental smoking increased the odds of an adolescent being in a higher smoking
stage by 26 percent across each transition point for both male and female (Lloyd-
Richardson, Papandonatos, Kazura, Stanton, & Niaura, 2002). However, ethnic
differences in parenting characteristics (parental smoking status, adolescents'
perceptions of parent-child communication, and parental monitoring) and adolescent
smoking might exist (Shakib et al., 2003). The most important long-term predictor of
daily smoking among young adults is mother's smoking (Oygard, Klepp, Tell, &
Vellar, 1995).
Connectedness to school and family were protective of smoking initiation (Lloyd-
Richardson, Papandonatos, Kazura, Stanton, & Niaura, 2002). Strong family ties were
mildly protective of adolescents transitioning to a higher smoking stage. But, those
students who had a poor connection to school were more likely to try smoking and
progress to become regular smokers and those in the higher grades were also more
likely to progress from trying smoking to regular/established smoking. These findings

18
are matched with the Social Development Model initiated by Hawkins and his
colleagues (Catalano, Haggerty, Oesterle, Fleming, & Hawkins, 2004). In the Social
Development Model, bonding is employed as composed of attachment and
commitment to a socializing unit. Involvement is seen as part of a socialization process
that leads to bonding, while beliefs in the social unit’s values are seen as a
consequence of bonding and as a mediator of the effect of bonding on behavioral
outcomes, including smoking progression. This model hypothesizes that children must
learn patterns of behavior, whether prosocial or antisocial, from their social
environment.
Adolescents with high exposure to cigarette advertising were significantly more
likely to be smokers, according to several measures of smoking behavior, than were
those with low exposure to cigarette advertising (Botvin, Goldberg, Botvin, &
Dusenbury, 1993). Youths reporting exposure to cigarette advertising were
significantly more likely to be current smokers and to expect to be smokers at 20 years
of age, after control for important social influence predictors (Braverman & Aaro,
2004; Sargent et al., 2002; Sargent, Dalton, Heatherton, & Beach, 2003).
Attitudes to tobacco advertising are related to some home and school factors, but
most significantly to tobacco and alcohol consumption, to amount of time at home
without adults, and to peer influence. The impact of cigarette ads is different between
two groups, the "vitalists" (those who liked the advertising because it was
entertaining and full of life) and the "credulous" (those who liked the advertising
because it was sincere and credible) (Santana, Gonzalez, Pinilla, Calvo, & Barber,
2003). Longitudinal studies consistently suggest that exposure to tobacco advertising

19
and promotion is associated with the likelihood that adolescents will start to smoke.
Based on the strength of this association, the consistency of findings across numerous
observational studies, temporality of exposure and smoking behaviors observed, as
well as the theoretical plausibility regarding the impact of advertising, we can
conclude that tobacco advertising and promotion increases the likelihood that
adolescents will start to smoke (Lovato, Linn, Stead, & Best, 2003). Specifically, The
tobacco advertising campaigns targeting women, which were launched in 1967, were
associated with a major increase in smoking uptake that was specific to females
younger than the legal age for purchasing cigarettes (Pierce, Lee, & Gilpin, 1994). A
study demonstrated that adolescents who experimented with cigarettes were better able
to recognize advertised products than those who had not, a selective exposure effect.
Conversely, subjects who were better at recognizing advertised brands were more
likely to have experimented with cigarettes, an effect due to their exposure to the
cigarette advertising (Klitzner, Gruenewald, & Bamberger, 1991; Unger, Johnson, &
Rohrbach, 1995)..
Mass media interventions targeting adolescents can reduce substance use behavior
for those segments (Worden et al., 1996). And another study showed that the radio
campaign had a modest influence on the expected consequences of smoking and friend
approval of smoking, the more expensive campaigns involving television was not
more effective than those with radio alone (Bauman, LaPrelle, Brown, Koch, &
Padgett, 1991; Rohrbach et al., 2002). Estimates of the cost-effectiveness ratios of a
mass media campaign in preventing the onset of smoking showed it to be

20
economically attractive and to compare favorably with other preventive and
therapeutic strategies (Secker-Walker, Worden, Holland, Flynn, & Detsky, 1997).
Availability of tobacco and the cost of tobacco products are associated with
smoking progression. It have been found that one of the best predictors of
experimentation with cigarettes was the perception that they were easily available, and
regular smoking appeared to be heavily influenced by cost (Robinson, Klesges,
Zbikowski, & Glaser, 1997). Availability of tobacco to young people is believed to be
an important factor in the onset of tobacco use. Despite legislation that prohibits
selling to minors, they are generally able to acquire cigarettes and other tobacco
products. Research showed that sources of cigarettes shift from social to commercial
with age and that sources of cigarettes for rural youths may be different than for urban
youths (Forster, Wolfson, Murray, Wagenaar, & Claxton, 1997).
Cultural orientation could be associated with smoking status. Tyas summarized
that the reports of equal or higher levels of smoking by females were primarily found
in studies with subjects from countries with a Western cultural orientation: England,
New Zealand and the United States rather than an "Eastern" one with higher smoking
levels among males: China, Japan, and Sri Lanka (Tyas & Pederson, 1998).
Research reported that acculturation was associated with the difference in risks of
smoking onset between Chinese American minors and the white adolescents (Chen,
Unger, & Johnson, 1999). Acculturation is a process in which members of one cultural
group adopt the beliefs and behaviors of another group. It is a multi-dimensional
process involving language, cultural beliefs and values and "structural assimilation"
(Hazuda, Haffner, Stern, & Eifler, 1988) . In general, it is the integration of members

21
of the minority group into the social structure of the majority group. Aspects of the
lifestyle and beliefs of particular cultural groups may affect smoking behavior. A
series of studies of Latino/Hispanic American youth indicated that compared with the
less acculturated Hispanic youth, the more acculturated ones experienced higher levels
of susceptibility to cigarette smoking, the greater risk of early smoking initiation and
the higher prevalence rate of cigarette smoking. Because acculturation and
socioeconomic status may be closely related, research studies must be careful to
distinguish acculturation effects from those that can be linked to income and
education.

Some other cultural factors are associated with adolescent smoking behavior as
well. Parenting styles and respect for elders are two important cultural factors
influencing smoking behavior, highlighted in the Surgeon Generals Report. Research
suggested that authoritarian parental messages protect African American youth against
higher rates of smoking (Koepke, Flay, & Johnson, 1990). Styling is often associated
with ethnicity. It can act as both a protective and risk factor for smoking.
Mermelstein found that Black girls in comparison to White girls were far more likely
to think that ‘not smoking’ enhanced their self image (Mermelstein, 1999) .
1.4. The Original Studies
The research utilized the data from the original WSPT and its complementary
genetic Epidemiologic study. The original specific aims of the Project WSPT were
primarily to test the direct, indirect, and contextual effects of policy interventions,

22
changed social norms, and social variables on reducing adolescent tobacco use in
Chinese adolescents. It was initiated jointly by the Center for Disease Control and
Prevention in Wuhan, China and the Department of Preventive Medicine at the
University of Southern California. Its complementary study investigated the functional
polymorphisms with a haplotype-based approach including characterization of the
underlying haplotype variation within each candidate gene using SNPs selected from
public databases.
1.4.1 Wuhan Smoking Prevention Trial
The Wuhan Smoking Prevention Trial (WSPT) is a longitudinal, randomized,
school-based smoking prevention trial. The trial was initiated jointly by the Center for
Disease Control and Prevention in Wuhan, China and the Department of Preventive
Medicine at the University of Southern California in 1998. The goal was to develop
and test the effectiveness of a school-based social normative smoking prevention
curriculum among adolescents in Wuhan, China. A middle school was randomly
selected from each of the 7 urban districts in Wuhan. Another middle school with
similar school size, teacher/student ratio, and academic rating in the same district was
selected later. One school from each matched pair was randomly assigned to the
program group. Four 7th grade classrooms from each school were randomly selected
to participate in the evaluation of the WSPT. Participating students completed a
baseline smoking survey conducted between December 1998 and January 1999.
Response rate at the baseline was 97% (98% for program group and 96% for control
group). All participating students in the program schools received 13 consecutive 45-

23
min classroom lessons with one lesson each week. The control schools maintained
their normal activities. The subjects were surveyed using a 200-item questionnaire at
four waves, baseline (7th grade), one-, two-, three- year follow-up (8th, 9th & 10th
grade).
The WSPT is a modified version of Project SMART (Graham, Johnson, Hansen,
Flay, & Gee, 1990; Hansen, Johnson, Flay, Graham, & Sobel, 1988). Project SMART
was translated to Chinese and then evaluated by Chinese-speaking researchers trained
in the US to ensure translation accuracy. To ensure the relevance and practicality of
the curriculum for implementation in Chinese classrooms, a team of public health
officials and educators from Wuhan evaluated the materials and made changes to
accommodate Chinese culture (e.g., session content specific to national smoking
prevalence rates, anti-tobacco policies, and pictures of cigarette advertisements was
replaced with similar information about China). Several other changes were made to
make the program more appealing and memorable to Chinese adolescents. Students
also make public commitments in front of their classmates not to smoke and discuss
the negative social and physical consequences of smoking to establish a social norm
among the students that smoking is unacceptable among their peers. Additional
emphasis was placed on avoiding household exposure to tobacco smoke.
The subjects were surveyed using a 200-item questionnaire at four waves, baseline
(7th grade), one-, two-, three- year follow-up (8th, 9th & 10th grade). The results
showed that boys smoked heavier than girls. The demographic characteristics of the
subjects are shown in Table 1:

24
Table 1: Demographic Characteristics at Baseline (N=2661)

Control
(N=1337)
Program
(N=1324)
Age (years) N 1334 1321
Mean±SD 12.4±0.7 12.6±0.7
Median 12.0 13.0
Range 11.0 – 16.0 11.0 - 18.0

Sex Male 703 (52.6%) 688 (52.0%)
Female 634 (47.4%) 636 (48.0%)

The percentages of subjects in different smoking stages at each measurement time
points (baseline and one-, two- year follow-ups) based on all urban eligible subjects
from 14 schools are summarized in Table 2. The three- year follow-up data were
collected from only 8 schools successfully. Therefore, first 3 waves of data (for linear
growth) were used in the growth curve analysis in the study three, but only focused on
the first two wave data (baseline to one-year follow-up) in the study one and two.

Table 2: Percentages of Subjects in Different Smoking Stages Across Waves
Baseline Wave 2 Wave 3
Gender Control Program Control Program Control Program
Male # of subjects 701 681 675 634 649 592
Never Smokers (%) 63.8% 53.2% 47.9% 39.6% 42.2% 32.8%
Life-time ever
smokers (%)
21.5% 27.2% 38.7% 45.0% 42.8% 48.5%
30-day ever smokers
(%)
14.7% 19.7% 13.5% 15.5% 14.9% 18.8%

25

Table 2: Continued
Baseline Wave 2 Wave 3
Gender Control Program Control Program Control Program
Female # of subjects 634 634 621 612 616 593
Never Smokers (%) 83.6% 79.8% 73.4% 70.1% 68.8% 65.6%
Life-time ever
smokers (%)
10.9% 13.4% 24.2% 26.0% 29.7% 31.9%
30-day ever smokers
(%)
5.5% 6.8% 2.4% 3.9% 1.5% 2.5%

1.4.2 Genetic Variants and Tobacco Use in Chinese Adolescents Study

The participants in the WSPT provided genetic material, making possible a genetic
epidemiologic study of polymorphisms in key candidate genes within the serotonin
and dopamine systems and their impacts on tobacco use. With high-throughput
genotyping method, this study genotyped over 1,535 single nucleotide polymorphisms
(SNPs) and functional polymorphisms in 57 key candidate genes within the serotonin
and dopamine systems in the sample of 2,661 students from WSPT. It used
information from the International HapMap Project (www.hapmap.org) to select
multiple SNPs to characterize the underlying genetic structure within each candidate
gene. We collected the genotyping data for repeat polymorphism and 15 SNPs from
MAOA gene.

26
1.4.3 Important Measures
The original surveys included demographic variables, perceived social norm,
perceived refusal self-efficacy, perceived attitude to cigarette use and passive
smoking, perceived consequences of tobacco use and policy violation, social support
for non-use from parents and peers, and frequency and amount of tobacco use. All of
these measures have been used before, with acceptable reliability and validity (Pentz
et al., 1989; Sussman et al., 1993; Unger et al., 1999). The following are the major
measures that were used in the dissertation research.
a) Outcome measures included cigarette use in their whole life, the number of
days smoking during the last 30 days, the number of cigarettes smoked per day during
the last 30 days, tobacco use within the first 30 minutes of waking, and attempt to quit.
Consistent with previous research (Choi, Pierce, Gilpin, Farkas, & Berry, 1997; B. R.
Flay, Phil, Hu, & Richardson, 1998; Jackson, Henriksen, Dickinson, Messer, &
Robertson, 1998; Lloyd-Richardson, Papandonatos, Kazura, Stanton, & Niaura, 2002),
the smoking stages assessed in this study were defined as:
• Never smokers - those who answered 'No' to the question regarding ever trying
a puff of cigarettes;
• Life-time ever smokers - those who admitted trying cigarette smoking,
although answering 'No' to the question of smoking within the past 30 days or
ever smoking regularly (i.e., daily smoking);
• Thirty-day ever smokers - those who reported smoking during the past 30
days;

27
b) Genotype measures were obtained for candidate genes within the serotonin and
dopamine pathways. Numerous population-based association studies have been
performed to explore the effects of a number of candidate genes, such as cytochrome
P450 (Batra, Patkar, Berrettini, Weinstein, & Leone, 2003; Walton, Johnstone,
Munafo, Neville, & Griffiths, 2001), dopamine receptor (DR) and transporter (Batra,
Patkar, Berrettini, Weinstein, & Leone, 2003; Duggirala, Almasy, & Blangero, 1999;
Erblich, Lerman, Self, Diaz, & Bovbjerg, 2004), serotonin transporter and nicotinic
acetylcholine receptor (Ishikawa et al., 1999; Lerman et al., 1998), on smoking
behavior. We obtained the genotypes on these previously investigated
polymorphisms. In addition, for additional candidate genes, we included tagging that
capture the underlying genetic diversity within the Han Chinese. These SNPs were
determined applying data available form the International HapMap Project
(www.hapmap.org). In addition to tagging SNPs, when available, putative functional
SNPs were included.
The number of repeat alleles can define each genotyping category. We used the
MAOA VNTR Genotyping with GeneScan Method to test the allele size. The MAOA
promoter 30-bp repeat was amplified by PCR with oligonucleotide primers: VIC-
MAOA_HermanF (5 ′-CAGAAACATGAGCACAAACGCCTCAGC-3 ′) which was
labeled with VIC and MAOA_mHermanR (5 ′-GACCGCCACTCAGAACGGACG-3 ′)
(Herman et al., 2005).
PCR reaction contained 50 ng of genomic DNA, 200nM of each primer, 1.25 unit
of Amplitaq Gold polymerase from ABI with Premix Buffer E from Epicentre
Biotechnologies in a total volume of 20 μl. Cycling conditions were initial

28
denaturation at 94 °C for 10 min, followed by 32 cycles of 34s at 94 °C, 34 s at 62 °C,
1.5min at 72 °C, and a final elongation step for 7 min at 72 °C. PCR products were
assayed on Applied Biosystem 3130xl Genetic Analyzer with GS500 LIZ size
standard (35-500 bp, ABI). Results were analyzed using GeneMarker v1.5
(SoftGenetics) software.

c) Psychosocial and intrapersonal measures include depression, hostility,
intention to smoking, social norm, social influences, refusal self-efficacy, attitude to
cigarette use, perceived academic grades, gender, age, etc.
Among these variables, depression was assessed by four items regarding to the
negative feelings and worries in the past week (1: never; 2: rarely; 3: occasionally; 4:
often). They are: 1) have you felt depressed in the last week? 2) have you felt alone in
the last week? 3) have you felt sad in the last week? 4) have you felt like crying out in
the last week? The alpha for these four items was 0.80 at baseline. The depression
score at baseline was the average over the answers to these four items. The score is
skewed right. The median was 1.5 among girls and 1.25 among boys. A separate
survey conducted within 1388 10th grade students in Chengdu, Wuhan, and Qingdao
China revealed a correlation of 0.74 between the ‘depression’ assessed with the 4 item
Wuhan scale and the 20 item CESD scale.(Radloff, 1977) Hostility/family conflict was
assessed by a summary score of the three-item responses (0: No; 1: Yes). The
questions are: 1) do you often get blamed because of doing dangerous things? 2) are
your parents picky or do they nag you about your behavior? 3) do you often argue
with family members? The alpha for these three items was 0.53 at baseline.

29

Some of other variables, such as gender, were treated as potential moderators.
Some were tested as potential mediators, such as refusal self-efficacy, social
influences, etc. Others might be controlled as covariates in the relevant analyses:

Potential Mediators
Refusal self-efficacy was measured by asking the respondents about perceived
ability to resist cigarette offer by a best friend (1= definitely not, 2= maybe not, 3=
maybe yes, 4= definitely yes). In the analyses, a dichotomous variable was created.
Those who answered ‘definitely yes’ was defined as 1, and the other three categories
were combined as 0.
Attitudes toward cigarette use were assessed by the mean of two-item responses
(1= definitely not, 2= maybe not, 3= maybe yes, 4= definitely yes) regarding
participants’ personal attitudes toward cigarette use. The questions are: 1) Is it
impolite to refuse a cigarette? 2) Is it courteous to offer cigarettes to guests?
Perceived social norm were assessed by asking the subjects the question, how
many out of 100 people of your age do you think smoke at least once a month? (0.
None, 1. About 10 people, 2. About 20 people, 3. About 30 people, 4. About 40
people, 5. About 50 people, 6. About 60 people, 7. About 70 people, 8. About 80
people, 9. About 90 people, 10. About 100 people)
Other important variables

30
Other important variables from the survey include: access to cigarettes was
assessed by asking the respondents whether they thought cigarette would be easy to
get (1=very difficult, 2=fairly difficult, 3=fairly easy, 4=very easy). Friends smoking
was assessed using the numbers of the best same-sex and opposite-sex friends who
smoked cigarettes, separately. Perceived friends’ approval of smoking was assessed by
the question, “have any one of your best friends ever asked you not to smoke” (0=no,
1=yes). Perceived parental approval of smoking was assessed by asking the
respondent about perceived parental reaction to lighting up a cigarette by the
respondent in front of them (0 = tell to stop, 1 = not tell to stop or have no reaction).
Cigarette offers were measured by asking whether the respondents who had never
smoked a whole cigarette or more had ever been offered a cigarette (0= no, 1= yes).
Social support was measuring by asking the respondent whether someone was
available to help when in trouble (0=no, 1=yes). School performance was evaluated by
asking the respondent about perceived academic grade last semester (1=very poor,
2=poor, 3=average, 4=good, 5=excellent). SES was measured by weekly allowance
the respondents received in the survey year. Perceived social consequences was
measured by asking the respondents a series of questions about perceived
consequences of tobacco use (1= definitely not, 2= maybe not, 3= maybe yes, 4=
definitely yes). They are 1) smoking can make teeth appear yellow; 2) smoking may
make you lose non-smoker friends; 3) smoking may make people smell weird; 4)
smokers have more young friends; 5) smoking makes young people look maturer; 6)
smoking makes young people look cool; 7) smoking makes young people more
attractive to the opposite sex. Perceived social influences were assessed with a series

31
of questions: 1) does your dad smoke? 2) does your mom smoke? 3) does anyone of
your brothers and sisters smoke? 4) how many of your male friends smoke? 5) do how
many of your female friends smoke?

1.5 THEORETICAL MODEL

Several theoretical models have been applied during the decades of prevention of
adolescent smoking progression. In this section, we review the major causal models
used in the previous smoking prevention studies, and present the theoretical model
developed for my studies.
1.5.1 Major Causal Models in Smoking Prevention
Some causal models that have been proposed in the literature, including social
cognitive theory, social influence theory, stress-coping theory, theory of planned
behavior, etc.
1) Social Cognitive Theory
According to this original social cognitive theory, cognition allows the adolescents
to use former experiences, rather than trial-and-error, to foresee possible consequences
of our acts, and behavior accordingly (Langlois, Petosa, & Hallam, 1999). Bandura’s
Triadic Model shows the relationship among overt behavior, environmental factors,
and personal factors. According to this method, self-regulation allows the adolescents
to choose behaviors that help them to avoid punishments and move towards long-term
objectives. It further can help the adolescents to form a peer group because of certain
dispositional attributes. For example, adolescents with certain dispositions attract

32
rebellious friends or provoke family conflict (van den Eijnden, Spijkerman, & Engels,
2006). And then, the social environment has the influences on smoking progression.
In a psychosocial smoking prevention programs used to decrease adolescent smoking
initiation, the Social Cognitive Theory constructs include behavioral capability to
resist positive images of smoking, refusal skill-efficacy, total positive refusal
expectations and importance, and total negative refusal expectations and importance,
etc. The smoking prevention program is expected to impact on student refusal skill-
efficacy and total positive refusal expectations and importance (Langlois, Petosa, &
Hallam, 1999).
Environmental
Influences
Personal Factors
(beliefs, expectations,
self-perceptions, dispositional attributes, etc.)
Overt
Behavior

Figure 1: Bandura's Triadic Model of Reciprocal Determinism

33
The dispositional attributes in the set of personal factors are influenced by genetic
and environmental factors. In the social cognitive model, external reinforcement isn’t
the only way in which behavior is acquired, maintained, or altered.
In my dissertation research, several paths in this model (e.g. environmental
influences -> smoking progression, environmental influences -> personal factors,
personal factors -> smoking progression) were tested. The research on the other paths
(e.g. smoking progression -> environmental influences, smoking progression ->
personal factors) was not conducted because of the aims of my studies.

2) Social influence theory
Social influence theory has been broadly used in adolescent smoking prevention.
Smoking
Progression
Attitude
Social
norm
Self
efficacy
Age
Gender
Education
performance.
Health status
…….
Parental smoking
Dispositional
Attributes

Figure 2: Social Influence Theory with Dispositional Attributes as Moderators

34
According to the above causal model followed by the social influence theory, the
motivation factors (attitude, social norm, and self-efficacy) are the major mediators,
which were investigated in the previous research (Lloyd-Richardson, Papandonatos,
Kazura, Stanton, & Niaura, 2002). In the above figure, it shows that dispositional
attributes moderate the effects of parental smoking on the mediators to adolescent
smoking progression. This model can be expanded to test the potential interaction
effects between dispositional attributes and any other risk/protective factors.
This is a subset of the theoretical model tested in my dissertation research. The
only difference is that this model only focuses on a specific interaction between
dispositional attributes and a risk/protective factor.

3) Stress-coping theory
Stress-coping theory has been applied in adolescent smoking prevention also
(Wills, 1986; Wills, Resko, Ainette, & Mendoza, 2004). The life changes and stresses
that occur during adolescence may have a substantial negative impact on emotional
well-being and result in the adoption of unhealthy or maladaptive behaviors. The
unsuccessful adjustment to these life changes has been posited to lead to psychological
distress. Stress and associated distress or depression are key factors in the smoking
inception. Stress, measured in various ways, is consistently associated with initiation
to smoking while the presence of stress can be implied through associations with its
outcome depression (distress).

35
Coping
Mechanisms
Outcomes
Long
-term
Problem
focused
Stress
appraisal
Short
-term
Influencing
factors (age,
gender, etc.)
Emotion
focused
Distributional
attributes
Theoretical Stress/Coping Model (Adapted from Lazarus and Folkman 1984)

Figure 3: Theoretical Stress/Coping Model
Supposed there is an association between the need to cope with stress and smoking
initiation. If such an association existed, both high levels of stress and low levels of
coping resources would be thought to be associated with tobacco use. It has been
found that for both males and females, stress (and the method in which it was coped
with) offered powerful predictors of smoking initiation (Koval & Pederson, 1999). For
males, rebelliousness was particularly significant. For females, although a factor,
rebelliousness was less significant than other predictors.
This approach is not related to my research very much. The curriculum used in
Wuhan smoking prevention trial (WSPT) was designed according to social influence
theory. This smoking prevention curriculum emphasizes the creating and reinforcing
non-smoking attitude, correcting social norms toward tobacco use, increasing self-

36
management and decision making skills, enhancing self-efficacy with regard to
smoking behavior. It did not emphasize the stress-coping skills.

4) Enhanced theory of planned behavior
Theory of planned behavior was derived from theory of reasoned action. Both of them have been
applied to adolescent smoking progression successfully (Bledsoe, 2006; Gerber, Newman, & Martin,
1988; Guo et al., 2006; Harakeh, Scholte, Vermulst, de Vries, & Engels, 2004; McGahee, Kemp, &
Tingen, 2000; Unger, Rohrbach, Howard-Pitney, Ritt-Olson, & Mouttapa, 2001). In the below figure,
the bubble with prototype factors for certain dispositional attracts is added in the conventional theory of
planned behavior(van den Eijnden, Spijkerman, & Engels, 2006).
Intention Use
Prototype
factor
Attitude
Knowledge
Subjective
norm
Perceived
Behavioural
control
Age
Gender
Education
performance.
Health status.
.
.
.
.
.
.
.

Figure 4: Enhanced Theory of Planned Behavior
This approach is based on the idea that social identification processes play an
important role during adolescent years. In line with this assumption, explanations of
young people’s smoking initiation should consider factors related to social identity,

37
such as social images related to smoking or smokers. The concept of social identity
and social images are considered when using the enhanced theory of planned behavior
as explanatory model for adolescent smoking. Therefore, a construct, prototype
factor, is added in the original theory of planned behavior. In the original theory,
human behavior is guided by three kinds of considerations: beliefs about the likely
outcomes of the behavior and the evaluations of these outcomes (behavioral beliefs),
beliefs about the normative expectations of others and motivation to comply with
these expectations (normative beliefs), and beliefs about the presence of factors that
may facilitate or impede performance of the behavior and the perceived power of these
factors (control beliefs). In their respective aggregates, behavioral beliefs produce a
favorable or unfavorable attitude toward the behavior; normative beliefs result in
perceived social pressure or subjective norm; and control beliefs give rise to perceived
behavioral control. In combination, attitude toward the behavior, subjective norm, and
perception of behavioral control lead to the formation of a behavioral intention. As a
general rule, the more favorable the attitude and subjective norm, and the greater the
perceived control, the stronger should be the person’s intention to perform the
behavior in question. Finally, given a sufficient degree of actual control over the
behavior, people are expected to carry out their intentions when the chance emerges.
Intention is thus assumed to be the immediate antecedent of behavior.
It assumes that adolescents are more likely to start smoking when they hold
favorable social images or prototypes of smoking peers. Adolescents who believe that
the smoking peer are sociable more frequently engage in smoking behavior. Moreover,

38
adolescents who hold the image that the smoking peer are rebellious are less inclined
to engage in smoking.
In this model, the dispositional attributes, which relate to prototype factors, are
influenced by genetic and environmental factors. However, in this model, prototype
factors impact the motivational factors directly.
When the smoking stage is the outcome, the intention to smoke can be
incorporated into the smoking stage by adding a stage of “susceptible to smoking”.
Therefore, in that case, it is unnecessary to have a construct for intention to smoking.
In the approach of enhanced theory of planned behavior, social identity and social
image may play important influence on adolescent smoking progression. I conducted
research on their roles to smoking progression in the study 2.

1.5.2 Theoretical Model Developed for this Study
The essential objective of WSPT was to develop and test the effectiveness of a
school-based social normative smoking prevention curriculum among adolescents in
Wuhan, China. This smoking prevention curriculum emphasizes the creating and
reinforcing non-smoking attitude, correcting social norms toward tobacco use,
increasing self-management and decision making skills, enhancing self-efficacy with
regard to smoking behavior. Therefore, these motivational factors (attitude to
cigarette use, social norm, refusal self-efficacy) are major factors treated as mediators.
It is assumed that the effects from the risk/protective factors to smoking progression
might be different between the groups defined by gender and dispositional attributes.

39
In the dissertation, I focused on the analyses based on social influence theory, in
which motivation factors (attitude to cigarette use, social norm, self-efficacy) are
considered as the major mediators, and dispositional attributes (depression and
hostility/family conflict) are considered as potential moderators.
The comprehensive model for analyzing the genetic and environmental influences
on smoking progression in adolescents is showed in Figure 5. It is an Attitude–Social
influence–self-Efficacy model and serves as the theoretical framework with an
intervention component as well. Originally, this model developed from the Theory of
Reasoned Action (Ajzen, 1988; Fishbein & Ajzen, 1975), but has incorporated
insights of various other theories, such as Social Cognitive Theory (Bandura, 1977,
1986), the Transtheoretical Model (Prochaska & DiClemente, 1984). The above
integrative model is applied to explain the gene and environmental influences on
behavioral change. Motivational factors, such as various attitudes, social norm and
self-efficacy, determine a person's behavior. Motivational factors are determined by
various demographic attributes, micro interpersonal factors and macro interpersonal
factors, which include information factors (the quality of messages, channels and
sources used) and awareness factors (knowledge, risk perceptions and cues to action).
Self-efficacy is the social skills with regard to smoking behavior. All of the major
three factors (attitude, social norm and self-efficacy) are determined by various
factors, such as demographics factors (age, gender, etc.), micro-level interpersonal
factors (peer, sibling smoking, susceptibility, etc.), macro-level interpersonal factors
(culture, policy, etc.). Dispositional attributes function as a moderator, which is
influenced by genetic factors and environmental factors. This model emphasizes the

40
creating and reinforcing non-smoking attitude, correcting social norm toward tobacco
use, increasing self-management and decision making skills, enhancing self-efficacy
with regard to smoking behavior. The genetic factor incorporated into the model is
used to explain the dispositional attributes. Based on current knowledge, this model
can depict the mechanism of smoking progression among adolescents more accurately.

Figure 5: Conceptual Model for the Analysis of Genetic and Environmental Factors on
Adolescent Smoking Progression

1.6 Significance
Prior research was limited in the following ways. First, few studies have been
conducted to assess the transition of smoking progression across the whole spectrum
of the change sequence. Most have only focused on the smoking onset. While Martin
et al. (Martin, Velicer, & Fava, 1996) examined the probability of movement among
Smoking Prevention
Program
Macro Level
Interpersonal
Factor
(Policy, culture, etc.)

Smoking
Progression
Attitude
Social
Influences
Self
Efficacy
Demographic
Attribute
(Age, gender, etc)
Dispositional
Attribute
(Hostility, depression, etc)
Genetic Factor
(Candidate Genes regulate
serotonin and dopamine)
Micro Level
Interpersonal Factor
(Peer, sibling smoking,
etc.)
Environmental Factor
(Demographic attributes,
Micro/Macro Interpersonal factor)
Genetic Factor
(Nicotine Metabolism)

41
the stages based on data from adult smokers (mean (± SD) age of the sample was 39.7
(±12.0) years), they focused on smoking cessation. Second, little is known about the
contributions of genetic factors on adolescent smoking progression. Therefore, little
work has been aimed at teasing out the genetic effects from psychosocial and cultural
factors when assessing adolescent smoking escalation. Third, they did not specifically
address whether some risk factors or protective factors act as moderators or mediators.
The different repeat allele sizes of a certain gene or different set of genes/SNPs
might be associated with different smoking transition patterns. In the study I extended
prior research in several ways. First, I tested the association between genetic factors
and adolescent smoking behavior/related dispositional attributes (depression,
hostility/family conflict) using repeat polymorphisms within a candidate gene,
MAOA. Second, in the study I modeled the progression patterns in adolescent
smoking and empirically identified the genetic and environmental factors for
adolescent smoking progression. Third, I investigated the moderator and mediating
effects among genetic, psychosocial and cultural variables on adolescent smoking
progression. The study enhanced our understanding in growth trajectories of
adolescent tobacco use. This provides information about etiological models and
intervention strategies for smoking control among adolescents. It assessed the
intervention effect and the effects of risk and protective factors more accurately by
teasing out the genetic factors from environmental factors. Thus, the study is helpful
for us to develop enhanced prevention interventions for smoking progression among
adolescents in the future. Furthermore, we might develop tailored prevention program

42
for certain people with specific genes by better understanding the underlying the
biological need and mechanism for smoking.
1.7 Primary Goals
The three research studies investigated in-depth the relationship between genetic
variants and smoking progression among adolescents, focusing on genetic factors
responsible for dispositional attributes, such as depression and hostility, act to both
influence an individual's tobacco use progression and to moderate the effectiveness of
interventions of smoking prevention. These studies provided insights in developing
more effective smoking prevention programs against tobacco progression among
adolescents. The following primary goals were fulfilled:
1) To investigate the relation of adolescent smoking progression and the gene
polymorphisms (repeat polymorphism) of MAOA, to identify what kind of
adolescents by the genetic variants at a higher risk of cigarette smoking
progression;
2) To investigate the environmental risk/protective factors to adolescent smoking
progression for the subjects with different allele size of MAOA;
3) To test the smoking prevention program effects, controlling for both genetic
and environmental factors;
4) To examine the moderating and mediating mechanisms with the data regarding
the genetic and environmental risk factors.

43
CHAPTER 2: RESEARCH STUDY 1 - ASSOCIATION
BETWEEN GENETIC POLYMORPHISMS OF MAO-A,
DISPOSITIONAL ATTRIBUTES AND SMOKING
BEHA VIOR

This study is the first step to analyze the genetic influence on adolescent smoking
progression. I conducted statistical analyses to investigate the association between
VNTR (Variable number tandem repeat) Polymorphism in Monoamine Oxidase A, the
dispositional attributes related to smoking behavior, such as depression and hostility,
and smoking progression.
Genetic polymorphism is a discontinuous genetic variation that results in the
occurrence of several different forms or types of individuals. Primarily, two types of
genetic mutation events create all forms of variations. They are the insertion or
deletion of one or more nucleotide(s) and single base mutation which substitutes one
nucleotide for another. Tandem Repeat Polymorphisms belongs to the insertion or
deletion of nucleotide(s). Single Nucleotide Polymorphisms (SNP) belongs to single
base mutation.
Variable Number of Tandem Repeat (VNTR) Polymorphisms exist because of the
repeats or variable number of tandem repeats. It is a very common class of
polymorphism, consisting of variable length of sequence motifs that are repeated in
tandem in a variable copy number.
SNPs are DNA sequence variations that occur when a single nucleotide (A, T, C,
or G) in the genome sequence is altered. For example, a SNP might change the DNA

44
sequence AAGGCTAA to ATGGCTAA. SNPs are the most common class of
polymorphisms.
In this study, the analyses were based on VNTR polymorphisms data.
2.1 Hypothesis

MAOA gene is made of 15 exons located on X-chromosome(Shih & Chen, 2004).
It provides the major enzymatic clearing step for serotonin and norepinephrine
(Meyer-Lindenberg et al., 2006). Research suggested that MAOA has important roles
in monoamine metabolism(Shih, Chen, & Ridd, 1999). This enzyme is important in
the breakdown of the amine neurotransmitters, including dopamine, which is thought
to mediate the reinforcing effects of nicotine and contribute to tobacco dependence. A
common variable number of tandem repeats polymorphism of the MAOA gene has
been described that strongly impacts transcriptional efficiency: enzyme expression is
relatively high for carriers of 3.5 or 4 repeats and lower for carriers of 2, 3, or 5
repeats(Sabol, Hu, & Hamer, 1998). High-activity allele carriers have higher enzyme
expression, lower amine concentration, and present higher scores on behavioral
measures of impulsivity than low-activity allele carriers (Passamonti et al., 2006). The
low activity of MAOA has been linked to enhanced reward and the sensation-seeking
temperament (Hallman, Sakurai, & Oreland, 1990; Johansson, Von Knorring, &
Oreland, 1983). A study also showed that functional MAOA uVNTR allele may act
as a genetic modifier of the severity of autism in males (Cohen et al., 2003).
Based on the knowledge regarding MAOA and the availability of data, I tested a
set of hypotheses regarding the association between gene polymorphism from MAOA

45
and adolescent behavior outcomes related to smoking progression, such as depression,
hostility/family conflict, smoking stages, the amount of monthly cigarette use, etc.
Since more than 98% the subjects have either 3 repeats or 4 repeats, the subjects with
less than 3 repeats or more than 4 repeats are excluded in the analyses.

The following are the hypotheses and the reasons I investigated them:
Hypothesis 1: Chinese adolescents with 3-repeat allele of MAOA have higher
score of hostility/family conflict at baseline.
The deficiency in the X-linked isolated MAOA gene, which represents a marked
disturbance of MAO metabolism, is associated with aggressive behavior (Cases et al.,
1995). Research found that those with low MAO-A activity are much more likely to
develop antisocial behavior, conduct disorder, a disposition toward violent behavior,
or conviction for violent offense than those with high MAO-A activity (Caspi et al.,
2002).
Studies have further provided evidence that the 3-repeat allele of the MAO-A-gene
promoter increases the risk of male adolescent criminal behavior, when interacting
with psychosocial factors(K. W. Nilsson et al., 2006), although, in another study with
129 Chinese Han males, neither antisocial alcoholism nor antisocial personality
disorder is found to be associated with genetic variants of the MAO-A gene (Lu, Lin,
Lee, Ko, & Shih, 2003). With a cohort of 370 healthy young Chinese females, a study
found the subjects who were homozygous for the 4-repeat allele of the MAOA tended
to have a higher total score on the Harm Avoidance, compared with the 3-repeat
carriers (Yu, Yang et al., 2005). We have an assumption that people with a lower

46
harm avoidance score or a higher aggressive/antisocial score are more likely to have a
higher score of hostility/family conflict. Therefore, I tested whether the adolescent
with 3 repeat allele of MAOA are more vulnerable to hostility/family conflict.
Hypothesis 2: Chinese adolescents with 3-repeat allele of MAOA are at less risk of
depression.
Research suggested that an excess of high-activity MAOA gene promoter alleles
resulting in an elevated MAOA activity is a risk factor for major depressive disorder
in females(Schulze et al., 2000). A study showed a significantly increased frequency
of 4-repeat allele in major depressive disorder (MDD) patients, especially in the
female population (Yu, Tsai et al., 2005). Therefore, I hypothesized that Chinese male
adolescents with 3-repeats MAOA alleles and female adolescents with the 3/3 or 3/4
genotypes are at the less risk of depression.
Hypothesis 3: Chinese adolescents with 3-repeat allele of MAOA smoke more
than those with 4-repeat allele of MAOA.
Central dopaminergic reward pathways give rise to dependence and are activated
by nicotine. Allelic variants in genes involved in dopamine metabolism may therefore
influence the amount of tobacco consumed by smokers. Research found male adults
with the 3-repeat genotype of the VNTR polymorphism had a significantly increased
risk of smoking significantly compared to those with the 4-repeat genotype (Jin et al.,
2006). The adjusted ORs were 2.0 (95% CI 1.0-4.1) in the current vs. former smokers
group, and 1.9 (95% CI 1.0-3.6) in the nicotine dependent vs. non-nicotine dependent
group respectively.

47
Two further hypotheses have been tested in this study as well.
Hypothesis 4: For Chinese adolescents with 3-repeat allele of MAOA, the
increase of the score of hostility/family conflict from baseline to two-year follow-up is
bigger than the others.
Hypothesis 5: For Chinese adolescents with 3-repeat allele of MAOA, the
increase of number of cigarettes smoked per day in the past 30 days from the baseline
to one-year follow-up is bigger than the others.
2.2 Statistics Method
Descriptive analyses were employed to check the distributions and normality of
the psychosocial variables. Univariate distributions of tobacco use, social influence
variables, and intrapersonal variables were examined by gender, program conditions
and genotype. Internal consistency reliability of the multi-item psychosocial scales
(e.g., depression) were calculated using Cronbach’s coefficient alpha. Wherever
applicable, group comparisons among gender, program conditions and genotypes were
done by t-test or Wilcoxon Rank Sum test for continuous, and Chi-square tests for
categorical variables.
Based on the baseline and one-year follow-up data (two-year follow-up data for
hostility/family conflict), the allele- and genotype-based analyses (allele-based for
males and genotype-based for females) of repeat polymorphism within the candidate
gene, MAOA, were employed with multilevel random effects models. The covariates
were age, education performance, peer smoking, and pocket money at baseline. Males
are hemizygous carriers of one MAOA allele. The dichotomous allele variable was

48
defined as 1 for 3-repeat allele, and 0 for 4-repeat allele. Females carry two alleles of
MAOA. In the multilevel random effects models for females, two dichotomous allele
variables were defined for the three types of repeat genotypes ‘3/3’, ‘3/4’ and ‘4/4’.
They are 1 and 0 for ‘3/3’ repeat genotype, 0 and 1 for ‘3/4’, two 0s for ‘4/4’. The
interaction between the genetic factor and prevention program was tested in the
multilevel random effects model as well.
2.3 Results

Among these 2661 participants at baseline, 2545 students (N = 1297 in the control
group, and N = 1248 in the program group) surveyed in 1-year follow-up, and 2455
students (N = 1267 in the control group, and N = 1188 in the program group) surveyed
at 2-year follow-up. It means, among the sample at the baseline, in year-one and year-
two follow-up, the attrition rates were 4.4% and 7.8%, respectively. The program
group had a higher attrition rate than the control group in each follow-up wave. In
year-one follow-up, while the difference in attrition between the program and control
groups was significant among males, it is not significant among females (Chou et al.,
2006). In year-two follow-up, the differences in attrition between the program and
control groups were significant among both males and females. In WSPT, we
collected cigarette use data at baseline and each follow-up wave, hostility/family
conflict data at baseline and year-two follow-up. No depression data in the same scale
as the baseline were collected in follow-up waves.
Among the 2661 participants at baseline, 2324 buccal cells DNA samples (87.3%)
were genotyped successfully to get genotyping records regarding VNTR

49
polymorphisms from female (552 in control and 563 in program) and male subjects
(629 in control and 580 in program), respectively.
The following table presents the baseline characteristic for these subjects:
Table 3: Basic Characteristics of Subjects in the Analysis Sample
Male Female

Control
(n=629)
Program
(n=580)
Control
(n=552)
Program
(n=563)
Smoking Stage Never
smokers
404 (64.2%) 305 (52.6%) 464 (84.1%) 443 (78.7%)
Life-time
ever smokers
133 (21.1%) 157 (27.1%) 58 (10.5%) 79 (14.0%)
30 day ever
smokers
68 (10.8%) 87 (15.0%) 20 (3.6%) 31 (5.5%)
Number of Days
smoking during last
30 days
N 623 577 552 561
Mean±SD 0.82±3.31 0.76±3.37 0.11±1.36 0.08±1.22
Number of
Cigarettes smoked
per day during last
30 days
N 594 509 527 533
Mean±SD 0.82±3.31 0.76±3.37 0.11±1.36 0.08±1.22
Perceived refusal
self-efficacy
Definitely
not
501 (79.7%) 428 (73.8%) 509 (92.2%) 496 (88.1%)
Maybe not 43 (6.8%) 63 (10.9%) 22 (4.0%) 28 (5.0%)
Maybe yes 64 (10.2%) 70 (12.1%) 17 (3.1%) 31 (5.5%)
Definitely
yes
17 (2.7%) 17 (2.9%) 4 (0.7%) 7 (1.2%)
Education
performance
Very poor 27 (4.3%) 44 (7.6%) 5 (0.9%) 19 (3.4%)
Poor 62 (9.9%) 81 (14.0%) 25 (4.5%) 44 (7.8%)
Average 134 (21.3%) 130 (22.4%) 74 (13.4%) 115 (20.4%)
Good 249 (39.6%) 213 (36.7%) 252 (45.7%) 252 (44.8%)
Excellent 154 (24.5%) 105 (18.1%) 194 (35.1%) 130 (23.1%)

50
Table 3: Continued

Male Female

Control
(n=629)
Program
(n=580)
Control
(n=552)
Program
(n=563)
Pocket money
(Yuan)
N 582 539 535 533
Mean±SD 22.3±32.7 21.1±29.8 20.0±29.7 18.9 ±24.4
Depression N 628 580 552 563
Mean±SD 1.56±0.67 1.57±0.64 1.77±0.79 1.76±0.78
Hostility/Family
conflict
N 612 570 548 559
Mean±SD 0.31±0.32 0.33±0.34 0.29±0.32 0.33±0.32
Peer smoking N 625 577 549 561
Mean±SD 1.08±2.51 1.32±2.64 0.62±2.03 0.49±1.56

Among the alleles reported in VNTR polymorphism in the MAOA promoter, only
the variants with 3 or 4 repeats are common in different ethnic populations (Sabol, Hu,
& Hamer, 1998), and this observation was confirmed in our sample as well. We
observed six different allele sizes among the subjects: 210, 241, 270, 297, 326, and
327. And the frequencies of common allele sizes (270 and 297) in MAOA were more
than 98% for both boys and girls. Table 4 presents the percentage of subjects with
different allele sizes. P values showed that there is no association between program
condition and allele sizes across the gender.

51
Table 4: Allele Size by Gender and Program Condition

Male Female
Allele Size
Control
(n=629)
Program
(n=580) p
Control
(n=1104)
Program
(n=1126) p
210 0 (0.0%) 0 (0.0%) 0.780 0 (0.0%) 2 (0.2%) 0.105
241 5 (0.8%) 2 (0.3%) 2 (0.2%) 10 (0.9%)
270 372 (59.1%) 350 (60.3%) 672 (60.9%) 673 (59.8%)
297 251 (39.9%) 226 (39.0%) 427 (38.7%) 436 (38.7%)
326 0 (0.0%) 1 (0.2%) 0 (0.0%) 1 (0.1%)
327 1 (0.2%) 1 (0.2%) 3 (0.3%) 4 (0.4%)

The repeated number corresponding to the allele sizes of 210, 241, 270, 297,
326/327 are 1, 2, 3, 4, 5. In all of the analyses regarding to the influence of genetic
variants, only the subjects with 3 or 4 repeats were used. That is, data from 1199 boys
were used in allele-based analyses, and data from 1104 girls were used in genotype-
based analyses.

2.3.1 Association between VNTR Polymorphisms of MAOA and Smoking Behavior
Boys
The associations between the predicators to smoking behavior and allele frequency
were checked and summarized in table 5 at below:

52
Table 5: Descriptive Statistics of Major Variables Related to Smoking Behavior by Allele Size
(Boys)

3 Repeat
(n=722)
4 Repeat
(n=477) P
Smoking Stage Never smokers 421 (58.3%) 281 (58.9%) 0.938
Life-time ever
smokers
177 (24.5%) 112 (23.5%)
30 day ever
smokers
93 (12.9%) 60 (12.6%)
Life time smoking No 415 (57.5%) 276 (57.9%) 0.871
Yes 299 (41.4%) 195 (40.9%)
# of days smoking in the past 30
days
0.92 (3.89) 0.90 (3.91) 0.924
# of cigarettes
smoked per day in the past 30 days
0.35 (1.75) 0.17 (0.54) 0.025*

Based on the results from the table 5, we know that the male adolescents with 3-
repeat allele of MAOA reported that they smoked more number of cigarettes per day
during last 30 days than those with 4 repeat allele (p=0.025).

Multilevel random effects models were applied to further investigate the
association between the allele size and the smoking behavior at baseline for male
subjects. We also tested the change of the number of cigarettes from baseline and
year-1 follow-up, and the change of number of days smoking during last month. The
results are presented in the table 6, and figures 6 and 7 as below:

53
Table 6: Effects of Allele Size, Program, and Their Interaction on Smoking Behavior (Boys)
Outcome Independent Variable Estimate±SE p
Number of cigarettes smoked per
day in the past 30 days at baseline
Program 0.004±0.141 .977
Allele Size 0.277±0.118 .019*
Program X Allele Size -0.104±0.171 .542
Number of smoking days in the past
30 days at baseline
Program -0.062±0.440 .890
Allele Size 0.095±0.305 .756
Program X Allele Size 0.233±0.444 .599
Change of number of cigarettes
smoked per day in the past 30 days
from baseline to year 1 follow-up
Program -0.070±0.149 .645
Allele Size -0.319±0.128 .013*
Program X Allele Size 0.141±0.186 .448
Change of number of days smoking
in the past 30 days from baseline to
year 1
follow-up
Program 0.026±0.421 .953
Allele Size -0.095±0.332 .774
Program X Allele Size 0.091±0.489 .853

Figure 6 showed that the genetic factor has influence on the number of cigarette
use per day during last 30 days for boys. For males with 3-repeat allele of MAOA,
they significantly smoked more cigarettes per day than the males with 4-repeat allele
of MAOA at baseline (p=0.019). From figure 7, we know that allele size had
significantly negative effect on the change of the number of cigarettes use from
baseline to the year-one follow-up for the males (p=0.013). No interaction between
the allele size of MAOA and prevention program exists for males.

54
0
0.1
0.2
0.3
0.4
0.5
3 Repeats 4 Repeats
# of Cigarette Use Per Day During
Last Month
Control Program
P for allele size=0.019

Figure 6: # of Cigarette Use Per Day During Last Month by Prevention Program and Allele Size
(Boys)
-0.2
-0.1
0
0.1
0.2
3 Repeats 4 Repeats
Change of Cigarette Use Per Day
During Last Month
Control Program
P for allele size=0.013

Figure 7: Increase of Cigarette Use Per Day During Last Month by Prevention Program and
Allele Size (Boys)

55

To understand the results more clearly, in Appendix, three sets of figures for males
were prepared to present the number of cigarette use per day during last month at
baseline (Figure 18 and 19), the number of cigarette use per day during last month at
baseline in year 1 follow-up (Figure 20 and 21), and the change of the number of
cigarette use per day during last month at baseline from baseline to year 1 follow-up
(Figure 22 and 23). In each set of figures, the first figure is by allele size. The second
figure is by allele size and program. The controlled risk/protective factors are the
same as in the model above, including age, education performance, peer smoking, and
pocket money at baseline.

Girls
The genotype distributions are presented below in table 7:
Table 7: Descriptive Statistics of Major Variables Related to Smoking Behavior by Genotype
(Girls)

Genotype Distributions

3/3
(n=406)
3/4
(n=520)
4/4
(n=167) p
Smoking Stage Never smokers 331 (81.5%) 415 (79.8%) 142 (85.0%) 0.398
Life-time ever smokers 42 (10.3%) 73 (14.0%) 19 (11.4%)
30 day ever smokers 21 (5.2%) 25 (4.8%) 5 (3.0%)
Life time smoking No 327 (80.5%) 413 (79.4%) 141 (84.4%) 0.371
Yes 75 (18.5%) 103 (19.8%) 25 (15.0%)
# of cigarettes
smoked in the past
month
0.04 (0.22) 0.04 (0.24) 0.01 (0.07) 0.218
# of smoking days
in the past month
0.15 (1.43) 0.18 (1.75) 0.05 (0.26) 0.592

56
We found that, without controlling for potential covariates, no association with
genotype variable is significant.

Multilevel random effects models were applied for females also to investigate the
influence of genotype on smoking behavior. The same set of outcome variables and
covariates used in allele-based analyses were tested in these models. No significant
association for Chinese female adolescents exists between Tandom repeat
polymorphisms of MAOA and smoking behavior.
The following table contains the results:
Table 8: Effect of Genotype, Program, and Their Interaction on Smoking Behavior (Girls)
Outcome Independent Variable Estimate±SE p
Number of cigarettes smoked per day in
the past 30 days at baseline
Program -0.005±0.035 0.880
Genotype 3/3 0.022±0.028 0.436
Program X Genotype 3/3 0.021±0.040 0.590
Genotype 3/4 0.014±0.027 0.603
Program X Genotype 3/4 0.014±0.038 0.713
Number of days smoking in the past
30 days at baseline
Program -0.127±0.249 0.620
Genotype 3/3 0.003±0.185 0.987
Program X Genotype 3/3 0.205±0.267 0.442
Genotype 3/4 -0.030±0.179 0.867
Program X Genotype 3/4 0.175±0.256 0.495
Change of number of cigarettes smoked
in the past 30 days from baseline to year 1
follow-up
Program 0.010±0.041 0.817
Genotype 3/3 -0.022±0.034 0.511
Program X Genotype 3/3 -0.000±0.048 0.997
Genotype 3/4 -0.016±0.033 0.631

57
Table 8: Continued

Outcome Independent Variable Estimate±SE p
Program X Genotype 3/4 -0.004±0.047 0.933
Change of number days smoking
in the past 30 days from baseline to year 1
follow-up
Program 0.010±0.041 0.817
Genotype 3/3 -0.022±0.034 0.511
Program X Genotype 3/3 -0.000±0.048 0.997
Genotype 3/4 -0.016±0.033 0.631
Program X Genotype 3/4 -0.004±0.047 0.933

2.3.2 Association between VNTR Polymorphisms of MAOA and Dispositional
Attributes
Boys
The associations between the predicators to dispositional attributes and allele
frequency were checked and summarized in table 9 at below:
Table 9: Descriptive Statistics of Major Variables Related to Dispositional Attributes by Allele
Size (Boys)

3 Repeat
(n=722)
4 Repeat
(n=477) p
Depression 1.58 (0.66) 1.54 (0.66) 0.394
Hostility/Family conflict 0.31 (0.33) 0.34 (0.32) 0.204
Based on the Anova test results, we know that both depression and hostility/family
conflict are not associated with allele size of MAOA for the Chinese male adolescents.
The additional analyses based on the dichotomized variables derived from the original
data and the medians of depression and hostility/family conflict were conducted. The
results confirmed that there is no association between the dispositional attributes and
the allele size of MAOA.

58
Multilevel random effects analyses were also conducted. The allele size of
MAOA in the Chinese male adolescents does not have significant effects on either
depression or hostility/family conflict.

Girls

The genotype distributions are presented the table below:
Table 10: Descriptive Statistics of Major Variables Related to Dispositional Attributes by
Genotype (Girl)

Genotype Distributions

3/3
(n=406)
3/4
(n=520)
4/4
(n=167) p
Depression 1.75 (0.77) 1.80 (0.79) 1.72 (0.82) 0.415
Hostility/Family conflict 0.28 (0.31) 0.33 (0.32) 0.32 (0.33) 0.071+

We found that, without controlling for potential covariates, no significant
difference of depression exists among the genotype groups. However, the association
between hostility/family conflict and repeat genotype is marginally significant
(p=0.071). Since the continuous data did not follow a normal distribution well,
additional analyses were conducted based on the dichotomized data transformed from
the original data and the corresponding medians. The results confirmed that there is no
association between the depression and the repeat genotype of MAOA (p=0.407), but
there is a significant association between hostility/family conflict and the repeat
genotype of MAOA (p=0.029).

59
The following table contains the results from multilevel random effects models:
Table 11: Effect of Genotype, Program, and Their Interaction on Dispositional Attributes (Girls)
Outcome Independent Variable Estimate±SE p
Depression at baseline Program -0.243±0.130 0.086+
Genotype 3/3 -0.098±0.101 0.334
Program X Genotype 3/3 0.299±0.145 0.040*
Genotype 3/4 -0.059±0.098 0.551
Program X Genotype 3/4 0.276±0.141 0.050*
Hostility/family conflict at baseline Program 0.041±0.050 0.426
Genotype 3/3 -0.069±0.041 0.097+
Program X Genotype 3/3 0.057±0.059 0.335
Genotype 3/4 0.027±0.040 0.501
Program X Genotype 3/4 -0.045±0.057 0.431
Change of hostility/family conflict from
baseline to year 2 follow-up
Program 0.038±0.059 0.531
Genotype 3/3 0.107±0.048 0.026*
Program X Genotype 3/3 -0.131±0.070 0.062+
Genotype 3/4 0.003±0.047 0.950
Program X Genotype 3/4 -0.015±0.068 0.821

It demonstrated that for Chinese female adolescents, Tandom repeat
polymorphisms of MAOA has significant impact only on the change of
hostility/family conflicts. And the interaction effects between program and repeat
polymorphisms exist on depression as well.

60

Figure 8: Interaction Effect Between Genotype and Program Condition among Females on
Depression at Baseline.

From Figure 8, we know that, for females, the interaction effect between program
and the repeat polymorphisms of MAOA on depression at baseline is significant. The
subjects with genotype ‘4/4’ have lower depression score in program group than those
in control group. The significant interaction means the randomization in this study was
not successful as designed. There is no significant difference on depression score at
baseline for the females with genotype ‘3/3’ or ‘3/4’.

61

Figure 9: Interaction Effect Between Genotype and Program Condition among Females on the
Change of Hostility/family Conflict.

According to Figure 9, the females with genotype ‘3/3’ had a higher increase of
hostility scores in the control group than those in the program group. It indicates that
the intervention program was successful to reduce the hostility/family conflict scores
for those females with genotype ‘3/3’.

Based on the results from the analyses I conducted, the following conclusion can
be made:
1) Hypothesis 1 is rejected for both girls and boys in this study. Tandom repeat
polymorphisms is not associated with hostility/family conflicts at baseline.

62
2) Hypothesis 2 is rejected for both girls and boys in this study. No association for
Chinese adolescents exists between depression and Tandom repeat polymorphisms.
But we found the females with genotype ‘4/4’ have lower depression score in program
group than those in control group at baseline.
3) Hypothesis 3 is accepted only for boys in this study. Chinese boy adolescents
with the 3-repeat allele of MAOA smoked more cigarettes per day during last month
than those with 4-repeat allele of MAOA (p=0.019).
4) Hypothesis 4 is rejected for boys. However, the increase of the score of
hostility/family conflict from the baseline to two-year follow-up is associated with
Tandom repeat polymorphisms of MAOA for girls. The females with 3/3 genotyping
of MAOA had larger increases in hostility scores in the control group than those in the
program group in the two-year follow-up.
5) Hypothesis 5 is accepted only for boys in this study. The increase of number of
cigarettes smoked per day in the past 30 days from the baseline to one-year follow-up
is more for Chinese male adolescents with the 3-repeat allele of MAOA than the other
males. For female, the hypothesis is rejected.

Overall, the above results showed that the genetic variants of MAOA is associated
with the amount of cigarette use at baseline for male adolescents, and can influence
the increase of the number of cigarettes use. It is also associated with the
hostility/family conflict and the change of hostility/family conflict.

63
CHAPTER 3: RESEARCH STUDY 2 - GENETIC AND
ENVIRONMENTAL RISK FACTORS ON SMOKING
PROGRESSION
Different sets of risk factors might be associated with smoking progression among
adolescents by gender and genotype. There are two sections in this study. In the first
section, I investigated the genetic and environmental risks related to the onset of
cigarette use based on the never smokers at the baseline. Multilevel logistic regression
models were applied to the data from first two waves to identify the different sets of
risk factors by gender. In the second section, I investigated the genetic and
environmental risks on the amount of cigarette consumption based on the life-time
ever smokers at the baseline with the random effects models. The dependent variable
is the number of cigarette smoked during the last 30 days for the life-time ever
smokers.

3.1 Hypothesis
The general hypothesis is that the different sets of risk factors, including genetic
and environmental factors, are associated with smoking progression among
adolescents, stratified by gender. In this study, research focuses on the investigating of
the interaction effects of the dispositional attributes (depression, hostility/family
conflicts), program, and genetic factors by gender, although other psychosocial
environmental risk factors (e.g., peer smoking, pocket money, etc.) were analyzed in
the models as well.

64
The major hypotheses include:
1. Program effect on the adolescent smoking progression is moderated by
hostility/family conflict.
2. Program effect on the adolescent smoking progression is moderated by
depression.
3. The effects of some environmental factors, such as dispositional attributes
(depression, hostility/family conflict), program condition, on smoking
progression might be moderated by repeat polymorphisms of MAOA.
3.2 Statistics Method

In WSPT, the observations in student level are nested within observations in
school level. Therefore, multilevel logistic regression models for dichotomous
dependent variables and multilevel random coefficient models for continuous
dependent variables were employed to assess the program effects and the effects from
genetic and environmental risk/protective factors for different type of smoking
progression pattern. The data from the first two waves were utilized for assessing the
intervention program effect and identifying the sets of potential genetic and
environmental risk factors.
Logic of multilevel models is as: 1) coefficients describing student level are
estimated within each school; 2) coefficients at student level are also analyzed at
school level.

65
The student level is as below:
Y
ij
= β
0j
+ Σβ
qj
x
qij
+e
ij
where
Y
ij
:

student level outcome smoking variable for student i in school j;
β
qj
: student level regression coefficients, q=0, 1 ,…, Q; β
0j
: refers to intercept;
x
qij
: student level predictor q for student i in school j;
e
ij
: student level random effect; and
σ
e
2
: variance of e
ij.

The school level is as:
β
qj
=r
q0
+ Σr
qs
w
sj
+u
qj

where
r
qs
: school level coefficients, q=0,1,…,Q;
w
sj
: school level predictors for school j;
u
qj
: school level random effect.

In both multilevel logistic regression models and multilevel random coefficient
models, all potential risk factors available were assessed. They include genotyping
variables, social, intrapersonal, and demographic variables, etc. Those established
risk/protective factors available in the survey were controlled.

There are two parts of analyses in this study. The first part is the application of
multilevel logistic regression models on the never smokers at the baseline, so that we

66
can understand more about the genetic and environmental risk factors on smoking
initiation. The second part is the application of multilevel random coefficient models
on the life-time ever smokers at the baseline. From this part of analyses, we can
understand more about the genetic and environmental risk factors on the progress of
the amount of cigarette use.
In each part, I first applied the statistical models to subjects with different genetic
types. For boys, allele size 3 is defined as 1, and allele size 4 is defined as 0. One
dichotomous variable was defined for girls. Repeat genotypes ‘3/3’ and ‘3/4’ were
defined as 1, and ‘4/4’ was defined as 0. Additional analyses, in which repeat
genotypes ‘3/3’ defined as 1, and ‘3/4’ and ‘4/4’ defined as 0, were conducted on
female also. To provide a means of formally testing the differences between these two
genetic groups, further analyses in each part were conducted. A full model with
genetic factor and additional interaction items were tested. In these analyses, the
interaction terms between genetic factor and the items adjusted in previous multilevel
logistic model were tested. They were three-way interactions (e.g., the interaction of
allele size, program and depression) or two-way interactions (e.g., the interaction of
allele size and program). All of focal terms and those additional interaction term with
p<=0.20 were kept in the final full model.
In all of the analyses, the covariates adjusted for at the individual level were grand-
mean centered. To properly handle clustered data, the SAS package provides a
procedure GLIMMIX, which is applicable for dichotomous outcome variables, and a
procedure MIXED, which is applicable for continuous outcome variables.

67
3.3 Results

The descriptive statistics on the adjusted covariates by gender and genetic factor
were presented in the table below:
Table 12: Descriptive Statistics on the Adjusted Covariates by Gender and Genetic Factor
Male Female
3 Repeats 4 Repeats
Genotype
3/3
Genotype
3/4 or 4/4
Age at baseline Mean ± SD 12.6 ± 0.67 12.5 ± 0.66 12.4 ± 0.68 12.5 ± 0.67
Depression Mean ± SD 1.58 ± 0.66 1.54 ± 0.66 1.75 ± 0.77 1.78 ± 0.80
Hostility/Family
conflict
Mean ± SD 0.31 ± 0.33 0.34 ± 0.32 0.28 ± 0.31 0.33 ± 0.32
# of Friends
Smoking
Mean ± SD 1.03 ± 2.32 1.43 ± 2.90 0.52 ± 1.68 0.58 ± 1.91
Allowance Mean ± SD 20.7 ± 28.11 22.5 ± 33.11 18.3 ± 24.65 19.9 ± 28.46
Prevention Program Control 372 (51.5%) 251 (52.6%) 205 (50.5%) 342 (49.8%)
Program 350 (48.5%) 226 (47.4%) 201 (49.5%) 345 (50.2%)
Perceived education
performance
Very poor 42 (5.8%) 29 (6.1%) 12 (3.0%) 12 (1.7%)
Poor 86 (11.9%) 57 (11.9%) 20 (4.9%) 46 (6.7%)
Average 156 (21.6%) 105 (22.0%) 68 (16.7%) 119 (17.3%)
Good 288 (39.9%) 169 (35.4%) 181 (44.6%) 312 (45.4%)
Excellent 142 (19.7%) 115 (24.1%) 123 (30.3%) 196 (28.5%)
Access to cigarettes No 429 (59.4%) 266 (55.8%) 228 (56.2%) 450 (65.5%)
Yes 260 (36.0%) 187 (39.2%) 152 (37.4%) 209 (30.4%)
Father smoking No 166 (23.0%) 104 (21.8%) 85 (20.9%) 124 (18.0%)
Yes 535 (74.1%) 352 (73.8%) 312 (76.8%) 541 (78.7%)
Mother smoking No 641 (88.8%) 426 (89.3%) 349 (86.0%) 606 (88.2%)
Yes 57 (7.9%) 34 (7.1%) 41 (10.1%) 59 (8.6%)
Participating in
physical activities
No 311 (43.1%) 207 (43.4%) 205 (50.5%) 293 (42.6%)
Yes 411 (56.9%) 270 (56.6%) 201 (49.5%) 394 (57.4%)

68
The intra-class correlation coefficients of the major smoking variables (i.e.,
smoking stages, life-time smoking, 30-day smoking, age at smoking initiation, and
smoking susceptibility) at baseline have been checked. They are from 0.018 to 0.039.
Therefore, the intraclass correlation of smoking behavior within schools was non-
negligible. Further multilevel analysis needs to be applied. The individual level
models incorporate a school-specific intercept. The higher-level analyses modeled the
school specific intercept and used school level as the random effects factor, thus,
allowed for comparing differences among school levels.

Smoking Initiation of Never Smokers at Baseline
Multilevel Logistic Regression model has been employed to assess the program
effects and the effects of a set of risk/protective factors on the onset of cigarette use
for the adolescents with 3 repeats and 4 repeats, respectively. The adjusted covariates
include age at baseline, perceived education performance, peer smoking, allowance,
access to cigarettes, father smoking, mother smoking, time to watch TV per day,
participating in physical activities, refusal self-efficacy, attitude to cigarette use,
social norm, social consequence, health consequence. The following table presents
the major results from the multilevel logistic regression models on boys:
Table 13: Major Results of Multilevel Logistic Regression Model for the Onset of Cigarette Use
by Allele Size (Boys)
3 Repeats 4 Repeats
RR 95% CI RR 95% CI
Depression 0.961 (0.871 - 1.060 ) 1.094 (0.965 - 1.241 )
Intervention program 0.890 (0.691 - 1.145 ) 1.205 (0.882 - 1.647 )

69
Table 13: Continued

3 Repeats 4 Repeats
RR 95% CI RR 95% CI
Program X depression 1.082 (0.920 - 1.271 ) 0.877 (0.710 - 1.082 )
Hostility/family conflict 1.065 (0.857 - 1.323 ) 1.260 (0.961 - 1.651 ) +
Program X hostility/family conflict 0.971 (0.692 - 1.362 ) 0.889 (0.598 - 1.320 )
Note: Analysis never smoker samples at baseline include 421 boys with 3 repeats and 281 with
4 repeats.

From table 13, we only found hostility/family conflict has marginally significant
effect on the onset of cigarette use (p=0.083) for those boys with 4 repeats. No
interaction effects exist.
The full estimates were presented in the table 24 in Appendix. From this table, we
know that, for a boy who carries 3-repeat allele of MAOA, perceived education
performance and time to watch TV per day, access to cigarettes are risk factors on
smoking initiation. If a boy has 4 repeats of MAOA, then the set of risk factors
become smoking friends, refusal self-efficacy, health consequences, access to
cigarettes.
To provide a means of formally testing the differences between the two genetic
groups, additional analyses were conducted. The genetic factor and additional
interaction items were controlled in the full model, whose results were presented in the
table 25 in Appendix in full estimates. The table below presents the major results
from the final full model.

70
Table 14: Major Results of Multilevel Logistic Regression Model for the Onset of Cigarette Use
(Boys)
RR 95% CI
Allele Size 1.320 (1.034 - 1.685 ) *
Depression 1.118 (0.992 - 1.261 ) +
Intervention program 1.290 (0.959 - 1.737 )
Hostility/Family conflict 1.283 (0.988 - 1.665 ) +
Allele size X depression 0.864 (0.740 - 1.008 ) +
Program X depression 0.871 (0.714 - 1.062 )
Allele size X program 0.693 (0.473 - 1.016 ) +
Allele size X hostility/family conflict 0.796 (0.567 - 1.117 )
Program X hostility/family conflict 0.840 (0.576 - 1.224 )
Allele Size X program X hostility/family conflict 1.144 (0.687 - 1.904 )
Allele Size X program X depression 1.212 (0.939 - 1.565 )
Allele Size X access to cigarettes 0.846 (0.726 - 0.985 ) *
Allele Size X time to watch TV per day 1.066 (1.015 - 1.120 ) *
Note: +: <0.10; *: p <.05; **: p<.01; ***: p<.001 (two-tailed tests).
From this table, we found that the boys with 3 repeats is 1.32 times more likely to
become a life time ever smokers in the year-one follow-up. The effect of access to
cigarettes on the smoking progression for male never smokers at baseline is different
for 3-repeat carriers and 4-repeat carriers. And the effect of time to watch TV per day
on the smoking progression for male never smokers at baseline is different for 3-repeat
carriers and 4-repeat carriers.

The following table 15 presents the major results from the multilevel logistic
regression models when the same risk factors were adjusted in the analysis for female
with genotype 3/3 or 3/4 s and genotype 4/4, respectively:

71
Table 15: Major Results of Multilevel Logistic Regression Model for the Onset of Cigarette Use
by Genotype (Girls)
Genotype (3/3 or 3/4) Genotype (4/4)
RR 95% CI RR 95% CI
Depression 1.047 (0.997 - 1.100 ) + 0.972 (0.888 - 1.063 )
Intervention program 1.034 (0.910 - 1.175 ) 1.002 (0.768 - 1.309 )
Program X depression 0.958 (0.893 - 1.029 ) 0.924 (0.780 - 1.095 )
Hostility/family conflict 1.048 (0.925 - 1.188 ) 0.886 (0.694 - 1.130 )
Program X hostility/family conflict 1.048 (0.849 - 1.293 ) 1.386 (0.943 - 2.039 )
Note: Analysis never smoker samples at baseline include 746 female with genotypes 3/3 or 3/4
and 142 females with genotype 4/4 . +: p<0.10

The table 26 with the full estimates can be found in Appendix. From that table, we
know that, for a girl with genotype 3/3 or 3/4, the set of risk factors for the onset of
smoking are depression score, perceived education performance, attitude to cigarette
use and pocket money. For the girls with genotype 4/4, peer smoking and access to
cigarettes have significant influence on the onset of cigarette use.
A full model with genetic factor and additional interaction items were found and
tested also. No interaction term is significant.
Table 16: Major Results of Multilevel Logistic Regression Model for the Onset of Cigarette Use
(Girls)
RR 95% CI
Repeat genotype 0.854 (0.703 - 1.037 )
Depression 0.972 (0.884 - 1.068 )
Intervention program 0.951 (0.727 - 1.244 )
Hostility/Family conflict 0.999 (0.781 - 1.277 )
Repeat genotype X depression 1.079 (0.970 - 1.199 )
Program X depression 0.955 (0.805 - 1.132 )
Repeat genotype X program 1.087 (0.809 - 1.461 )

72
Table 16: Continued

RR 95% CI
Repeat genotype X hostility/family conflict 1.037 (0.789 - 1.363 )
Program X hostility/family conflict 1.209 (0.816 - 1.792 )
Repeat genotype X program X hostility/family
conflict
0.876 (0.559 - 1.370 )
Repeat genotype X program X depression 1.004 (0.835 - 1.208 )
Note: +: <0.10; *: p <.05; **: p<.01; ***: p<.001 (two-tailed tests).

In the additional analysis with repeat genotype of MAOA defined in recessive
mode. No significant effects on the major variables and interaction items were found.

Smoking Progression of Life-time ever Smokers at Baseline
To investigate the genetic and environmental risks on the amount of cigarette
consumption, random effects models were applied based on the life-time ever smokers
in follow-up year 1. The dependent variable is the number of cigarette smoked per
day during the last 30 days based on the life-time ever smokers at baseline. Because
of limited sample size of life-time ever smokers at baseline, I only applied this model
on male samples.
Table 17: Major Results of Multilevel Random Effects Models for Life-time Ever Smokers at
Baseline by Allele Size (Boys)
3 Repeats 4 Repeats
Estimate SE DF p Estimate SE DF p
# of cigarettes
smoked per day in
the past month at
baseline
-.023 0.026 165 0.365 0.463 0.127 101 0.000 ***
Intervention
program
-.007 0.126 12 0.956 -.320 0.124 12 0.024 *
Depression -.021 0.132 165 0.872 -.311 0.136 101 0.024 *

73

Table 17: Continued

3 Repeats 4 Repeats
Estimate SE DF p Estimate SE DF p
Program X
depression
-.308 0.176 165 0.081 + 0.472 0.193 101 0.016 *
Hostility/Family
conflict
-.290 0.253 165 0.253 -.101 0.260 101 0.698
Program X
hostility/family
conflict
0.146 0.331 165 0.660 -.268 0.405 101 0.510
Note: +: <0.10; *: p <.05; **: p<.01; ***: p<.001 (two-tailed tests).
Analysis sample include 270 boys with 3 repeats and 172 with 4 repeats.

The results on boys were presented at Table 17 above. The interaction effects
between program and depression on the smoking consumption for the boys with 4-
repeat allele of MAOA is significant as well (p=0.024). And the program significantly
reduced the number of cigarettes use per day during past 30 days (p=0.024) for boys
with 4 repeats. The corresponding table with full estimates is the table 28 in
Appendix. From that table, we know that, the set of potential risk factors for the boys
with 3-repeat allele of MAOA include perceived education performance (p=0.061),
the peer smoking (p=0.062), and attitude to cigarette use (0.048), health consequence
(p=0.079). The set of risk factors for the boys with 4-repeat allele of MAOA include
only the amount of cigarette use during the past month at baseline (p=0.007), the
intervention program (p=0.024), age at baseline (0.074), access to cigarettes
(p=0.037), depression (p=0.024), social norm (p=0.036) and health consequence
(p=0.054).

74
Figure 10 showed the interacting effect between program and depression on the
monthly cigarette use among male carriers with 4 repeats. We found that for the boys
with 4 repeats, those in the program group had significantly fewer cigarette use per
day than in the control group (p=0.024). However, for the boys in the program group,
the higher the depression score they got, the more cigarettes they smoked. The
intervention program worked better for those subjects who are not in depression
condition.

0
0.3
0.6
0.9
1.2
Depression score
# of Cigarette Use per Day during past month
Control Program

Figure 10: # of Cigarette Use per Day During last 30 Days for Boys with 4-repeats.

A full model with genetic factor and additional interaction items were tested also.
Its complete estimates were put in table 29 in Appendix. The major estimates were
reported in the table below.
p for program: 0.024
p for depression:0.024
p for interaction: 0.016

75
Table 18: Major Results of Multilevel Random Effects Models among Life-time Ever Smokers at
Baseline (Boys)
Statistics Parameter
Estimate SE DF p
# of cigarettes smoked per day in the past month at baseline -0.011 0.026 290 0.683
Allele Size 0.077 0.205 290 0.709
Intervention program -0.094 0.229 12 0.69
Allele size X program -0.033 0.285 290 0.909
Depression -0.238 0.167 290 0.155
Allele size X depression 0.276 0.213 290 0.196
Program X depression 0.428 0.233 290 0.067 +
Allele size X depression X program -0.842 0.291 290 0.004 **
Hostility/Family conflict 0.053 0.318 290 0.868
Allele size X hostility/family conflict -0.305 0.41 290 0.458
Program X hostility/family conflict -0.431 0.468 290 0.358
Allele size X hostility X program 0.577 0.576 290 0.317
Allele size X attitude to cigarette use 0.329 0.178 290 0.066 +
Allele size X health consequence 0.392 0.164 290 0.018 *

Since the interaction effect between depression and program is significant, the
figures regarding cigarette use (Figure 24 for the baseline, Figure 25 for the year 1
follow-up, and 26 for the change from baseline to year 1 follow-up) were presented in
Appendix, by allele size, program, and depression score. In the figures, the depression
score was dichotomized to low and high only with the median as the dividing point, so
that we can interpret the results easier. The controlled risk/protective factors include
age, education performance, peer smoking, and pocket money at baseline.

76
Because of limited number of female life-time ever smokers (n=185) at baseline, I
did not conduct the analysis to investigate the risk factors for subjects with different
repeat genotype, respectively. I only conduct analyses for the full model with genetic
factor and additional interaction items adjusted. The effect of # of cigarettes smoked
per day in the past month at baseline (p=0.044), hostility/family conflict (p=0.036), the
interaction effect between genetic factor and hostility/family conflict (p=0.027), and
the interaction effect between genetic factor and peer smoking (p=0.044) are
significant. The higher score of hostility/family conflict a girl had, the more cigarettes
she smoked during the past month. And the effect of hostility/family conflict was
moderated by repeat genotype of MAOA. The major results were presented in table
19. The corresponding table with full estimates is the table 30 in Appendix.
Table 19: Major Results of Multilevel Random Effect Model among Life-time Ever Smokers
(Girls)
Statistics Parameter
Estimate SE DF p
# of cigarettes smoked per day in the past month at baseline 0.057 0.028 105 0.044 *
Repeat genotype 0.084 0.053 105 0.114
Intervention program 0.069 0.058 12 0.258
Repeat genotype X program -0.103 0.078 105 0.193
Depression 0.101 0.066 105 0.129
Repeat genotype X depression -0.067 0.051 105 0.187
Program X depression 0.017 0.11 105 0.88
Repeat genotype X program X depression -0.024 0.094 105 0.798
Hostility/Family conflict 0.231 0.109 105 0.036 *
Repeat genotype X Hostility -0.292 0.13 105 0.027 *
Hostility X program 0.022 0.184 105 0.905
Repeat genotype X hostility X program 0.085 0.211 105 0.687

77
In the additional analysis with repeat genotype of MAOA defined in recessive
mode. No significant effects on the major variables and interaction items were found.

From the about analysis results, we draw the conclusions: The different sets of
risk factors, including genetic and environmental factors, might be associated with
smoking progression among both boys and girls. For the male life-time ever smokers
at baseline with high (4-repeat) activity of MAOA, the intervention program reduced
their amount of cigarette use during the past month.
Regarding the smoking initiation, there is no interaction between program and
depression or hostility/family conflict for both boys and girls. The boys with 3 repeats
is 1.32 times more likely to become a life time ever smokers in the year-one follow-up.
The effects of availability of cigarette and time to watch TV per day on smoking
initiation on smoking initiation depend on the allele size. The easier to get cigarette,
the less likely for a boys with 3 repeats to become a life-time ever smoker in two-year
follow-up. However, the easier to get cigarette, the more likely for a boys with 4
repeats to become a life-time ever smoker. Time to watch TV per day has
significantly positive influence on smoking progression for those male never smokers
with 3 repeats.
Regarding to the change of the number of cigarette use per day during last month,
for a boy, the effects from availability of cigarette, attitude to cigarette use and health
consequences depend on the allele size. The three-way interaction among program,
depression and allele size of MAOA exists. For a girl, the effects from
hostility/family conflict depend on the repeat genotype. For a boy with 3-repeats,

78
health consequences can increase his monthly cigarette use. However, for the 4-
repeats carriers, health consequences can reduce the monthly cigarette use.

79
CHAPTER 4: RESEARCH STUDY 3 - MEDIATED
MODERATION MODEL FOR GENETIC AND
ENVIRONMENTAL INFLUENCES ON ADOLESCENT
SMOKING PROGRESSION
In the study one, the associations of the repeat polymorphisms of MAOA with
smoking behavior and its related dispositional attributes (depression and
hostility/family conflict) were tested. In the study two, the genetic and environmental
risk/protective factors on adolescent smoking progression were examined. In the
study three, I conducted further analyses to investigate the mechanism of the
influences of genetic and environmental factors on the smoking progression.
According to the theoretical causal model presented for the dissertation research in
Chapter 1, the motivation factors (attitude, social norm, and self-efficacy) are the
major mediators, which were investigated in the previous research followed by the
social influence theory. The dispositional attributes might moderate the effects of
parental smoking on the mediators to adolescent smoking progression. In the research
study 3, the repeat polymorphisms of MAOA was incorporated in the theoretical
model. This study tried to present a causal model related to the adolescents’ behavior
disease (e.g. smoking progression), which are determined by genetic differences,
environmental differences and their interactions, and with motivational factors as
mediators.
Our substantive knowledge is rarely complete enough to know whether and where
the interaction effects related to a potential moderator exist. Ignoring the difference
from the moderator at any path leads to potential for biased and inconsistent

80
estimators. Thus, it is important to check this type of “group” differences. Latent
growth curve analysis, a type of structural equation modeling (SEM), with multiple
group method is an appropriate method to verify this type of causal model. Path
model and structural equation model were used to test mediation hypotheses, while a
multi-sample approach stratified by the genetic factor, and interaction terms derived
from the genetic factor and the second variable, were used to examine moderation
hypotheses.
Intrapersonal variables such as attitude to cigarette use, social influence, and
refusal self-efficacy were tested as mediators. The detailed procedures (Baron &
Kenny, 1986) were followed to estimate the total, direct, and mediated effects and to
test the statistical significance.

4.1 Hypothesis
Genetic factors, environmental factors and their potential interactions may predict
the adolescent smoking progression. Some factors, such as attitude, social influences,
refusal self-efficacy might function as mediators.
The major hypothesis in this research study is that the motivational factors (refusal
self-efficacy, social norm, attitude to cigarette use) mediate the effects of genetic and
some environmental factors, such as program condition and dispositional attributes, on
smoking progression for both boys and girls. I focused on LGCM with refusal self-
efficacy.

81
Since LGCM is the statistical method applied on the first three-wave data,
additional hypotheses can be tested related to growth rate. They are:
1. The program can reduce the growth rate of the amount of monthly cigarette use
for both boys and girls.
2. Dispositional attributes (depression and hostility/family conflict) are associated
with initial status and linear growth rate of smoking progression and time-
varying motivational factors (refusal self-efficacy, social norm, attitude to
cigarette use).

4.2 Statistics Method
Latent growth curve analysis is multilevel SEM (Muthen & Shedden, 1999)..
SEM can examine both measurement model and the structure model simultaneously.
Maximum likelihood estimate method will be performed to obtain parameter
estimates. Model with goodness of fit indices (ratio of the maximum likelihood chi-
square value to the degrees of freedom < 2.0 or the comparative fit index (CFI) > .90,
and Root Mean Square Error of Approximation (RMSEA) <0.05) is considered
empirically acceptable to fit the data (Newcomb, 1990, 1994). RMSEA is the most
widely used measure of fit. Ideally, RMSEA should be below .05, and a value that is
not below .05 is considered problematic. Each construct will be measured by one or
two manifest variables. Multiple group comparisons will be conducted to test the
differences in the paths across certain allele size or genotype.

82
Latent growth modeling has become a common way to model individual
differences in growth in substance use (Andrews & Duncan, 1998). The growth factor
means represent the typical growth trajectory in the sample, and variance around these
means represents heterogeneity in growth.
In this sub-study, parallel growth model with time-variant mediator were tested.
It is supposed that attitude to cigarette use, social norm and refusal self-efficacy
function as mediators. The mediator can be confirmed with the analysis following the
conventional procedures:
Step 1 Conduct a regression analysis with X predicting Y

01
YB BX e =+ +
Step 2 Conduct a regression analysis with X predicting Z

01
Z BBX e =+ + .
Step 3 Conduct a regression analysis with Z predicting Y

01
YB BZ e =+ + .
Step 4 Conduct a multiple regression analysis with X and Z predicting Y.

01 2
YB BX BZ e =+ + +

Suppose the dependent variable is the amount of cigarette use per day in the past
month, and there are one time-varying mediator and one time-invariant covariate in
the model. The following set of equations can be used to present the model in the first
level and second level:
1
st
level
(For the amount of cigarette use in the past month)
y
ti
=
11
0
22 2
1
0 1
1
1
ii
i
ili i
i
mi mli mi
y
y
y
ε
η
λ ε
η
λ ε
⎛⎞ ⎛ ⎞ ⎛ ⎞
⎛⎞
⎜⎟ ⎜ ⎟ ⎜ ⎟
⎜⎟
⎜⎟ ⎜ ⎟ ⎜ ⎟
=+
⎜⎟
⎜⎟ ⎜ ⎟ ⎜ ⎟
⎜⎟
⎜⎟ ⎜ ⎟ ⎜ ⎟
⎝⎠
⎝⎠ ⎝ ⎠ ⎝ ⎠
## # #

83
(For mediator)
y
dti
=
11
0
22 2
1
0 1
1
1
di d i
di
di dli d i
di
dmi dmli dmi
y
y
y
ε
η
λε
η
λε
⎛⎞ ⎛ ⎞ ⎛ ⎞
⎛⎞
⎜⎟ ⎜ ⎟ ⎜ ⎟
⎜⎟
⎜⎟ ⎜ ⎟ ⎜ ⎟
=+
⎜⎟
⎜⎟ ⎜ ⎟ ⎜ ⎟
⎜⎟
⎜⎟ ⎜ ⎟ ⎜ ⎟
⎝⎠
⎝⎠ ⎝ ⎠ ⎝ ⎠
## # #
Where
i and m indicate a subject and the total number of repeated measurements on a
subject, respectively. y
ti
and y
dti
are the outcome variables at time-point t for subject i,
λ
mli
and λ
dmli
are the time-related linear growth scores for time-point t, and ε
mi
and

ε
dmi

are the residuals at time-point t for subject i (t=1 to m). For a study with evenly spaced
times, the linear growth scores can be set as 0, 1, 2, 3, …, m.

2
nd
level
(For the amount of cigarette use in the past month)

0 00 0 00 0 0 0
1 10 10 1 10 11 12 1 1 1
1
20 20 21 2 20 21 22 2 2 2 2
000 0 0
00
0
i id di wn i
N
i s id d d di niwn i
n
ssi ddd di wn i i
W
η β ηβ η β ζ
η β βη βββη β ζ
β ββ η β β β η β ζ η
=
⎛⎞ ⎛⎞ ⎛ ⎞⎛ ⎞ ⎛ ⎞⎛ ⎞ ⎛ ⎞ ⎛ ⎞
⎜⎟ ⎜⎟ ⎜ ⎟⎜ ⎟ ⎜ ⎟⎜ ⎟ ⎜ ⎟ ⎜ ⎟
=+ + + +
⎜⎟ ⎜⎟ ⎜ ⎟⎜ ⎟ ⎜ ⎟⎜ ⎟ ⎜ ⎟ ⎜ ⎟
⎜⎟ ⎜ ⎟⎜ ⎟ ⎜ ⎟⎜ ⎟ ⎜ ⎟ ⎜ ⎜⎟
⎝⎠ ⎝ ⎠⎝ ⎠ ⎝ ⎠⎝ ⎠ ⎝ ⎠ ⎝ ⎠ ⎝⎠
∑
⎟

(For mediator)

0 00 0 0 0
1 10 11 1 1 1
1
20 21 22 2 2 2 2
000
00
0
diddi dwndi
N
di d m d i ni dwn d i
n
dmm di dwn di di
W
ηβη βζ
η ββ η β ζ
βββ η β ζ η
=
⎛⎞ ⎛⎞⎛ ⎞⎛ ⎞ ⎛ ⎞⎛ ⎞
⎜⎟ ⎜⎟⎜ ⎟⎜ ⎟ ⎜ ⎟⎜ ⎟
=+ + +
⎜⎟ ⎜⎟⎜ ⎟⎜ ⎟ ⎜ ⎟⎜ ⎟
⎜⎟⎜ ⎟⎜ ⎟ ⎜ ⎟⎜ ⎟ ⎜⎟
⎝⎠⎝ ⎠⎝ ⎠ ⎝ ⎠⎝ ⎠ ⎝⎠
∑

84
Where
β
wmn0
, β
wmn1
and β
wmn2
show the impact of time-invariant covariates, w
n
(n=1
to N), on the growth parameters, η
d0i
, η
d1i
, and η
d2i
, of the mediator, respectively. And
β
d00
, β
d10
, β
d11
, β
d12
, β
d20
, β
d21
and β
d22
reflect the impact of mediator on the growth
of smoking progression.

In WSPT, we have measures on the number of cigarettes use per day during last 30
days, refusal self-efficacy, and other potential mediators at three time points (baseline
and two follow-ups). The smoking initial status and progression slopes can be
estimated by gender and allele size/genotype, controlling for other time-invariant
variables. The following figure shows the parallel growth curve model with one time-
varying mediator and only two latent growth variables (intercept and linear growth
term):

85
Intercept
(Mediator)
Slope
(Rate of Mediator)
Intercept
(Smoking Progression)
Slope
(Rate of Smoking
Progression)
M (t
0
)
M (t
1
)
M (t
2
)
S (t
0
) S (t
1
) S (t
2
)
X
1
X
n
X
2
.
.
.

Figure 11: LGCM on Smoking Progression with a Time-Varying Mediator

To provide a means of formally testing the genetic influences on the initial status
and slope of smoking progression and potential mediator, additional Latent Growth
Curve Analyses were conducted based on boys and girls, respectively. In these
additional analyses, the genetic factor and the interaction terms among genetic factor,
program condition, depression and hostility/family/conflict were adjusted as covariates
as well.
MPLUS 4.0 was used in the analyses. Full information maximum likelihood
estimation with missing values (FIML) was applied to use all available data. For
example, some adolescents were surveyed at baseline and all follow-up waves, but
others may have skipped year-1, or year-2 follow-up. All available information was
used with this approach.

86
4.3 Results
Latent growth curve analysis was applied to investigate the mechanism of the
influences of genetic and environmental factors on the smoking progression. Based on
the model introduced in the Figure 9, refusal self-efficacy, perceived social influences,
attitude to cigarette use were considered as a mediator, respectively.
Before modeling the growth of tobacco use, we examined the descriptive statistics
of the time-varying variables, including the number of cigarettes use during last 30
days, refusal self-efficacy (REF), social norm (SN), and attitude to cigarette use
(ACU) across each wave, stratified by gender and genotype. The results for boys were
presented in table 20.
Table 20: Descriptive Statistics of Time-varying Variables across Waves (Boys)

3 Repeats
(mean±SD)
4 Repeats
(mean±SD)
Control Program Control Program
# of cigarettes smoked per day
during last month
Baseline 0.38 (2.04) 0.32 (1.38) 0.14 (0.48) 0.20 (0.60)
Wave 2 0.28 (1.29) 0.22 (0.68) 0.32 (1.06) 0.21 (1.00)
Wave 3 0.32 (1.75) 0.79 (2.58) 0.57 (1.81) 0.33 (1.04)
Refusal Self Efficacy Baseline 0.80 (0.40) 0.76 (0.43) 0.80 (0.40) 0.71 (0.46)
Wave 2 0.72 (0.45) 0.72 (0.45) 0.72 (0.45) 0.71 (0.46)
Wave 3 0.69 (0.46) 0.62 (0.49) 0.65 (0.48) 0.69 (0.46)
Attitude to cigarette use Baseline 1.74 (0.78) 1.74 (0.77) 1.70 (0.80) 1.83 (0.82)
Wave 2 1.86 (0.83) 1.78 (0.80) 1.82 (0.81) 1.85 (0.83)
Wave 3 1.78 (0.79) 1.85 (0.79) 1.85 (0.81) 1.82 (0.84)
Social norm to cigarette use Baseline 2.28 (2.51) 2.43 (2.78) 2.18 (2.60) 2.38 (2.66)
Wave 2 2.74 (2.75) 2.52 (2.55) 2.48 (2.39) 2.71 (2.73)
Wave 3 2.72 (2.54) 3.31 (3.01) 2.93 (2.65) 3.12 (2.95)
The numbers of subjects with eligible genetic and survey data at baseline, wave 2, and wave 3 are
1185, 1151, and 1082.

87
The results of the descriptive statistics for girls were presented in table 21.
Table 21: Descriptive Statistics of Time-varying Variables across Waves (Girl)

Genotype 3/3 or 3/4
(mean±SD)
Genotype 4/4
(mean±SD)
Control Program Control Program
# of cigarettes smoked per day
during last month
Baseline 0.03 (0.21) 0.05 (0.26) 0.01 (0.05) 0.01 (0.08)
Wave 2 0.02 (0.10) 0.04 (0.26) 0.01 (0.08) 0.01 (0.08)
Wave 3 0.01 (0.06) 0.04 (0.45) 0.00 (0.00) 0.06 (0.41)
Refusal Self Efficacy Baseline 0.92 (0.27) 0.88 (0.33) 0.93 (0.26) 0.93 (0.26)
Wave 2 0.91 (0.28) 0.87 (0.34) 0.87 (0.34) 0.89 (0.32)
Wave 3 0.89 (0.31) 0.85 (0.36) 0.81 (0.40) 0.86 (0.35)
Attitude to cigarette use Baseline 1.64 (0.70) 1.69 (0.72) 1.60 (0.70) 1.52 (0.74)
Wave 2 1.65 (0.71) 1.67 (0.72) 1.75 (0.75) 1.52 (0.71)
Wave 3 1.76 (0.72) 1.78 (0.73) 1.77 (0.77) 1.64 (0.70)
Social norm to cigarette use Baseline 1.92 (2.33) 2.11 (2.42) 1.99 (2.14) 1.61 (2.08)
Wave 2 2.13 (2.14) 2.17 (2.21) 2.40 (2.35) 1.99 (2.18)
Wave 3 2.62 (2.26) 2.94 (2.45) 2.51 (2.29) 2.32 (2.04)
The numbers of subjects with genetic and survey data at baseline, wave 2, and wave 3 are
1084, 1071, and 1041.

Latent Growth Curve Analysis Stratified by Gender and Genetic Variants
In the first part, LGCM was applied stratified by gender and genetic variants
(Boys: 3 repeats and 4 repeats; Girls: Genotype 3/3 or 3/4 and genotype 4/4).
1) Refusal self-efficacy as the mediator
The first LGCM model was tested with refusal self-efficacy as the mediator. The
tests for the model fit showed that it is a good fit (CFI = .965, RMSEA = .031). The
model is acceptable. The same set of covariates was adjusted. When both of the
parallel processes (smoking progression and refusal self-efficacy) are linear, we
obtained the results presented in the table 21. Figure 12 presents the change of

88
cigarette use per day during last month from baseline to two-year follow-up by
program condition and allele size for boys.
0
0.15
0.3
0.45
0.6
Baseline Wave 2 Wave 3
Number of Cigarette smoking per day
Control & 3 Repeats Program & 3 Repeats
Control & 4 Repeats Program & 4 Repeats

Figure 12: Growth Curve of Cigarette Monthly Use across Prevention Program and Allele Size
(Boys)

From the table 22, we know that, intercept of refusal self-efficacy (RSE) is
significantly associated with intercept of smoking progression (SMK) for all groups of
subjects, and for boys with 3 repeats, slope of RSE is significantly associated with
intercept of smoking progression as well. The slope of RSE is significantly associated
with the slope of SMK for boys (3 repeats: p=0.013; 4 repeats: p=0.005). Several
risk/protective factors can significantly influence the smoking progression by
impacting the initial status and linear growth of RSE.
From Figure 12, we know that, for boys both with 3 repeats or with 4 repeats, the
subjects in program condition showed slower smoking progression than those in the
control group.

89
Table 22: Major Estimated Parameters from the Model with Refusal Self-efficacy as Mediator by Gender and Genetic Factor
Boy Girl
Low (3-repeat) activity High (4-repeat) activity Genotype of 3/3 or 3/4 Genotype of 4/4
Estimate SE p Estimate SE p Estimate SE p Estimate SE p
Intercept of SMK

Intercept of RSE
-2.020 0.575 0.000 *** -0.888 0.256 0.001 ** -0.656 0.257 0.011 * -0.340 0.113 0.003 **
Slope of RSE
2.742 1.330 0.039 * 0.087 0.304 0.774 0.188 0.210 0.371 3.070 8.278 0.711
Slope of SMK

Intercept of RSE
0.456 0.424 0.282 -0.722 0.306 0.018 * 0.221 0.138 0.108 0.122 0.066 0.067 +
Slope of RSE
-3.278 1.326 0.013 * -1.412 0.503 0.005 ** -0.268 0.177 0.130 -3.167 8.380 0.705
Effects on Intercept of SMK

Intervention Program
0.336 0.323 0.297 -0.026 0.153 0.863 -0.073 0.057 0.197 -0.006 0.173 0.973
Depression
0.144 0.173 0.405 0.067 0.073 0.355 0.015 0.022 0.500 -0.006 0.068 0.934
Program X Depression
-0.278 0.243 0.253 -0.052 0.120 0.666 0.032 0.035 0.362 -0.008 0.095 0.935
Hostility/family conflict
-0.125 0.333 0.707 -0.118 0.116 0.309 -0.031 0.062 0.619 0.165 0.353 0.641
Program X Hostility
-0.411 0.460 0.372 0.216 0.203 0.287 0.032 0.078 0.681 -0.130 0.418 0.756
Effects on Slope of SMK

Intervention Program
-0.142 0.268 0.596 -0.049 0.163 0.765 -0.028 0.036 0.428 0.044 0.171 0.798
Depression
-0.149 0.135 0.270 0.024 0.082 0.771 -0.013 0.016 0.424 0.013 0.067 0.849
Program X Depression
0.191 0.191 0.317 0.047 0.102 0.643 0.022 0.023 0.324 -0.006 0.094 0.951
Hostility/family conflict
0.072 0.289 0.803 0.063 0.171 0.712 -0.016 0.040 0.688 -0.156 0.356 0.661
Program X Hostility
0.092 0.421 0.828 -0.358 0.255 0.162 0.011 0.052 0.831 0.090 0.421 0.831
Effects on Intercept of RSE

Intervention Program
0.043 0.068 0.522 -0.064 0.081 0.428 -0.023 0.055 0.677 0.028 0.052 0.588
Depression
-0.049 0.031 0.116 -0.023 0.035 0.508 0.000 0.025 0.991 -0.018 0.018 0.332
Program X Depression
-0.024 0.046 0.600 -0.043 0.057 0.445 0.001 0.033 0.978 -0.048 0.033 0.139
Hostility/family conflict
-0.110 0.066 0.093 + -0.266 0.075 0.000 *** -0.106 0.068 0.118 -0.058 0.049 0.231
Program X Hostility
-0.018 0.094 0.846 0.292 0.117 0.013 * 0.004 0.098 0.963 0.037 0.083 0.653

90
Table 22: Continued
Boy Girl
Low (3-repeat) activity High (4-repeat) activity Genotype of 3/3 or 3/4 Genotype of 4/4
Estimate SE p Estimate SE p Estimate SE p Estimate SE p
Effects on Slope of RSE

Intervention Program
-0.020 0.049 0.677 0.060 0.067 0.376 -0.024 0.047 0.613 0.016 0.034 0.637
Depression
0.021 0.025 0.389 -0.002 0.030 0.943 -0.016 0.021 0.452 0.006 0.014 0.674
Program X Depression
0.003 0.033 0.918 0.023 0.047 0.622 0.005 0.029 0.873 -0.008 0.020 0.667
Hostility/family conflict
0.047 0.049 0.335 0.053 0.058 0.360 -0.011 0.052 0.833 -0.042 0.033 0.206
Program X Hostility
0.022 0.069 0.754 -0.137 0.094 0.143 0.009 0.073 0.901 0.047 0.050 0.341
+: p<0.10; *: p<0.05; **: p<0.01; ***: p<0.001

91
Figure 13 presents the change of cigarette use per day during last month from
baseline to two-year follow-up by program condition and genotype for girls.
0.01
0.02
0.03
0.04
0.05
Baseline Wave 2 Wave 3
# of Cigarette Use Per Day during Last Month
Control & Genotype 3/3 or 3/4 Control & Genotype 4/4
Program & Genotype 3/3 or 3/4 Program & Genotype 4/4

Figure 13: Growth Curve of Cigarette Monthly Use across Prevention Program and Allele Size
(Girls)

From Figure 13, we know that, for girls in the program group, their amount of
cigarette monthly use decrease from baseline to wave 3. And those subjects in the
control group smoked more from baseline to wave 3.
Figure 14 presents the growth curve of refusal self-efficacy by gender and genetic
type. The refusal self-efficacy ability in all groups decreased across three waves.

92
0.5
0.6
0.7
0.8
0.9
1
Control & 3 Repeats
Program & 3 Repeats
Control & 4 Repeats
Program & 4 Repeats
Control & Genotype 3/3 or 3/4
Program & Genotype 3/3 or 3/4
Control & Genotype 4/4
Program & Genotype 4/4
Score of Refusal Self-efficacy
Baseline
Wave 2
Wave 3

Figure 14: Growth Curve of Refusal Self-efficacy across Waves by Prevention Program and
Genetic Type (Boys with 3 or 4 repeats and Girls with 3/3, 3/4 or 4/4 genotype)

2) Social norm as the mediator
In the second model, social norm was treated as a mediator in LGCM. The same
set of covariates was adjusted. The results were presented in the table 32 in Appendix.
The tests for the model fit showed that it is a good fit (CFI = .950, RMSEA =
.032). The model is acceptable. Figure 15 presents the growth curve of social norm
by gender and genetic type. The score of social norm in all groups increased across
three waves.

93
1.5
2
2.5
3
3.5
Control & 3 Repeats
Program & 3 Repeats
Control & 4 Repeats
Program & 4 Repeats
Control & Genotype 3/3 or 3/4
Program & Genotype 3/3 or 3/4
Control & Genotype 4/4
Program & Genotype 4/4
Score of Social Norms
Baseline
Wave 2
Wave 3

Figure 15: Growth Curve of Social Norm across Waves by Prevention Program and Genetic Type
(Boys with 3 or 4 repeats and Girls with 3/3, 3/4 or 4/4 genotype)

3) Attitude to cigarette use as the mediator
The third LGCM model was tested with attitude to cigarette use as the time
varying mediator. The results were presented in the table 33 in Appendix. The tests
for the model fit showed that it is a good fit (CFI = .942, RMSEA = .035). The model
is acceptable as well.

94
1.5
1.6
1.7
1.8
1.9
2
Control &
3 Repeats
Control &
4 Repeats
Program &
Genotype
3/3 or 3.4
Program &
Genotype
4/4
Score of Attitude to Cigarette Use
Baseline
Wave 2
Wave 3

Figure 16: Growth Curve of Attitude to Cigarette Use across Waves by Prevention Program and
Genetic Type (Boys with 3 or 4 repeats and Girls with 3/3, 3/4 or 4/4 genotype)

Figure 16 presents the growth curve of attitude to cigarette use by gender and
genetic type. The attitude to cigarette use in all groups for females was improved
across three waves. For boys except the group with 4 repeats in the control group,
their attitude to cigarette use got improved also. And the boys with 4 repeats in the
program group got biggest improvement in their attitude to cigarette use.

Latent Growth Curve Analysis Stratified by Gender Only
To formally testing the genetic influences on the initial status and slope of
smoking progression and potential mediator, additional Latent Growth Curve
Analyses were conducted based on boys and girls, respectively. The genetic factor
and the interaction terms among genetic factor, program condition, depression and
hostility/family/conflict were adjusted as covariates in these additional analyses.

95
1) Refusal self-efficacy as the mediator
The first LGCM model was tested with refusal self-efficacy as the mediator for
boys and girls, respectively. The tests for the model fit showed that it is a good fit for
both boys (CFI = 0.938, RMSEA = .033) and girls (CFI =0 .970, RMSEA=0.019).
The models are acceptable. The results were presented in the table 23 below. Its full
estimates were in table 34 in Appendix.
From the table 23, we know that, the slope of RSE is significantly associated with
the slope of SMK for boys (p=001). The bigger increase of refusal self-efficacy a boy
has, the less increase of smoking progression. Depression is significantly associated
with the slope of SMK (p=0.050).
Table 23: Major Estimated Parameters from the Model with Refusal Self-efficacy as Mediator by
Gender
Boy Girl
Estimate SE p Estimate SE p
Intercept of SMK

Intercept of RSE
-1.575 0.388 0.000 *** -0.384 0.095 0.000 ***
Slope of RSE
1.183 0.607 0.051 + 0.543 0.425 0.201
Slope of SMK

Intercept of RSE
0.061 0.298 0.838 0.114 0.086 0.184
Slope of RSE
-2.185 0.637 0.001 ** -0.674 0.397 0.089 +
Effects on Intercept of SMK

Intervention Program
-0.003 0.261 0.991 0.030 0.054 0.578
Depression
0.368 0.255 0.149 0.012 0.036 0.746
Program X Depression
-0.130 0.157 0.408 -0.017 0.024 0.480
Hostility/family conflict
-0.189 0.504 0.707 0.023 0.074 0.759
Program X Hostility
-0.082 0.262 0.754 0.011 0.051 0.829
Genetic Variant
0.116 0.193 0.549 0.001 0.043 0.988
Genetic V. X Depression
-0.197 0.137 0.150 0.004 0.026 0.871
Genetic V. X Hostility
0.089 0.276 0.748 0.003 0.051 0.956
Genetic V. X Program
0.029 0.146 0.842 -0.015 0.031 0.626
Effects on Slope of SMK

Intervention Program
0.258 0.262 0.324 -0.054 0.044 0.219
Depression
-0.389 0.199 0.050 * -0.003 0.028 0.918
Program X Depression
0.103 0.129 0.424 0.032 0.020 0.118

96
Table 23: Continued

Boy Girl
Estimate SE p Estimate SE p
Hostility/family conflict
0.079 0.442 0.859 -0.040 0.064 0.533
Program X Hostility
-0.176 0.251 0.484 -0.031 0.043 0.476
Genetic Variant
-0.153 0.180 0.393 -0.003 0.035 0.927
Genetic V. X Depression
0.220 0.118 0.061 + -0.010 0.020 0.621
Genetic V. X Hostility
-0.010 0.243 0.967 0.012 0.047 0.793
Genetic V. X Program
-0.190 0.132 0.150 0.020 0.030 0.495
Effects on Intercept of RSE

Intervention Program
0.060 0.077 0.440 -0.049 0.063 0.434
Depression
-0.082 0.056 0.144 -0.026 0.040 0.505
Program X Depression
-0.029 0.036 0.426 -0.028 0.025 0.253
Hostility/family conflict
-0.174 0.118 0.138 -0.188 0.102 0.064 +
Program X Hostility
0.103 0.074 0.161 0.023 0.063 0.720
Genetic Variant
-0.022 0.055 0.683 -0.054 0.059 0.357
Genetic V. X Depression
0.030 0.036 0.404 0.012 0.031 0.684
Genetic V. X Hostility
0.001 0.075 0.991 0.094 0.077 0.222
Genetic V. X Program
-0.045 0.043 0.295 0.049 0.042 0.250
Effects on Slope of RSE

Intervention Program
-0.077 0.062 0.214 -0.020 0.043 0.643
Depression
0.028 0.045 0.537 0.012 0.028 0.668
Program X Depression
0.009 0.028 0.758 -0.002 0.017 0.907
Hostility/family conflict
0.123 0.089 0.170 -0.008 0.066 0.903
Program X Hostility
-0.037 0.056 0.503 0.019 0.041 0.644
Genetic Variant
0.015 0.044 0.728 -0.019 0.039 0.628
Genetic V. X Depression
-0.011 0.029 0.705 -0.013 0.022 0.563
Genetic V. X Hostility
-0.055 0.058 0.342 -0.010 0.051 0.849
Genetic V. X Program
0.065 0.032 0.047 * 0.022 0.030 0.478
+: p<0.10; *: p<0.05; **: p<0.01; ***: p<0.001

For boys, the interaction effect between genetic variants and program on RSE is
significant (p=0.047). The figure below showed the change of refusal self-efficacy
from baseline to two-year follow-up by program condition and allele size.

97
0.5
0.6
0.7
0.8
0.9
Baseline Wave 2 Wave 3
Score of Refusal Self-Efficacy
Control & 3 Repeats Program & 3 Repeats
Control & 4 Repeats Program & 4 Repeats

Figure 17: Growth Curve of Cigarette Monthly Use across Prevention Program and Allele Size
(Boys)

2) Social norm as the mediator
In the second model, social norm was treated as a mediator in LGCM. The same
set of covariates was adjusted.
The tests for the model fit showed that it is not a very good fit for boys (CFI =
.897, RMSEA = .033), but it is a good fit for girls (CFI = .958, RMSEA = 0.020),
although both models are acceptable. The results were presented in the table 35 in
Appendix.

3) Attitude to cigarette use as the mediator
The third LGCM model was tested with attitude to cigarette use as the mediator.
The models for either boys or girls are acceptable, although the tests for the model fit
P for interaction of allele size X program = 0.047

98
showed that it is not a very good fit for boys (CFI = .889, RMSEA = .035), but it is a
good fit for girls (CFI = .963, RMSEA =0.020). The results were presented in table 36
in Appendix.

From the results, we found that the model fits the data from boys the best with the
refusal self-efficacy treated as a mediator, compared to the models with social norm or
attitude to cigarette use treated as a mediator. For the girls, all of the models fit the
data very well.
We concluded that the initial status and slope of smoking progression of Chinese
adolescents’ were not influenced by genetic variants, directly. Depression and
hostility/family conflict can influence the initial status and slope of smoking
progression and the time-varying motivational factors. And the effects of program and
dispositional attributes on smoking progression and motivational factors might be
moderated by repeat polymorphisms of MAOA.
The results verified that refusal self-efficacy/attitude to cigarette use/social norm
mediate the effects of genetic factor, program condition, and some other
environmental factors on smoking progression for both boys and girls.

99
CHAPTER 5: CONCLUSION

5.1 Conclusion from Three Studies
The three studies of this dissertation, utilizing a longitudinal sample from the
Wuhan Smoking Prevention Trial (WSPT) and its complementary genetic
epidemiologic study, explored the patterns and risk factors for progressions of
adolescent smoking and investigated the potential moderating and mediating
mechanisms among genetic and environmental risk factors.
From Study 1, the association between genetic variants and smoking behavior,
depression, and hostility/family conflict was examined. Chinese male adolescents
with the 3-repeat allele of MAOA smoked more cigarettes per day during last month
than those with 4-repeat allele of MAOA (p=0.019). And they had less increase of
cigarettes use from baseline to the year-one follow-up (p=0.013). We also found that
the females with 3/3 repeat genotype of MAOA also had larger increases in hostility
scores than those with 4/4 repeat genotype in the two-year follow-up (p=0.026).
Study 2 investigated the potential risk and protective factors for smoking
progression among adolescents. Regarding to smoking initiation, for a boy who carries
3-repeat allele of MAOA, perceived education performance (RR=0.95, p=0.047) and
time to watch TV per day (RR=1.049, p=0.003), access to cigarettes (RR=1.109,
p=0.042) are risk factors. For a boys with 4-repeat allele of MAOA, the set of risk
factors include peer smoking (RR=1.049, p=0.005), refusal self-efficacy (RR=0.779,
p=0.024). For a girl with genotype 3/3, the set of risk factors are perceived education

100
performance (RR=0.942, p=0.047), attitude to cigarette use (RR=1.142, p=0.010) and
social norm (RR=1.029, p=0.031). For the girls with genotype 3/4 or 4/4, father
smoking status (RR=1.077, p=0.048), refusal self-efficacy (RR=0.791, p=0.018) have
significant influence on the onset of cigarette use. Regarding to the amount of
cigarette use, for the boys with 4 repeats, those in the program group smoked
significantly fewer than in the control group (p=0.031). It suggests that a different set
of risk/ protective factors exist for those adolescents with 3- or 4- repeat of MAOA
allele. We also found that, among never smokers at baseline, the boys with 3 repeats
is more likely to become a life time ever smokers in the year-one follow-up.
Study 3 applied latent growth curve analysis to investigate the mechanism of the
influences of genetic and environmental factors on the smoking progression. Refusal
self-efficacy was treated as a mediator. Intercept of refusal self-efficacy is
significantly associated with initial status of smoking progression for all groups of
subjects. For boys with 3 repeats, slope of refusal self-efficacy is significantly
associated with initial status of smoking progression as well (p=0.039). The slope of
refusal self-efficacy is significantly associated with the slope of smoking progression
for boys (3 repeats: p=0.013; 4 repeats: p=0.005). Several risk/protective factors can
influence the smoking progression by impacting the initial status and linear growth of
refusal self-efficacy. Perceived social norm and attitude to cigarette use were treated
as a mediator, respectively, in additional models as well. The results proved that
refusal self-efficacy/attitude to cigarette use/social norm can mediate the effects of
genetic factor, peer smoking, program condition, and other environmental factors on
smoking progression for both boys and girls.

101
These findings suggest that the genetic polymorphism of MAOA is associated
with Chinese adolescent smoking progression.

5.2 Strengths
Understanding the risk factors, including genetic, psychosocial and other
environmental factors, for adolescent smoking progressions, is important for
identifying adolescents with high risk, and designing efficient prevention curriculum
for deterring smoking development and nicotine dependence.
The gene environment interaction can explain partially the differences of
prevalence of tobacco smoking and smoking progression patterns between different
groups defined by intervention program, or other psychosocial characteristics. It
occurs when the effect of an environmental factor on smoking progression is
conditional on a person’s genotype (or conversely, when environmental experience
moderates genes’ effects on smoking progression). To investigate the gene
environment interaction, abundant genetic and environmental data are needed.
The strengths of conducted studies include 1) WSPT is a well designed school-
based study; 2) abundant genetic and environmental data have been collected from
WSPT and its complementary project; 3) longitudinal data make the research on the
adolescent smoking progression possible.

102
5.3 Limitations
Several limitations in the design of WSPT need to be considered. First, all the
measures including cigarette use, exposure to cigarette advertising, having tobacco
promotion items and other smoking correlates in data are based on self-report. No data
of smoking utilized in the dissertation were obtained from biological samples, such as
saliva or expired air, as bogus pipeline verification (Hansen, Malotte, & Fielding,
1985; Murray & Perry, 1987), or from other sources such as reports of family
members or friends. Underreporting of smoking may have happened. Nevertheless,
studies showed that self reports of smoking cigarettes by adolescents are usually
reliable and valid compared with biological measures, and the validity of self-reports
of smoking was generally comparable across African-American, Hispanic, and White
adolescents (Oetting & Beauvais, 1990; Wills & Cleary, 1997). Accuracy of self-
reporting cigarette use among adolescents is mainly determined by whether they are
assured anonymity and confidentiality (Dolcini, Adler, & Ginsberg, 1996; Hansen,
Malotte, & Fielding, 1985; Murray & Perry, 1987; Williams et al., 1979). Gilpin et al.
(1994) and Pierce, Gilpin, et al. (1998) indicated that the estimates of smoking
prevalence using random-digit dialed telephone surveys were reliable as well.
Genes do not have a continuous effect in our bodies. They may be turned on and
off, both during our overall development and within the lifetime of an individual cell.
Multiple genetic factors from different genes might exist for adolescent smoking
progression. They may interact with each other and have different effects depending
on which other factors are present in the individual’s genotype. The three studies in
dissertation focus on Monoamine Oxidase A only. However, numerous population-

103
based association studies demonstrated the effects of a number of candidate genes,
such as dopamine receptor (DR) and transporter genes (Batra, Patkar, Berrettini,
Weinstein, & Leone, 2003; Duggirala, Almasy, & Blangero, 1999; Erblich, Lerman,
Self, Diaz, & Bovbjerg, 2004), serotonin transporter and nicotinic acetylcholine
receptor (Ishikawa et al., 1999; Lerman et al., 1998), cytochrome P450 (Batra, Patkar,
Berrettini, Weinstein, & Leone, 2003; Walton, Johnstone, Munafo, Neville, &
Griffiths, 2001), on smoking behavior. It showed that multiple genes might be
associated with adolescent smoking behavior. Therefore, possible gene-gene
interaction on smoking progression should be investigated in the further analysis.
Genetic heterogeneity may also make it difficult to ensure phenotype
homogeneity. Genetic heterogeneity, which occurs when mutations in many different
genes can all cause a similar or even identical phenotype, can confound efforts to
collect a genetically similar population.
Estimates of statistical techniques are useful in understanding the relative
contribution of different types of influence and their relation to each other. They are
also useful for understanding why smoking progression and dispositional attributes are
associated with. They do not, however, lead directly to predictive information
regarding individuals, nor do they give reliable estimates of how strongly predictive a
genetic test might be if it were developed.
The depression and hostility/family conflicts can be complicated by the fact that
symptomatic expression. Their measure, interpretation and social response to them
can differ between cultural groups. Especially for the measure of hostility/family
conflict in these three studies, its interpretation is not very clear yet. According to

104
previous studies, aggressiveness is associated with MAOA with the 3-repeat allele
carriers are more aggressive. My assumption is that hostility/family conflict is a type
of dispositional attributes similar to aggressiveness. But the present study showed that
the 3-repeat allele carriers have marginally lower scores of hostility/family conflict at
baseline (p=0.097), although it also demonstrated that the 3-repeat allele carriers have
bigger increase of scores of hostility/family conflict. If hostility/family conflict is
similar to aggressiveness, why do we have the different result from the previous
research? We know that Overt aggression scale (OAS) is a scale used to measure
aggressiveness in human studies(Zammit et al., 2004). The aggressive behavior
measured with this scale indicates a dispositional attributes for those subjects showed
to the environment, which might be mainly outside of their families, however, the
hostility/family conflict is a dispositional attributes for the participants showed only
inside the family. Although both of aggressiveness and hostility/family conflict might
be associated with genetic variants of MAOA, the directions of genetic influences can
be opposite. Further researches of hostility/family conflict are needed.
The results in the present study also suggest that the male adolescents who have 3-
repeat allele of MAOA smoked more number of cigarettes at 7th grade. The findings
can help us to understand the mechanism of adolescent smoking progression.
However, it has ethical limitation on the potential implications for the design of
tobacco control programs in China, even if we can identify the carriers with 3-repeat
allele of MAOA. The major implication is on the development of more specific
phenotypes to increase the genetic signal. For example, according to the study 1 in the
dissertation study, the association between the repeat polymorphisms of MAOA and

105
hostility/family conflict was found only for girls. However, we know, from many
previous animal/human studies, the repeat polymorphisms of MAOA is associated
with aggressiveness. Therefore, we think the hostility/family conflict might be a good
construct to measure the aggressiveness for female Chinese adolescents. For male
Chinese adolescents, more research are needed to develop the construct to measure the
aggressiveness.

106
BIBLIOGRAPHY

Abell, C. W. (1987). Monoamine oxidase A and B from human liver and brain. Methods Enzymol, 142,
638-650.
Ajzen, I. (1988). Attitudes, personality, and behavior (U.S. ed.). Chicago, IL: Dorsey Press.
Andrews, J. A., & Duncan, S. C. (1998). The effect of attitude on the development of adolescent
cigarette use. J Subst Abuse, 10(1), 1-7.
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. American
Journal of 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.
Ausems, M., Mesters, I., van Breukelen, G., & De Vries, H. (2003). Do Dutch 11-12 years olds who
never smoke, smoke experimentally or smoke regularly have different demographic
backgrounds and perceptions of smoking? European Journal of Public Health, 13(2), 160-167.
Bandura, A. (1977). Social learning theory. Englewood Cliffs, N.J.: Prentice Hall.
Bandura, A. (1986). Social foundations of thought and action : a social cognitive theory. Englewood
Cliffs, N.J.: Prentice-Hall.
Baron, R. M., & Kenny, D. A. (1986). The moderator-mediator variable distinction in social
psychological research: conceptual, strategic, and statistical considerations. Journal of
Personality & Social Psychology, 51(6), 1173-1182.
Batra, V., Patkar, A. A., Berrettini, W. H., Weinstein, S. P., & Leone, F. T. (2003). The genetic
determinants of smoking.[see comment]. Chest, 123(5), 1730-1739.
Bauman, K. E., LaPrelle, J., Brown, J. D., Koch, G. G., & Padgett, C. A. (1991). The influence of three
mass media campaigns on variables related to adolescent cigarette smoking: results of a field
experiment. American Journal of Public Health, 81(5), 597-604.
Bierut, L. J., Dinwiddie, S. H., Begleiter, H., Crowe, R. R., Hesselbrock, V., Nurnberger, J. I., Jr., et al.
(1998). Familial transmission of substance dependence: alcohol, marijuana, cocaine, and
habitual smoking: a report from the Collaborative Study on the Genetics of Alcoholism. Arch
Gen Psychiatry, 55(11), 982-988.
Bledsoe, L. K. (2006). Smoking cessation: an application of theory of planned behavior to
understanding progress through stages of change. Addict Behav, 31(7), 1271-1276.

107
Botvin, G. J., & Botvin, E. M. (1992). Adolescent tobacco, alcohol, and drug abuse: prevention
strategies, empirical findings, and assessment issues. Journal of Developmental & Behavioral
Pediatrics, 13(4), 290-301.
Botvin, G. J., Goldberg, C. J., Botvin, E. M., & Dusenbury, L. (1993). Smoking behavior of adolescents
exposed to cigarette advertising. Public Health Reports, 108(2), 217-224.
Braverman, M. T., & Aaro, L. E. (2004). Adolescent smoking and exposure to tobacco marketing under
a tobacco advertising ban: findings from 2 Norwegian national samples. American Journal of
Public Health, 94(7), 1230-1238.
Bricker, J. B., Peterson, A. V., Jr., Andersen, M. R., Rajan, K. B., Leroux, B. G., & Sarason, I. G.
(2005). Childhood friends who smoke: Do they influence adolescents to make smoking
transitions? Addict Behav.
Burt, R. D., Dinh, K. T., Peterson, A. V., Jr., & Sarason, I. G. (2000). Predicting adolescent smoking: a
prospective study of personality variables. Preventive Medicine, 30(2), 115-125.
Carmelli, D., Swan, G. E., Robinette, D., & Fabsitz, R. (1992). Genetic influence on smoking--a study
of male twins.[see comment]. New England Journal of Medicine, 327(12), 829-833.
Cases, O., Seif, I., Grimsby, J., Gaspar, P., Chen, K., Pournin, S., et al. (1995). Aggressive behavior and
altered amounts of brain serotonin and norepinephrine in mice lacking MAOA.[see comment].
Science, 268(5218), 1763-1766.
Caspi, A., McClay, J., Moffitt, T. E., Mill, J., Martin, J., Craig, I. W., et al. (2002). Role of genotype in
the cycle of violence in maltreated children. Science, 297(5582), 851-854.
Catalano, R. F., Haggerty, K. P., Oesterle, S., Fleming, C. B., & Hawkins, J. D. (2004). The importance
of bonding to school for healthy development: findings from the Social Development Research
Group. J Sch Health, 74(7), 252-261.
CDC. (2002). Annual smoking-attributable mortality, years of potential life lost, and economic costs--
United States, 1995-1999. MMWR - Morbidity & Mortality Weekly Report, 51(14), 300-303.
CDC. (2005). Tobacco use, access, and exposure to tobacco in media among middle and high school
students--United States, 2004. MMWR - Morbidity & Mortality Weekly Report, 54(12), 297-
301.
Chao, J., & Nestler, E. J. (2004). Molecular neurobiology of drug addiction. Annual Review of
Medicine, 55, 113-132.
Chen, X., & Unger, J. B. (1999). Hazards of smoking initiation among Asian American and non-Asian
adolescents in California: a survival model analysis. Preventive Medicine, 28(6), 589-599.

108
Chen, X., Unger, J. B., & Johnson, C. A. (1999). Is acculturation a risk factor for early smoking
initiation among Chinese American minors? A comparative perspective. Tobacco Control,
8(4), 402-410.
Choi, W. S., Harris, K. J., Okuyemi, K., & Ahluwalia, J. S. (2003). Predictors of smoking initiation
among college-bound high school students. Annals of Behavioral Medicine, 26(1), 69-74.
Choi, W. S., Pierce, J. P., Gilpin, E. A., Farkas, A. J., & Berry, C. C. (1997). Which adolescent
experimenters progress to established smoking in the United States. American Journal of
Preventive Medicine, 13(5), 385-391.
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.
Cloninger, C. R. (1988). A unified biosocial theory of personality and its role in the development of
anxiety states: a reply to commentaries. Psychiatric Developments, 6(2), 83-120.
Cohen, I. L., Liu, X., Schutz, C., White, B. N., Jenkins, E. C., Brown, W. T., et al. (2003). Association
of autism severity with a monoamine oxidase A functional polymorphism. Clin Genet, 64(3),
190-197.
Colder, C. R., Mehta, P., Balanda, K., Campbell, R. T., Mayhew, K. P., Stanton, W. R., et al. (2001).
Identifying trajectories of adolescent smoking: an application of latent growth mixture
modeling. Health Psychology, 20(2), 127-135.
Comings, D. E. (1998). Polygenic inheritance and micro/minisatellites.[see comment]. Molecular
Psychiatry, 3(1), 21-31.
Comings, D. E., & Blum, K. (2000). Reward deficiency syndrome: genetic aspects of behavioral
disorders. Progress in Brain Research, 126, 325-341.
Comings, D. E., Ferry, L., Bradshaw-Robinson, S., Burchette, R., Chiu, C., & Muhleman, D. (1996).
The dopamine D2 receptor (DRD2) gene: a genetic risk factor in smoking. Pharmacogenetics,
6(1), 73-79.
Corrigall, W. A., Franklin, K. B., Coen, K. M., & Clarke, P. B. (1992). The mesolimbic dopaminergic
system is implicated in the reinforcing effects of nicotine. Psychopharmacology, 107(2-3),
285-289.
Critchley, J., & Capewell, S. (2004). Smoking cessation for the secondary prevention of coronary heart
disease.[update of Cochrane Database Syst Rev. 2003;(4):CD003041; PMID: 14583958].
Cochrane Database of Systematic Reviews(1), CD003041.
Dierker, L. C., Avenevoli, S., Stolar, M., & Merikangas, K. R. (2002). Smoking and depression: an
examination of mechanisms of comorbidity. American Journal of Psychiatry, 159(6), 947-953.

109
Duggirala, R., Almasy, L., & Blangero, J. (1999). Smoking behavior is under the influence of a major
quantitative trait locus on human chromosome 5q. Genetic Epidemiology, 17 Suppl 1, S139-
144.
Epstein, J. A., Griffin, K. W., & Botvin, G. J. (2000). A model of smoking among inner-city
adolescents: the role of personal competence and perceived social benefits of smoking.
Preventive Medicine, 31(2 Pt 1), 107-114.
Erblich, J., Lerman, C., Self, D. W., Diaz, G. A., & Bovbjerg, D. H. (2004). Stress-induced cigarette
craving: effects of the DRD2 TaqI RFLP and SLC6A3 VNTR polymorphisms.
Pharmacogenomics Journal, 4(2), 102-109.
Etter, J. F., Pelissolo, A., Pomerleau, C., & De Saint-Hilaire, Z. (2003). Associations between smoking
and heritable temperament traits. Nicotine & Tobacco Research, 5(3), 401-409.
Everett, S. A., Warren, C. W., Sharp, D., Kann, L., Husten, C. G., & Crossett, L. S. (1999). Initiation of
cigarette smoking and subsequent smoking behavior among U.S. high school students.
Preventive Medicine, 29(5), 327-333.
Fagan, P., Eisenberg, M., Stoddard, A. M., Frazier, L., & Sorensen, G. (2001). Social influences, social
norms, social support, and smoking behavior among adolescent workers. American Journal of
Health Promotion, 15(6), 414-421.
Fergusson, D. M., & Horwood, L. J. (1995). Transitions to cigarette smoking during adolescence.
Addictive Behaviors, 20(5), 627-642.
Fishbein, M., & Ajzen, I. (1975). Belief, attitude, intention, and behavior : an introduction to theory
and research. Reading, Mass.: Addison-Wesley Pub. Co.
Flay, B. (1993). Youth tobacco use: risk patterns and control. In C. T. Orleans & J. Slade (Eds.),
Nicotine addiction : principles and management (pp. 653-661). New York: Oxford University
Press.
Flay, B. R., Hu, F. B., Siddiqui, O., Day, L. E., Hedeker, D., Petraitis, J., et al. (1994). Differential
influence of parental smoking and friends' smoking on adolescent initiation and escalation of
smoking. Journal of Health & Social Behavior, 35(3), 248-265.
Flay, B. R., Phil, D., Hu, F. B., & Richardson, J. (1998). Psychosocial predictors of different stages of
cigarette smoking among high school students. Preventive Medicine, 27(5 Pt 3), A9-18.
Forster, J. L., Wolfson, M., Murray, D. M., Wagenaar, A. C., & Claxton, A. J. (1997). Perceived and
measured availability of tobacco to youths in 14 Minnesota communities: the TPOP Study.
Tobacco Policy Options for Prevention. American Journal of Preventive Medicine, 13(3), 167-
174.
Gerber, R. W., Newman, I. M., & Martin, G. L. (1988). Applying the theory of reasoned action to early
adolescent tobacco chewing. J Sch Health, 58(10), 410-413.

110
Gerra, G., Avanzini, P., Zaimovic, A., Sartori, R., Bocchi, C., Timpano, M., et al. (1999).
Neurotransmitters, neuroendocrine correlates of sensation-seeking temperament in normal
humans. Neuropsychobiology, 39(4), 207-213.
Glassman, A. H., Helzer, J. E., Covey, L. S., Cottler, L. B., Stetner, F., Tipp, J. E., et al. (1990).
Smoking, smoking cessation, and major depression. JAMA, 264(12), 1546-1549.
Goldstein, R. Z., & Volkow, N. D. (2002). Drug addiction and its underlying neurobiological basis:
neuroimaging evidence for the involvement of the frontal cortex. American Journal of
Psychiatry, 159(10), 1642-1652.
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.
Preventive Medicine, 19(3), 305-313.
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 Tob Res, 8(2), 245-255.
Guo, Q., Johnson, C. A., Unger, J. B., Lee, L., Xie, B., Chou, C. P., et al. (2006). Utility of the theory of
reasoned action and theory of planned behavior for predicting Chinese adolescent smoking.
Addict Behav.
Hallman, J., Sakurai, E., & Oreland, L. (1990). Blood platelet monoamine oxidase activity, serotonin
uptake and release rates in anorexia and bulimia patients and in healthy controls. Acta
Psychiatrica Scandinavica, 81(1), 73-77.
Hansen, W. B., Johnson, C. A., Flay, B. R., Graham, J. W., & Sobel, J. (1988). Affective and social
influences approaches to the prevention of multiple substance abuse among seventh grade
students: results from project SMART. Preventive Medicine, 17(2), 135-154.
Harakeh, Z., Scholte, R. H., Vermulst, A. A., de Vries, H., & Engels, R. C. (2004). Parental factors and
adolescents' smoking behavior: an extension of The theory of planned behavior. Prev Med,
39(5), 951-961.
Harrell, J. S., Bangdiwala, S. I., Deng, S., Webb, J. P., & Bradley, C. (1998). Smoking initiation in
youth: the roles of gender, race, socioeconomics, and developmental status. Journal of
Adolescent Health, 23(5), 271-279.
Hawkins, J. D., Catalano, R. F., & Miller, J. Y. (1992). Risk and protective factors for alcohol and other
drug problems in adolescence and early adulthood: implications for substance abuse
prevention. Psychological Bulletin, 112(1), 64-105.
Hazuda, H. P., Haffner, S. M., Stern, M. P., & Eifler, C. W. (1988). Effects of acculturation and
socioeconomic status on obesity and diabetes in Mexican Americans. The San Antonio Heart
Study. American Journal of Epidemiology, 128(6), 1289-1301.
Heath, A. C., & Martin, N. G. (1993). Genetic models for the natural history of smoking: evidence for a
genetic influence on smoking persistence. Addictive Behaviors, 18(1), 19-34.

111
Heishman, S. J., Balfour, D. J., Benowitz, N. L., Hatsukami, D. K., Lindstrom, J. M., & Ockene, J. K.
(1997). Society for Research on Nicotine and Tobacco. Addiction, 92(5), 615-633.
Herman, A. I., Kaiss, K. M., Ma, R., Philbeck, J. W., Hasan, A., Dasti, H., et al. (2005). Serotonin
transporter promoter polymorphism and monoamine oxidase type A VNTR allelic variants
together influence alcohol binge drinking risk in young women. Am J Med Genet B
Neuropsychiatr Genet, 133(1), 74-78.
Ishikawa, H., Ohtsuki, T., Ishiguro, H., Yamakawa-Kobayashi, K., Endo, K., Lin, Y. L., et al. (1999).
Association between serotonin transporter gene polymorphism and smoking among Japanese
males. Cancer Epidemiology, Biomarkers & Prevention, 8(9), 831-833.
Jackson, C., Henriksen, L., Dickinson, D., Messer, L., & Robertson, S. B. (1998). A longitudinal study
predicting patterns of cigarette smoking in late childhood. Health Education & Behavior,
25(4), 436-447.
Jin, Y., Chen, D., Hu, Y., Guo, S., Sun, H., Lu, A., et al. (2006). Association between monoamine
oxidase gene polymorphisms and smoking behaviour in Chinese males. Int J
Neuropsychopharmacol, 9(5), 557-564.
Johansson, F., Von Knorring, L., & Oreland, L. (1983). Platelet MAO activity in patients with chronic
pain syndrome. Relationship to personality traits, endorphins in CSF and plasma cortisol.
Medical Biology, 61(2), 101-104.
Johnson, J. G., Cohen, P., Pine, D. S., Klein, D. F., Kasen, S., & Brook, J. S. (2000). Association
between cigarette smoking and anxiety disorders during adolescence and early adulthood.[see
comment]. JAMA, 284(18), 2348-2351.
Josendal, O., Aaro, L. E., & Bergh, I. H. (1998). Effects of a school-based smoking prevention program
among subgroups of adolescents. Health Education Research, 13(2), 215-224.
Kandel, D., Chen, K., Warner, L. A., Kessler, R. C., & Grant, B. (1997). Prevalence and demographic
correlates of symptoms of last year dependence on alcohol, nicotine, marijuana and cocaine in
the U.S. population. Drug & Alcohol Dependence, 44(1), 11-29.
Kendler, K. S., Neale, M. C., Sullivan, P., Corey, L. A., Gardner, C. O., & Prescott, C. A. (1999). A
population-based twin study in women of smoking initiation and nicotine dependence.
Psychological Medicine, 29(2), 299-308.
Kessler, D. A., Natanblut, S. L., Wilkenfeld, J. P., Lorraine, C. C., Mayl, S. L., Bernstein, I. B., et al.
(1997). Nicotine addiction: a pediatric disease. Journal of Pediatrics, 130(4), 518-524.
Klitzner, M., Gruenewald, P. J., & Bamberger, E. (1991). Cigarette advertising and adolescent
experimentation with smoking. British Journal of Addiction, 86(3), 287-298.
Kobus, K. (2003). Peers and adolescent smoking. Addiction, 98 Suppl 1, 37-55.

112
Koepke, D., Flay, B. R., & Johnson, C. A. (1990). Health behaviors in minority families: The case of
cigarette smoking. Family Community Health, 13(1), 35-43.
Koob, G. F., & Nestler, E. J. (1997). The neurobiology of drug addiction. Journal of Neuropsychiatry &
Clinical Neurosciences, 9(3), 482-497.
Koopmans, J. R., Slutske, W. S., Heath, A. C., Neale, M. C., & Boomsma, D. I. (1999). The genetics of
smoking initiation and quantity smoked in Dutch adolescent and young adult twins. Behavior
Genetics, 29(6), 383-393.
Koval, J. J., & Pederson, L. L. (1999). Stress-coping and other psychosocial risk factors: a model for
smoking in grade 6 students. Addict Behav, 24(2), 207-218.
Langlois, M. A., Petosa, R., & Hallam, J. S. (1999). Why do effective smoking prevention programs
work? Student changes in social cognitive theory constructs. J Sch Health, 69(8), 326-331.
Laviolette, S. R., & van der Kooy, D. (2003). Blockade of mesolimbic dopamine transmission
dramatically increases sensitivity to the rewarding effects of nicotine in the ventral tegmental
area. Molecular Psychiatry, 8(1), 50-59.
Lerman, C., Caporaso, N. E., Audrain, J., Main, D., Bowman, E. D., Lockshin, B., et al. (1999).
Evidence suggesting the role of specific genetic factors in cigarette smoking.[see comment].
Health Psychology, 18(1), 14-20.
Lerman, C., Shields, P. G., Audrain, J., Main, D., Cobb, B., Boyd, N. R., et al. (1998). The role of the
serotonin transporter gene in cigarette smoking. Cancer Epidemiology, Biomarkers &
Prevention, 7(3), 253-255.
Lessov, C. N., Swan, G. E., Ring, H. Z., Khroyan, T. V., & Lerman, C. (2004). Genetics and drug use
as a complex phenotype. Subst Use Misuse, 39(10-12), 1515-1569.
Leventhal, H., & Cleary, P. D. (1980). The smoking problem: a review of the research and theory in
behavioral risk modification. Psychological Bulletin, 88(2), 370-405.
Li, M. D. (2003). The genetics of smoking related behavior: a brief review. American Journal of the
Medical Sciences, 326(4), 168-173.
Lloyd-Richardson, E. E., Papandonatos, G., Kazura, A., Stanton, C., & Niaura, R. (2002).
Differentiating stages of smoking intensity among adolescents: stage-specific psychological
and social influences. Journal of Consulting & Clinical Psychology, 70(4), 998-1009.
Lovato, C., Linn, G., Stead, L. F., & Best, A. (2003). Impact of tobacco advertising and promotion on
increasing adolescent smoking behaviours. Cochrane Database of Systematic Reviews(4),
CD003439.

113
Lu, R. B., Lin, W. W., Lee, J. F., Ko, H. C., & Shih, J. C. (2003). Neither antisocial personality disorder
nor antisocial alcoholism is associated with the MAO-A gene in Han Chinese males.
Alcoholism: Clinical & Experimental Research, 27(6), 889-893.
Madden, P. A., Pedersen, N. L., Kaprio, J., Koskenvuo, M. J., & Martin, N. G. (2004). The
epidemiology and genetics of smoking initiation and persistence: crosscultural comparisons of
twin study results. Twin Research, 7(1), 82-97.
Mannino, D. M. (2003). Chronic obstructive pulmonary disease: definition and epidemiology.
Respiratory Care, 48(12), 1185-1191; discussion 1191-1183.
Martin, R. A., Velicer, W. F., & Fava, J. L. (1996). Latent transition analysis to the stages of change for
smoking cessation. Addictive Behaviors, 21(1), 67-80.
McGahee, T. W., Kemp, V., & Tingen, M. (2000). A theoretical model for smoking prevention studies
in preteen children. Pediatr Nurs, 26(2), 135-138, 141.
McGinnis, J. M., & Foege, W. H. (1993). Actual causes of death in the United States.[see comment].
JAMA, 270(18), 2207-2212.
McMahon, R. J. (1999). Child and adolescent psychopathology as risk factors for subsequent tobacco
use. Nicotine & Tobacco Research, 1 Suppl 2, S45-50; discussion S69-70.
Mermelstein, R. (1999). Explanations of ethnic and gender differences in youth smoking: a multi-site,
qualitative investigation. The Tobacco Control Network Writing Group. Nicotine & Tobacco
Research, 1 Suppl 1, S91-98.
Meyer-Lindenberg, A., Buckholtz, J. W., Kolachana, B., A, R. H., Pezawas, L., Blasi, G., et al. (2006).
Neural mechanisms of genetic risk for impulsivity and violence in humans. Proc Natl Acad Sci
U S A, 103(16), 6269-6274.
Munafo, M., Clark, T., Johnstone, E., Murphy, M., & Walton, R. (2004). The genetic basis for smoking
behavior: a systematic review and meta-analysis. Nicotine Tob Res, 6(4), 583-597.
Murray, D. M., Varnell, S. P., & Blitstein, J. L. (2004). Design and analysis of group-randomized trials:
a review of recent methodological developments. American Journal of Public Health, 94(3),
423-432.
Muthen, B., & Shedden, K. (1999). Finite mixture modeling with mixture outcomes using the EM
algorithm. Biometrics, 55(2), 463-469.
Nezami, E., Unger, J., Tan, S., Mahaffey, C., Ritt-Olson, A., Sussman, S., et al. (2005). The influence
of depressive symptoms on experimental smoking and intention to smoke in a diverse youth
sample. Nicotine & Tobacco Research, 7(2), 243-248.

114
Nilsson, K. W., Sjoberg, R. L., Damberg, M., Leppert, J., Ohrvik, J., Alm, P. O., et al. (2006). Role of
monoamine oxidase A genotype and psychosocial factors in male adolescent criminal activity.
Biol Psychiatry, 59(2), 121-127.
Nilsson, W. K. (2006). Gene-Environment Interaction in Adolescent Deviant Behavior. Acta
Universitatis Upsaliensis, UPPSALA.
Niu, T., Chen, C., Ni, J., Wang, B., Fang, Z., Shao, H., et al. (2000). Nicotine dependence and its
familial aggregation in Chinese. International Journal of Epidemiology, 29(2), 248-252.
Oreland, L., Hallman, J., & Damberg, M. (2004). Platelet MAO and personality--function and
dysfunction. Curr Med Chem, 11(15), 2007-2016.
Osler, M., Holst, C., Prescott, E., & Sorensen, T. I. (2001). Influence of genes and family environment
on adult smoking behavior assessed in an adoption study. Genetic Epidemiology, 21(3), 193-
200.
Oygard, L., Klepp, K. I., Tell, G. S., & Vellar, O. D. (1995). Parental and peer influences on smoking
among young adults: ten-year follow-up of the Oslo youth study participants. Addiction, 90(4),
561-569.
Passamonti, L., Fera, F., Magariello, A., Cerasa, A., Gioia, M. C., Muglia, M., et al. (2006).
Monoamine oxidase-a genetic variations influence brain activity associated with inhibitory
control: new insight into the neural correlates of impulsivity. Biol Psychiatry, 59(4), 334-340.
Patel, D. R., & Homnick, D. N. (2000). Pulmonary effects of smoking. Adolescent Medicine State of the
Art Reviews, 11(3), 567-576.
Patil, N., Berno, A. J., Hinds, D. A., Barrett, W. A., Doshi, J. M., Hacker, C. R., et al. (2001). Blocks of
limited haplotype diversity revealed by high-resolution scanning of human chromosome
21.[see comment]. Science, 294(5547), 1719-1723.
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.
Paus, T. (2005). Mapping brain maturation and cognitive development during adolescence. Trends in
Cognitive Sciences, 9(2), 60-68.
Pedersen, W., & Lavik, N. J. (1991). Role modelling and cigarette smoking: vulnerable working class
girls? A longitudinal study. Scandinavian Journal of Social Medicine, 19(2), 110-115.
Pederson, L. L., Koval, J. J., McGrady, G. A., & Tyas, S. L. (1998). The degree and type of relationship
between psychosocial variables and smoking status for students in grade 8: is there a dose-
response relationship? Preventive Medicine, 27(3), 337-347.

115
Pentz, M. A., MacKinnon, D. P., Dwyer, J. H., Wang, E. Y., Hansen, W. B., Flay, B. R., et al. (1989).
Longitudinal effects of the midwestern prevention project on regular and experimental
smoking in adolescents. Preventive Medicine, 18(2), 304-321.
Pierce, J. P., Choi, W. S., Gilpin, E. A., Farkas, A. J., & Merritt, R. K. (1996). Validation of
susceptibility as a predictor of which adolescents take up smoking in the United States. Health
Psychology, 15(5), 355-361.
Pierce, J. P., Gilpin, E. A., Emery, S. L., White, M. M., Rosbrook, B., Berry, C. C., et al. (1998). Has
the California tobacco control program reduced smoking? JAMA, 280(10), 893-899.
Pierce, J. P., Lee, L., & Gilpin, E. A. (1994). Smoking initiation by adolescent girls, 1944 through 1988.
An association with targeted advertising.[see comment]. JAMA, 271(8), 608-611.
Prochaska, J. O., & DiClemente, C. C. (1984). The transtheoretical approach : crossing traditional
boundaries of therapy. Homewood, Ill.: Dow Jones-Irwin.
Prokhorov, A. V., Warneke, C., de Moor, C., Emmons, K. M., Mullin Jones, M., Rosenblum, C., et al.
(2003). Self-reported health status, health vulnerability, and smoking behavior in college
students: implications for intervention. Nicotine & Tobacco Research, 5(4), 545-552.
Radloff, L. S. (1977). The CES-D scale: A self-report depression scale for research in the general
population. Applied Psychological Measurement, 1(385-401).
Richardson, J. L., Radziszewska, B., Dent, C. W., & Flay, B. R. (1993). Relationship between after-
school care of adolescents and substance use, risk taking, depressed mood, and academic
achievement.[see comment]. Pediatrics, 92(1), 32-38.
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.
Rius, C., Fernandez, E., Schiaffino, A., Borras, J. M., & Rodriguez-Artalejo, F. (2004). Self perceived
health and smoking in adolescents. Journal of Epidemiology & Community Health, 58(8), 698-
699.
Robinson, L. A., Klesges, R. C., Zbikowski, S. M., & Glaser, R. (1997). Predictors of risk for different
stages of adolescent smoking in a biracial sample. Journal of Consulting & Clinical
Psychology, 65(4), 653-662.
Rohrbach, L. A., Howard-Pitney, B., Unger, J. B., Dent, C. W., Howard, K. A., Cruz, T. B., et al.
(2002). Independent evaluation of the California Tobacco Control Program: relationships
between program exposure and outcomes, 1996-1998. Am J Public Health, 92(6), 975-983.
Sabol, S. Z., Hu, S., & Hamer, D. (1998). A functional polymorphism in the monoamine oxidase A
gene promoter. Hum Genet, 103(3), 273-279.

116
Sabol, S. Z., Nelson, M. L., Fisher, C., Gunzerath, L., Brody, C. L., Hu, S., et al. (1999). A genetic
association for cigarette smoking behavior. Health Psychology, 18(1), 7-13.
SAMHSA. (2005). Results from the 2004 National Survey on Drug Use and Health. Rockville, MD:
Office of Applied Studies.
Santana, Y., Gonzalez, B., Pinilla, J., Calvo, J. R., & Barber, P. (2003). Young adolescents, tobacco
advertising, and smoking. Journal of Drug Education, 33(4), 427-444.
Sargent, J. D., Dalton, M. A., Beach, M. L., Mott, L. A., Tickle, J. J., Ahrens, M. B., et al. (2002).
Viewing tobacco use in movies: does it shape attitudes that mediate adolescent smoking?
American Journal of Preventive Medicine, 22(3), 137-145.
Sargent, J. D., Dalton, M. A., Heatherton, T., & Beach, M. (2003). Modifying exposure to smoking
depicted in movies: a novel approach to preventing adolescent smoking. Archives of Pediatrics
& Adolescent Medicine, 157(7), 643-648.
Schulze, T. G., Muller, D. J., Krauss, H., Scherk, H., Ohlraun, S., Syagailo, Y. V., et al. (2000).
Association between a functional polymorphism in the monoamine oxidase A gene promoter
and major depressive disorder. Am J Med Genet, 96(6), 801-803.
Secker-Walker, R. H., Worden, J. K., Holland, R. R., Flynn, B. S., & Detsky, A. S. (1997). A mass
media programme to prevent smoking among adolescents: costs and cost effectiveness.
Tobacco Control, 6(3), 207-212.
Shakib, S., Mouttapa, M., Johnson, C. A., Ritt-Olson, A., Trinidad, D. R., Gallaher, P. E., et al. (2003).
Ethnic variation in parenting characteristics and adolescent smoking. Journal of Adolescent
Health, 33(2), 88-97.
Shih, J. C., & Chen, K. (2004). Regulation of MAO-A and MAO-B gene expression. Curr Med Chem,
11(15), 1995-2005.
Shih, J. C., Chen, K., & Ridd, M. J. (1999). Role of MAO A and B in neurotransmitter metabolism and
behavior. Pol J Pharmacol, 51(1), 25-29.
Shih, J. C., & Thompson, R. F. (1999). Monoamine oxidase in neuropsychiatry and behavior. American
Journal of Human Genetics, 65(3), 593-598.
Slomkowski, C., Rende, R., Novak, S., Lloyd-Richardson, E., & Niaura, R. (2005). Sibling effects on
smoking in adolescence: evidence for social influence from a genetically informative design.
Addiction, 100(4), 430-438.
Stice, E., & Shaw, H. (2003). Prospective relations of body image, eating, and affective disturbances to
smoking onset in adolescent girls: how Virginia slims. Journal of Consulting & Clinical
Psychology, 71(1), 129-135.

117
Sun, W., Andreeva, V. A., Unger, J. B., Conti, D. V., Chou, C. P., Palmer, P. H., et al. (2006). Age-
related smoking progression among adolescents in China. J Adolesc Health, 39(5), 686-693.
Sun, W., Cook, A. V., Unger, B. J., Conti, D., Sun, P., Chou, C.-P., et al. (In Press). Age-Related
Smoking Progression among Middle and Upper School Students in China. Journal of
adolescence Health.
Sussman, S., Brannon, B. R., Dent, C. W., Hansen, W. B., Johnson, C. A., & Flay, B. R. (1993).
Relations of coping effort, coping strategies, perceived stress, and cigarette smoking among
adolescents. International Journal of the Addictions, 28(7), 599-612.
Swan, G. E. (1999). Multiple risk factors for the initiation of smoking: the public health imperative for
multidisciplinary genetic epidemiological investigations of nicotine addiction. Nicotine &
Tobacco Research, 1 Suppl 2, S71-73; discussion S69-70.
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.
True, W. R., Heath, A. C., Scherrer, J. F., Waterman, B., Goldberg, J., Lin, N., et al. (1997). Genetic
and environmental contributions to smoking. Addiction, 92(10), 1277-1287.
Tsai, S. J., Hong, C. J., Yu, Y. W., & Chen, T. J. (2004). Association study of catechol-O-
methyltransferase gene and dopamine D4 receptor gene polymorphisms and personality traits
in healthy young chinese females. Neuropsychobiology, 50(2), 153-156.
Turner, L., Mermelstein, R., & Flay, B. (2004). Individual and contextual influences on adolescent
smoking. Annals of the New York Academy of Sciences, 1021, 175-197.
Tyas, S. L., & Pederson, L. L. (1998). Psychosocial factors related to adolescent smoking: a critical
review of the literature. Tobacco Control, 7(4), 409-420.
Tyndale, R. F., & Sellers, E. M. (2001). Variable CYP2A6-mediated nicotine metabolism alters
smoking behavior and risk. Drug Metab Dispos, 29(4 Pt 2), 548-552.
Unger, J. B., Johnson, C. A., & Rohrbach, L. A. (1995). Recognition and liking of tobacco and alcohol
advertisements among adolescents: relationships with susceptibility to substance use. Prev
Med, 24(5), 461-466.
Unger, J. B., Johnson, C. A., Stoddard, J. L., Nezami, E., & Chou, C. P. (1997). Identification of
adolescents at risk for smoking initiation: validation of a measure of susceptibility. Addictive
Behaviors, 22(1), 81-91.
Unger, J. B., Rohrbach, L. A., Howard-Pitney, B., Ritt-Olson, A., & Mouttapa, M. (2001). Peer
influences and susceptibility to smoking among California adolescents. Subst Use Misuse,
36(5), 551-571.

118
Unger, J. B., Rohrbach, L. A., Howard, K. A., Boley Cruz, T., Johnson, C. A., & Chen, X. (1999).
Attitudes toward anti-tobacco policy among California youth: associations with smoking
status, psychosocial variables and advocacy actions. Health Education Research, 14(6), 751-
763.
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. Preventive Medicine, 34(4), 476-484.
USDHHS. (1994). Preventing tobacco use among young people. A report of the Surgeon General.
Executive summary. Morbidity & Mortality Weekly Report. Recommendations & Reports,
43(RR-4), 1-10.
van den Eijnden, R. J., Spijkerman, R., & Engels, R. C. (2006). Relative contribution of smoker
prototypes in predicting smoking among adolescents: a comparison with factors from the
theory of planned behavior. Eur Addict Res, 12(3), 113-120.
Vink, J. M., Willemsen, G., Engels, R. C., & Boomsma, D. I. (2003). Smoking status of parents,
siblings and friends: predictors of regular smoking? Findings from a longitudinal twin-family
study. Twin Research, 6(3), 209-217.
Walton, R., Johnstone, E., Munafo, M., Neville, M., & Griffiths, S. (2001). Genetic clues to the
molecular basis of tobacco addiction and progress towards personalized therapy. Trends in
Molecular Medicine, 7(2), 70-76.
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.
Whalen, C. K., Jamner, L. D., Henker, B., & Delfino, R. J. (2001). Smoking and moods in adolescents
with depressive and aggressive dispositions: evidence from surveys and electronic diaries.
Health Psychology, 20(2), 99-111.
Wills, T. A. (1986). Stress and coping in early adolescence: relationships to substance use in urban
school samples. Health Psychol, 5(6), 503-529.
Wills, T. A., Resko, J. A., Ainette, M. G., & Mendoza, D. (2004). Smoking onset in adolescence: a
person-centered analysis with time-varying predictors. Health Psychol, 23(2), 158-167.
Worden, J. K., Flynn, B. S., Solomon, L. J., Secker-Walker, R. H., Badger, G. J., & Carpenter, J. H.
(1996). Using mass media to prevent cigarette smoking among adolescent girls. Health
Education Quarterly, 23(4), 453-468.
Yang, G., Fan, L., Tan, J., Qi, G., Zhang, Y., Samet, J. M., et al. (1999). Smoking in China: findings of
the 1996 National Prevalence Survey. Jama, 282(13), 1247-1253.

119
Yorulmaz, F., Akturk, Z., Dagdeviren, N., & Dalkilic, A. (2002). Smoking among adolescents: relation
to school success, socioeconomic status nutrition and self-esteem. Swiss Medical Weekly,
132(31-32), 449-454.
Yu, Y. W., Tsai, S. J., Hong, C. J., Chen, T. J., Chen, M. C., & Yang, C. W. (2005). Association study
of a monoamine oxidase a gene promoter polymorphism with major depressive disorder and
antidepressant response. Neuropsychopharmacology, 30(9), 1719-1723.
Yu, Y. W., Yang, C. W., Wu, H. C., Tsai, S. J., Hong, C. J., Chen, M. C., et al. (2005). Association
study of a functional MAOA-uVNTR gene polymorphism and personality traits in Chinese
young females. Neuropsychobiology, 52(3), 118-121.
Zammit, S., Jones, G., Jones, S. J., Norton, N., Sanders, R. D., Milham, C., et al. (2004).
Polymorphisms in the MAOA, MAOB, and COMT genes and aggressive behavior in
schizophrenia. Am J Med Genet B Neuropsychiatr Genet, 128(1), 19-20.

120
APPENDIX
Table 24: Results of Multilevel Logistic Regression Model for the Onset of Cigarette Use by Allele
Size (Boys)
3 repeats 4 repeats
RR 95% CI RR 95% CI
Intercept 1.287 (1.093 - 1.516 ) * 1.056 (0.869 - 1.284 )
Depression 0.961 (0.871 - 1.060 ) 1.094 (0.965 - 1.241 )
Intervention program 0.890 (0.691 - 1.145 ) 1.205 (0.882 - 1.647 )
Program X depression 1.082 (0.920 - 1.271 ) 0.877 (0.710 - 1.082 )
Hostility/Family conflict 1.065 (0.857 - 1.323 ) 1.260 (0.961 - 1.651 ) +
Program X hostility/family conflict 0.971 (0.692 - 1.362 ) 0.889 (0.598 - 1.320 )
Age at baseline 0.984 (0.895 - 1.083 ) 1.023 (0.938 - 1.116 )
Perceived education performance 0.950 (0.903 - 0.998 ) * 0.946 (0.876 - 1.023 )
Peer Smoking 1.034 (0.978 - 1.093 ) 1.047 (1.012 - 1.083 ) **
Allowance 0.999 (0.998 - 1.001 ) 0.999 (0.997 - 1.000 )
Access to cigarettes 0.898 (0.813 - 0.993 ) * 1.122 (0.995 - 1.266 ) +
Father smoking 1.006 (0.905 - 1.117 ) 1.050 (0.924 - 1.194 )
Mother smoking 0.941 (0.750 - 1.180 ) 1.045 (0.777 - 1.406 )
Time to watch TV per day 1.049 (1.016 - 1.083 ) ** 0.990 (0.951 - 1.029 )
Participating in physical activities 0.983 (0.897 - 1.078 ) 1.018 (0.911 - 1.138 )
Refusal self-efficacy 1.009 (0.903 - 1.128 ) 1.151 (1.006 - 1.317 ) *
Attitude to cigarette use 1.022 (0.916 - 1.140 ) 1.035 (0.890 - 1.203 )
Social norm 0.996 (0.974 - 1.018 ) 1.004 (0.977 - 1.032 )
Social consequence 1.018 (0.935 - 1.109 ) 0.946 (0.828 - 1.082 )
Health consequence 1.043 (0.959 - 1.134 ) 1.116 (0.995 - 1.251 ) +
Note: Analysis never smoker samples at baseline include 421 boys with 3 repeats and 281 with
4 repeats.

121
Table 25: Results of Multilevel Logistic Regression Model for the Onset of Cigarette Use (Boys)
RR 95% CI
Intercept 0.994 (0.822 - 1.204 )
Allele Size 1.320 (1.034 - 1.685 ) *
Depression 1.118 (0.992 - 1.261 ) +
Intervention program 1.290 (0.959 - 1.737 )
Hostility/Family conflict 1.283 (0.988 - 1.665 ) +
Age at baseline 1.004 (0.948 - 1.064 )
Perceived education performance 0.953 (0.915 - 0.992 ) *
Peer Smoking 1.040 (1.012 - 1.070 ) **
Allowance 0.999 (0.998 - 1.000 )
Access to cigarettes 1.098 (0.978 - 1.233 )
Father smoking 1.019 (0.940 - 1.105 )
Mother smoking 0.978 (0.821 - 1.165 )
Time to watch TV per day 0.986 (0.950 - 1.023 )
Participating in physical activities 0.999 (0.931 - 1.072 )
Refusal self-efficacy 1.064 (0.982 - 1.152 )
Attitude to cigarette use 1.035 (0.949 - 1.129 )
Social norm 0.999 (0.984 - 1.015 )
Social consequence 0.996 (0.916 - 1.083 )
Health consequence 1.072 (1.002 - 1.147 ) *
Allele size X depression 0.864 (0.740 - 1.008 ) +
Program X depression 0.871 (0.714 - 1.062 )
Interaction of allele and program 0.693 (0.473 - 1.016 ) +
Allele size X hostility/family conflict 0.796 (0.567 - 1.117 )
Program X hostility/family conflict 0.840 (0.576 - 1.224 )
Allele Size X program X hostility/family conflict 1.144 (0.687 - 1.904 )
Allele Size X program X depression 1.212 (0.939 - 1.565 )
Allele Size X access to cigarettes 0.846 (0.726 - 0.985 ) *
Allele Size X time to watch TV per day 1.066 (1.015 - 1.120 ) *
*: p <.05; **: p<.01; ***: p<.001 (two-tailed tests).

122
Table 26: Results of Multilevel Logistic Regression Model for the Onset of Cigarette Use by
Genotype (Girls)
Genotype (3/3 or 3/4) Genotype (4/4)
RR 95% CI RR p
Intercept 1.039 (0.951 - 1.135 ) 1.207 (1.013 - 1.439 ) +
Depression 1.047 (0.997 - 1.100 ) + 0.972 (0.888 - 1.063 )
Intervention program 1.034 (0.910 - 1.175 ) 1.002 (0.768 - 1.309 )
Program X depression 0.958 (0.893 - 1.029 ) 0.924 (0.780 - 1.095 )
Hostility/Family conflict 1.048 (0.925 - 1.188 ) 0.886 (0.694 - 1.130 )
Program X hostility/family conflict 1.048 (0.849 - 1.293 ) 1.386 (0.943 - 2.039 )
Age at baseline 1.001 (0.963 - 1.041 ) 1.013 (0.922 - 1.112 )
Perceived education performance 0.961 (0.926 - 0.999 ) * 1.028 (0.954 - 1.107 )
Peer Smoking 0.992 (0.976 - 1.007 ) 0.950 (0.897 - 1.005 ) +
Allowance 1.002 (1.000 - 1.003 ) * 1.001 (0.998 - 1.004 )
Access to cigarettes 0.967 (0.915 - 1.023 ) 1.142 (1.010 - 1.292 ) *
Father smoking 1.018 (0.955 - 1.085 ) 1.121 (0.978 - 1.283 )
Mother smoking 1.009 (0.900 - 1.130 ) 0.934 (0.778 - 1.121 )
Time to watch TV per day 1.011 (0.992 - 1.031 ) 1.011 (0.960 - 1.066 )
Participating in physical activities 1.010 (0.960 - 1.063 ) 1.133 (0.974 - 1.319 )
Refusal self-efficacy 1.050 (0.926 - 1.191 ) 1.222 (0.892 - 1.675 )
Attitude to cigarette use 1.058 (0.993 - 1.128 ) + 1.136 (0.945 - 1.365 )
Social norm 1.008 (0.993 - 1.023 ) 1.021 (0.993 - 1.051 )
Social consequence 0.994 (0.942 - 1.050 ) 0.922 (0.791 - 1.074 )
Health consequence 1.002 (0.943 - 1.065 ) 1.011 (0.875 - 1.167 )
Note: Analysis never smoker samples at baseline include 746 female with genotypes 3/3 or 3/4
and 142 with genotype 4/4. *: p <.05; **: p<.01; ***: p<.001 (two-tailed tests).

123
Table 27: Results of Multilevel Logistic Regression Model for the Onset of Cigarette Use (Girls)
RR 95% CI
Intercept 1.212 (1.017 - 1.444 ) +
Repeat genotype 0.854 (0.703 - 1.037 )
Depression 0.972 (0.884 - 1.068 )
Intervention program 0.951 (0.727 - 1.244 )
Hostility/Family conflict 0.999 (0.781 - 1.277 )
Age at baseline 1.006 (0.963 - 1.050 )
Perceived education performance 0.971 (0.935 - 1.008 )
Peer Smoking 0.988 (0.973 - 1.003 )
Allowance 1.001 (1.000 - 1.002 ) **
Access to cigarettes 1.084 (0.960 - 1.224 )
Father smoking 1.035 (0.976 - 1.097 )
Mother smoking 0.992 (0.891 - 1.104 )
Time to watch TV per day 1.011 (0.992 - 1.030 )
Participating in physical activities 1.025 (0.978 - 1.075 )
Refusal self-efficacy 1.071 (0.986 - 1.165 )
Attitude to cigarette use 1.172 (1.018 - 1.351 ) *
Social norm 1.009 (0.997 - 1.020 )
Social consequence 0.993 (0.944 - 1.044 )
Health consequence 1.002 (0.948 - 1.058 )
Repeat genotype X depression 1.079 (0.970 - 1.199 )
Program X depression 0.955 (0.805 - 1.132 )
Repeat genotype X program 1.087 (0.809 - 1.461 )
Repeat genotype X hostility/family conflict 1.037 (0.789 - 1.363 )
Program X hostility/family conflict 1.209 (0.816 - 1.792 )
Repeat genotype X program X hostility/family conflict 0.876 (0.559 - 1.370 )
Repeat genotype X program X depression 1.004 (0.835 - 1.208 )
Repeat genotype X access to cigarettes 0.897 (0.786 - 1.024 )
Repeat genotype X attitude to cigarettes use 0.897 (0.769 - 1.046 )
*: p <.05; **: p<.01; ***: p<.001 (two-tailed tests).

124
Table 28: Results of Random Effects Models for Life-time Ever Smokers at Baseline by Allele Size (Boys)
3 repeats 4 repeats
Estimate SE DF p Estimate SE DF p
Intercept 0.400 0.095 12 0.001 ** 0.681 0.098 12 0.000 ***
# of cigarettes smoked per day in the past month at
baseline
-.023 0.026 165 0.365 0.463 0.127 101 0.000 ***
Intervention program -.007 0.126 12 0.956 -.320 0.124 12 0.024 *
Depression -.021 0.132 165 0.872 -.311 0.136 101 0.024 *
Program X depression -.308 0.176 165 0.081 + 0.472 0.193 101 0.016 *
Hostility/Family conflict -.290 0.253 165 0.253 -.101 0.260 101 0.698
Program X hostility/family conflict 0.146 0.331 165 0.660 -.268 0.405 101 0.510
Age at baseline 0.016 0.093 165 0.860 -.203 0.112 101 0.074 +
Perceived education performance -.119 0.063 165 0.061 + -.055 0.067 101 0.409
Peer Smoking 0.068 0.036 165 0.062 + 0.072 0.080 101 0.368
Allowance 0.002 0.003 165 0.570 0.002 0.003 101 0.482
Access to cigarettes 0.184 0.134 165 0.171 -.258 0.122 101 0.037 *
Father smoking 0.116 0.137 165 0.398 0.022 0.163 101 0.891
Mother smoking 0.140 0.286 165 0.626 0.115 0.334 101 0.732
Participating in physical activities -.077 0.121 165 0.524 0.147 0.128 101 0.254
Time to watch TV per day 0.027 0.035 165 0.444 0.052 0.040 101 0.197

125
Table 28: Continued

3 repeats 4 repeats
Estimate SE DF p Estimate SE DF p
Refusal self-efficacy -.259 0.167 165 0.122 0.122 0.323 101 0.706
Attitude to cigarette use 0.226 0.114 165 0.048 * -.044 0.122 101 0.718
Social norm 0.045 0.032 165 0.155 0.079 0.037 101 0.036 *
Social consequence -.030 0.108 165 0.778 0.041 0.113 101 0.719
Health consequence 0.239 0.135 165 0.079 + -.231 0.119 101 0.054 +
Note: *: p <.05; **: p<.01; ***: p<.001 (two-tailed tests).
Analysis sample include 270 boys with 3 repeats and 172 with 4 repeats.

126
Table 29: Results of Random Effects Models for Life-time Ever Smokers at Baseline (Boys)
Statistics Parameter
Estimate SE DF p
Intercept 0.477 0.16 12 0.011 *
# of cigarettes smoked per day in the past month at baseline -0.011 0.026 290 0.683
Allele Size 0.077 0.205 290 0.709
Intervention program -0.094 0.229 12 0.69
Allele size X program -0.033 0.285 290 0.909
Depression -0.238 0.167 290 0.155
Allele size X depression 0.276 0.213 290 0.196
Program X depression 0.428 0.233 290 0.067 +
Allele size X depression X program -0.842 0.291 290 0.004 **
Hostility/Family conflict 0.053 0.318 290 0.868
Allele size X hostility/family conflict -0.305 0.41 290 0.458
Program X hostility/family conflict -0.431 0.468 290 0.358
Allele size X hostility X program 0.577 0.576 290 0.317
Age at baseline -0.037 0.083 290 0.66
Perceived education performance -0.136 0.052 290 0.009 **
Peer Smoking 0.077 0.039 290 0.048 *
Allowance 0.002 0.003 290 0.433
Access to cigarettes -0.211 0.17 290 0.216
Father smoking 0.124 0.116 290 0.288
Mother smoking 0.089 0.302 290 0.768
Participating in physical activities -0.015 0.093 290 0.874
Time to watch TV per day 0.044 0.027 290 0.108
Refusal self-efficacy -0.188 0.158 290 0.235
Attitude to cigarette use -0.07 0.137 290 0.612
Social norm 0.048 0.022 290 0.03 *
Social consequence -0.007 0.089 290 0.938
Health consequence -0.16 0.143 290 0.265

127
Table 29: Continued

Statistics Parameter
Estimate SE DF p
Allele Size X access to cigarettes 0.325 0.196 290 0.098 +
Allele size X attitude to cigarette use 0.329 0.178 290 0.066 +
Allele size X health consequence 0.392 0.164 290 0.018 *

128

Table 30: Results of Random Effects Models for Life-time Ever Smokers at Baseline (Girls)
Statistics Parameter
Estimate SE DF p
Intercept 0.033 0.04 12 0.423
# of cigarettes smoked per day in the past month at baseline 0.057 0.028 105 0.044 *
Repeat genotype 0.084 0.053 105 0.114
Intervention program 0.069 0.058 12 0.258
Repeat genotype X program -0.103 0.078 105 0.193
Depression 0.101 0.066 105 0.129
Repeat genotype X depression -0.067 0.051 105 0.187
Program X depression 0.017 0.11 105 0.88
Repeat genotype X program X depression -0.024 0.094 105 0.798
Hostility/Family conflict 0.231 0.109 105 0.036 *
Repeat genotype X Hostility -0.292 0.13 105 0.027 *
Hostility X program 0.022 0.184 105 0.905
Repeat genotype X hostility X program 0.085 0.211 105 0.687
Age at baseline -0.08 0.049 105 0.109
Perceived education performance -0.021 0.031 105 0.508
Peer Smoking -0.033 0.015 105 0.03 *
Allowance 0.001 0.001 105 0.494
Access to cigarettes 0.015 0.023 105 0.512
Father smoking 0.063 0.065 105 0.338
Mother smoking 0.051 0.029 105 0.089 +
Participating in physical activities 0.012 0.012 105 0.299
Time to watch TV per day 0.06 0.025 105 0.019 *
Refusal self-efficacy 0.015 0.028 105 0.584
Attitude to cigarette use 0.009 0.025 105 0.723
Social norm -0.007 0.005 105 0.119
Social consequence -0.044 0.043 105 0.314
Health consequence 0.013 0.024 105 0.582

129

Table 30: Continued

Statistics Parameter
Estimate SE DF p
Repeat genotype X peer smoking 0.045 0.016 105 0.006 **
Repeat genotype X pocket money -0.002 0.002 105 0.255

130
Table 31: Estimated Parameters from the Model with Refusal Self-Efficacy as Mediator by Gender and Genetic Factor
Boy Girl
Low (3-repeat) activity High (4-repeat) activity Genotype of 3/3 or 3/4 Genotype of 4/4
Estimate SE p Estimate SE p Estimate SE p Estimate SE p
Means

Intercept of SMK
1.271 1.133 0.262 0.756 0.586 0.197 0.636 0.275 0.021 0.426 0.585 0.467
Slope of SMK
-0.136 0.963 0.888 0.885 0.625 0.157 -0.224 0.170 0.188 -0.188 0.573 0.743
Intercept of RSE
0.949 0.261 0.000 *** 1.385 0.353 0.000 *** 0.770 0.247 0.002 ** 0.851 0.209 0.000 ***
Slope of RSE
-0.130 0.177 0.462 0.001 0.267 0.996 -0.037 0.222 0.867 -0.052 0.129 0.689
Intercept of SMK

Intercept of RSE
-2.020 0.575 0.000 *** -0.888 0.256 0.001 ** -0.656 0.257 0.011 * -0.340 0.113 0.003 **
Slope of RSE
2.742 1.330 0.039 * 0.087 0.304 0.774 0.188 0.210 0.371 3.070 8.278 0.711
Slope of SMK

Intercept of RSE
0.456 0.424 0.282 -0.722 0.306 0.018 * 0.221 0.138 0.108 0.122 0.066 0.067 +
Slope of RSE
-3.278 1.326 0.013 * -1.412 0.503 0.005 ** -0.268 0.177 0.130 -3.167 8.380 0.705
Effects on Intercept of SMK

Intervention Program
0.336 0.323 0.297 -0.026 0.153 0.863 -0.073 0.057 0.197 -0.006 0.173 0.973
Depression
0.144 0.173 0.405 0.067 0.073 0.355 0.015 0.022 0.500 -0.006 0.068 0.934
Program X Depression
-0.278 0.243 0.253 -0.052 0.120 0.666 0.032 0.035 0.362 -0.008 0.095 0.935
Hostility/family conflict
-0.125 0.333 0.707 -0.118 0.116 0.309 -0.031 0.062 0.619 0.165 0.353 0.641
Program X Hostility
-0.411 0.460 0.372 0.216 0.203 0.287 0.032 0.078 0.681 -0.130 0.418 0.756
Age
0.070 0.069 0.312 0.005 0.040 0.910 0.008 0.016 0.590 -0.004 0.034 0.907
Education Performance
0.005 0.065 0.945 -0.024 0.023 0.286 -0.020 0.018 0.265 -0.007 0.034 0.839
Peer Smoking
0.003 0.049 0.957 0.001 0.024 0.980 -0.009 0.010 0.359 -0.055 0.160 0.733
Pocket Money
-0.003 0.003 0.171 0.000 0.001 0.907 0.000 0.000 0.273 0.000 0.002 0.793
Access to cigarettes
0.036 0.116 0.758 0.094 0.050 0.062 + -0.029 0.027 0.285 0.013 0.066 0.841
Father smoking
-0.161 0.171 0.345 -0.031 0.053 0.565 -0.007 0.024 0.777 0.037 0.084 0.664
Mother smoking
0.186 0.286 0.516 0.014 0.142 0.920 -0.021 0.040 0.589 0.054 0.219 0.804
Physical Activity
-0.236 0.137 0.085 + 0.057 0.046 0.218 -0.034 0.027 0.202 -0.016 0.047 0.735
Time to watch TV per day
0.023 0.038 0.549 0.009 0.019 0.628 0.004 0.007 0.569 -0.001 0.026 0.955

131
Table 31: Continued

Boy Girl
Low (3-repeat) activity High (4-repeat) activity Genotype of 3/3 or 3/4 Genotype of 4/4
Estimate SE p Estimate SE p Estimate SE p Estimate SE p
Effects on Slope of SMK

Intervention Program
-0.142 0.268 0.596 -0.049 0.163 0.765 -0.028 0.036 0.428 0.044 0.171 0.798
Depression
-0.149 0.135 0.270 0.024 0.082 0.771 -0.013 0.016 0.424 0.013 0.067 0.849
Program X Depression
0.191 0.191 0.317 0.047 0.102 0.643 0.022 0.023 0.324 -0.006 0.094 0.951
Hostility/family conflict
0.072 0.289 0.803 0.063 0.171 0.712 -0.016 0.040 0.688 -0.156 0.356 0.661
Program X Hostility
0.092 0.421 0.828 -0.358 0.255 0.162 0.011 0.052 0.831 0.090 0.421 0.831
Age
-0.037 0.066 0.573 -0.009 0.035 0.789 0.005 0.011 0.633 0.005 0.033 0.886
Education Performance
-0.013 0.052 0.808 -0.037 0.029 0.202 -0.014 0.012 0.219 -0.005 0.034 0.879
Peer Smoking
0.035 0.046 0.453 0.021 0.028 0.451 0.001 0.006 0.841 0.056 0.161 0.729
Pocket Money
0.002 0.002 0.340 -0.001 0.001 0.526 0.000 0.000 0.350 0.000 0.001 0.755
Access to cigarettes
-0.032 0.104 0.758 -0.085 0.048 0.076 + 0.009 0.017 0.606 -0.020 0.065 0.762
Father smoking
0.114 0.148 0.440 -0.034 0.062 0.582 0.024 0.017 0.142 -0.041 0.084 0.625
Mother smoking
0.131 0.246 0.595 -0.051 0.119 0.671 0.022 0.036 0.543 -0.081 0.220 0.711
Physical Activity
0.200 0.126 0.111 0.030 0.056 0.593 0.004 0.016 0.783 0.007 0.045 0.885
Time to watch TV per day
0.003 0.038 0.945 0.015 0.019 0.403 -0.005 0.005 0.310 0.011 0.025 0.675
Effects on Intercept of RSE

Intervention Program
0.043 0.068 0.522 -0.064 0.081 0.428 -0.023 0.055 0.677 0.028 0.052 0.588
Depression
-0.049 0.031 0.116 -0.023 0.035 0.508 0.000 0.025 0.991 -0.018 0.018 0.332
Program X Depression
-0.024 0.046 0.600 -0.043 0.057 0.445 0.001 0.033 0.978 -0.048 0.033 0.139
Hostility/family conflict
-0.110 0.066 0.093 + -0.266 0.075 0.000 *** -0.106 0.068 0.118 -0.058 0.049 0.231
Program X Hostility
-0.018 0.094 0.846 0.292 0.117 0.013 * 0.004 0.098 0.963 0.037 0.083 0.653
Age
-0.005 0.020 0.811 -0.038 0.027 0.166 0.013 0.018 0.445 0.009 0.016 0.578
Education Performance
0.048 0.013 0.000 *** 0.035 0.015 0.015 * 0.023 0.015 0.123 0.018 0.013 0.162
Peer Smoking
-0.048 0.008 0.000 *** -0.057 0.007 0.000 *** -0.024 0.012 0.037 * -0.034 0.010 0.001 **
Pocket Money
-0.001 0.000 0.302 0.000 0.001 0.460 0.000 0.000 0.981 0.000 0.000 0.567
Access to cigarettes
-0.057 0.031 0.060 + -0.012 0.037 0.739 -0.059 0.028 0.037 * -0.040 0.025 0.111

132

Table 31: Continued

Boy Girl
Low (3-repeat) activity High (4-repeat) activity Genotype of 3/3 or 3/4 Genotype of 4/4
Estimate SE p Estimate SE p Estimate SE p Estimate SE p
Father smoking
-0.023 0.031 0.453 -0.052 0.038 0.171 -0.010 0.027 0.703 -0.009 0.029 0.757
Mother smoking
0.039 0.056 0.489 -0.078 0.088 0.373 0.004 0.049 0.943 -0.062 0.049 0.204
Physical Activity
-0.045 0.028 0.103 0.013 0.034 0.698 -0.007 0.027 0.786 -0.004 0.023 0.863
Time to watch TV per day
-0.025 0.009 0.005 ** -0.002 0.011 0.854 -0.012 0.008 0.156 -0.008 0.009 0.335
Effects on Slope of RSE

Intervention Program
-0.020 0.049 0.677 0.060 0.067 0.376 -0.024 0.047 0.613 0.016 0.034 0.637
Depression
0.021 0.025 0.389 -0.002 0.030 0.943 -0.016 0.021 0.452 0.006 0.014 0.674
Program X Depression
0.003 0.033 0.918 0.023 0.047 0.622 0.005 0.029 0.873 -0.008 0.020 0.667
Hostility/family conflict
0.047 0.049 0.335 0.053 0.058 0.360 -0.011 0.052 0.833 -0.042 0.033 0.206
Program X Hostility
0.022 0.069 0.754 -0.137 0.094 0.143 0.009 0.073 0.901 0.047 0.050 0.341
Age
-0.003 0.013 0.831 -0.010 0.020 0.626 -0.001 0.016 0.939 0.002 0.010 0.854
Education Performance
0.004 0.010 0.664 -0.002 0.012 0.859 0.015 0.013 0.257 0.003 0.007 0.667
Peer Smoking
0.003 0.005 0.539 0.017 0.005 0.001 ** 0.008 0.009 0.375 0.019 0.005 0.000 ***
Pocket Money
0.000 0.000 0.852 0.000 0.000 0.298 -0.001 0.000 0.001 ** 0.000 0.000 0.511
Access to cigarettes
0.012 0.021 0.555 0.004 0.028 0.896 -0.009 0.023 0.688 -0.005 0.016 0.738
Father smoking
0.015 0.023 0.503 0.007 0.030 0.817 0.027 0.025 0.283 -0.008 0.017 0.660
Mother smoking
-0.010 0.038 0.793 0.038 0.062 0.543 0.057 0.034 0.092 + -0.024 0.029 0.398
Physical Activity
0.057 0.019 0.003 * 0.021 0.027 0.427 0.005 0.020 0.817 0.000 0.014 0.982
Time to watch TV per day
-0.006 0.006 0.348 -0.005 0.008 0.565 -0.005 0.007 0.462 0.002 0.005 0.685
+: p<0.10; *: p<0.05; **: p<0.01; ***: p<0.001

133
Table 32: Estimated Parameters from the Model with Social Norm as Mediator by Gender and Genetic Factor
Boy Girl
Low (3-repeat) activity High (4-repeat) activity Genotype of 3/3 or 3/4 Genotype of 4/4
Estimate SE p Estimate SE p Estimate SE p Estimate SE p
Means

Intercept of SMK
-0.634 0.879 0.471 0.708 1.142 0.535 0.538 1.816 0.766 0.058 0.131 0.658
Slope of SMK
0.168 0.894 0.851 -0.688 1.200 0.566 1.262 5.269 0.810 -0.058 0.119 0.627
Intercept of SN
-0.852 1.834 0.642 -2.745 2.125 0.197 2.333 1.875 0.213 -0.005 1.689 0.998
Slope of SN
1.253 1.349 0.353 3.447 1.602 0.031 * -1.463 1.263 0.247 2.247 1.125 0.046 *
Intercept of SMK

Intercept of SN
0.134 0.119 0.259 0.132 0.061 0.031 * -0.006 0.019 0.460 0.022 0.015 0.139
Slope of SN
-0.180 0.194 0.353 -0.244 0.270 0.366 0.273 1.230 0.410 -0.037 0.038 0.339
Slope of SMK

Intercept of SN
0.011 0.113 0.925 0.137 0.064 0.031 * -0.027 0.013 0.975 -0.008 0.008 0.332
Slope of SN
0.450 0.217 0.038 * 0.270 0.288 0.348 0.848 3.572 0.416 0.061 0.044 0.169
Effects on Intercept of SMK

Intervention Program
0.409 0.304 0.179 -0.128 0.221 0.563 -0.076 0.113 0.615 0.050 0.039 0.195
Depression
0.282 0.162 0.082 + -0.085 0.114 0.459 0.022 0.062 0.473 0.015 0.016 0.335
Program X Depression
-0.324 0.249 0.194 0.189 0.199 0.343 0.037 0.063 0.582 -0.022 0.021 0.299
Hostility/family conflict
0.059 0.331 0.859 0.245 0.163 0.132 0.316 1.236 0.425 0.045 0.034 0.185
Program X Hostility
-0.287 0.396 0.468 -0.457 0.389 0.240 -0.078 0.485 0.383 -0.006 0.051 0.904
Age
0.035 0.056 0.537 -0.018 0.062 0.779 -0.055 0.241 0.412 -0.009 0.010 0.362
Education Performance
-0.077 0.059 0.190 -0.122 0.073 0.094 + -0.012 0.102 0.363 -0.003 0.009 0.770
Peer Smoking
0.086 0.042 0.041 * 0.034 0.018 0.056 + 0.014 0.025 0.562 0.010 0.010 0.272
Pocket Money
-0.005 0.003 0.141 0.000 0.001 0.793 0.001 0.002 0.447 0.000 0.000 0.726
Access to cigarettes
0.150 0.101 0.136 -0.052 0.115 0.651 -0.013 0.106 0.366 -0.003 0.018 0.862
Father smoking
-0.098 0.136 0.474 -0.079 0.086 0.360 0.079 0.325 0.419 0.027 0.022 0.232
Mother smoking
0.012 0.299 0.969 -0.067 0.212 0.753 -0.130 0.529 0.421 -0.010 0.028 0.734
Physical Activity
0.045 0.108 0.677 0.008 0.069 0.908 -0.083 0.257 0.455 -0.017 0.016 0.289
Time to watch TV per day
0.052 0.036 0.146 -0.005 0.030 0.878 0.011 0.014 0.678 0.002 0.009 0.813

134
Table 32: Continued

Boy Girl
Low (3-repeat) activity High (4-repeat) activity Genotype of 3/3 or 3/4 Genotype of 4/4
Estimate SE p Estimate SE p Estimate SE p Estimate SE p
Effects on Slope of SMK

Intervention Program
-0.351 0.306 0.251 -0.003 0.212 0.987 -0.072 0.314 0.412 -0.024 0.029 0.398
Depression
-0.205 0.136 0.133 0.022 0.129 0.864 0.025 0.179 0.375 -0.009 0.012 0.430
Program X Depression
0.278 0.225 0.216 -0.031 0.200 0.878 0.039 0.167 0.415 0.021 0.016 0.175
Hostility/family conflict
0.004 0.296 0.988 0.165 0.209 0.429 0.837 3.575 0.415 -0.029 0.024 0.221
Program X Hostility
0.105 0.380 0.783 -0.294 0.454 0.517 -0.315 1.392 0.412 -0.040 0.039 0.315
Age
-0.010 0.064 0.874 0.037 0.060 0.536 -0.162 0.705 0.413 0.010 0.010 0.322
Education Performance
0.019 0.054 0.722 0.001 0.080 0.991 0.053 0.301 0.390 -0.011 0.009 0.213
Peer Smoking
-0.001 0.042 0.974 0.011 0.017 0.514 0.013 0.073 0.392 -0.005 0.005 0.381
Pocket Money
0.005 0.003 0.112 -0.001 0.001 0.253 0.002 0.007 0.416 0.000 0.000 0.846
Access to cigarettes
-0.096 0.106 0.368 -0.071 0.124 0.566 -0.066 0.309 0.406 0.008 0.015 0.616
Father smoking
0.006 0.126 0.963 -0.083 0.088 0.347 0.251 0.929 0.431 -0.034 0.027 0.216
Mother smoking
0.279 0.262 0.286 -0.050 0.175 0.776 -0.353 1.508 0.415 -0.008 0.022 0.716
Physical Activity
-0.049 0.105 0.639 -0.016 0.056 0.779 -0.165 0.742 0.410 0.009 0.014 0.516
Time to watch TV per day
-0.006 0.039 0.872 0.033 0.022 0.130 -0.004 0.040 0.361 0.007 0.007 0.265
Effects on Intercept of SN

Intervention Program
-0.764 0.461 0.098 + 0.390 0.560 0.486 -0.103 0.556 0.394 -0.188 0.379 0.620
Depression
0.008 0.207 0.970 0.765 0.264 0.004 ** 0.044 0.206 0.407 0.238 0.152 0.116
Program X Depression
0.453 0.312 0.147 -0.530 0.388 0.172 0.167 0.297 0.565 0.062 0.204 0.762
Hostility/family conflict
0.733 0.438 0.094 + -0.413 0.489 0.399 0.868 0.571 0.898 0.572 0.357 0.110
Program X Hostility
-0.381 0.616 0.536 1.239 0.741 0.095 + -0.323 0.749 0.505 -0.035 0.511 0.946
Age
0.193 0.138 0.162 0.205 0.160 0.200 -0.091 0.148 0.589 0.089 0.123 0.469
Education Performance
-0.093 0.088 0.290 0.047 0.086 0.582 -0.159 0.121 0.851 -0.077 0.097 0.425
Peer Smoking
0.174 0.041 0.000 *** 0.173 0.048 0.000 *** 0.150 0.065 0.983 0.157 0.051 0.002 **
Pocket Money
0.011 0.004 0.004 ** 0.002 0.004 0.671 0.002 0.004 0.544 0.004 0.003 0.128
Access to cigarettes
0.235 0.201 0.241 0.583 0.231 0.012 * 0.132 0.215 0.590 0.209 0.176 0.236

135
Table 32: Continued

Boy Girl
Low (3-repeat) activity High (4-repeat) activity Genotype of 3/3 or 3/4 Genotype of 4/4
Estimate SE p Estimate SE p Estimate SE p Estimate SE p
Father smoking
0.369 0.205 0.072 + 0.646 0.224 0.004 ** 0.574 0.212 0.995 -0.067 0.196 0.730
Mother smoking
0.155 0.352 0.660 0.570 0.470 0.225 -0.186 0.430 0.506 0.517 0.328 0.115
Physical Activity
-0.106 0.185 0.568 0.171 0.209 0.413 0.160 0.212 0.653 0.063 0.153 0.680
Time to watch TV per day
0.118 0.059 0.045 * 0.024 0.073 0.743 0.177 0.071 0.990 0.173 0.060 0.004 **
Effects on Slope of SN

Intervention Program
0.675 0.353 0.056 + -0.479 0.414 0.248 0.047 0.320 0.377 0.329 0.271 0.225
Depression
-0.072 0.137 0.601 -0.302 0.175 0.085 + -0.039 0.125 0.449 0.046 0.123 0.705
Program X Depression
-0.258 0.226 0.253 0.541 0.273 0.048 * -0.015 0.191 0.349 -0.100 0.155 0.520
Hostility/family conflict
-0.291 0.292 0.320 0.284 0.316 0.369 -1.005 0.331 0.998 0.054 0.273 0.844
Program X Hostility
-0.233 0.430 0.588 -0.972 0.516 0.059 + 0.379 0.455 0.685 -0.267 0.381 0.484
Age
-0.049 0.100 0.622 -0.120 0.118 0.308 0.198 0.101 0.960 -0.146 0.082 0.075 +
Education Performance
-0.052 0.069 0.457 -0.245 0.066 0.000 *** -0.083 0.078 0.771 -0.034 0.066 0.611
Peer Smoking
0.002 0.039 0.959 0.017 0.032 0.591 -0.018 0.040 0.513 -0.045 0.034 0.192
Pocket Money
-0.008 0.003 0.005 ** 0.000 0.002 0.852 -0.002 0.002 0.663 -0.001 0.002 0.661
Access to cigarettes
-0.017 0.150 0.909 -0.334 0.158 0.034 * 0.079 0.137 0.571 -0.218 0.131 0.096 +
Father smoking
0.101 0.153 0.511 -0.046 0.178 0.796 -0.259 0.139 0.951 0.259 0.146 0.076 +
Mother smoking
-0.219 0.260 0.401 -0.312 0.315 0.322 0.422 0.228 0.948 -0.010 0.271 0.970
Physical Activity
0.082 0.138 0.551 * -0.068 0.145 0.641 0.202 0.136 0.892 -0.061 0.114 0.592
Time to watch TV per day
0.031 0.043 0.472 -0.049 0.051 0.339 0.004 0.041 0.353 -0.054 0.042 0.196
+: p<0.10; *: p<0.05; **: p<0.01; ***: p<0.001

136
Table 33: Estimated Parameters from the Model with Attitude to Cigarette Use as Mediator by Gender and Genetic Factor
Boy Girl
Low (3-repeat) activity High (4-repeat) activity Genotype of 3/3 or 3/4 Genotype of 4/4
Estimate SE p Estimate SE p Estimate SE p Estimate SE p
Means

Intercept of SMK
-1.457 0.932 0.118 -0.381 0.443 0.389 -0.049 0.192 0.798 -0.062 0.097 0.518
Slope of SMK
0.448 0.864 0.604 -0.323 0.499 0.518 0.043 0.150 0.776 0.074 0.062 0.231
Intercept of ACU
1.975 0.526 0.000 *** -0.449 0.711 0.527 *** 1.760 0.655 0.007 ** 0.989 0.551 0.072 +
Slope of ACU
0.093 0.365 0.799 0.426 0.483 0.378 -0.297 0.405 0.463 0.139 0.379 0.714
Intercept of SMK

Intercept of ACU
0.268 0.245 0.273 0.165 0.111 0.137 0.086 0.044 0.052 + 0.040 0.024 0.101
Slope of ACU
-0.305 0.626 0.626 -0.045 0.178 0.802 -0.081 0.113 0.472 -0.005 0.078 0.952
Slope of SMK

Intercept of ACU
0.134 0.258 0.604 0.169 0.111 0.127 -0.034 0.036 0.342 0.000 0.017 0.991
Slope of ACU
0.147 0.667 0.826 0.710 0.238 0.003 ** 0.085 0.097 0.379 0.041 0.069 0.556
Effects on Intercept of SMK

Intervention Program
0.220 0.303 0.468 0.082 0.177 0.645 -0.060 0.048 0.208 0.034 0.032 0.300
Depression
0.277 0.141 0.050 * 0.092 0.076 0.229 0.002 0.014 0.896 0.014 0.014 0.301
Program X Depression
-0.243 0.242 0.315 -0.037 0.135 0.783 0.032 0.028 0.256 -0.016 0.018 0.379
Hostility/family conflict
0.027 0.323 0.933 0.055 0.096 0.568 0.032 0.047 0.498 0.052 0.028 0.062 +
Program X Hostility
-0.224 0.396 0.572 -0.092 0.163 0.571 -0.003 0.074 0.965 0.001 0.050 0.991
Age
0.074 0.048 0.122 0.012 0.033 0.717 0.004 0.014 0.773 -0.002 0.007 0.736
Education Performance
-0.059 0.062 0.336 -0.056 0.022 0.011 * -0.029 0.019 0.129 -0.004 0.008 0.653
Peer Smoking
0.101 0.036 0.005 ** 0.047 0.016 0.003 ** 0.011 0.007 0.136 0.015 0.008 0.055 +
Pocket Money
-0.004 0.002 0.134 0.000 0.001 0.680 0.000 0.000 0.296 0.000 0.000 0.852
Access to cigarettes
0.157 0.105 0.133 0.065 0.063 0.300 0.008 0.018 0.643 0.005 0.012 0.670
Father smoking
-0.003 0.201 0.990 -0.002 0.054 0.963 -0.004 0.017 0.796 0.017 0.016 0.279
Mother smoking
0.107 0.262 0.684 0.081 0.124 0.518 -0.009 0.040 0.820 0.004 0.027 0.867
Physical Activity
-0.005 0.106 0.963 0.032 0.045 0.474 -0.035 0.024 0.144 -0.012 0.015 0.420
Time to watch TV per day
0.050 0.034 0.135 0.013 0.019 0.500 0.008 0.005 0.102 0.007 0.006 0.239

137
Table 33: Continued

Boy Girl
Low (3-repeat) activity High (4-repeat) activity Genotype of 3/3 or 3/4 Genotype of 4/4
Estimate SE p Estimate SE p Estimate SE p Estimate SE p
Effects on Slope of SMK

Intervention Program
-0.045 0.252 0.859 -0.186 0.184 0.312 -0.032 0.033 0.326 -0.002 0.022 0.932
Depression
-0.249 0.119 0.037 * -0.013 0.089 0.886 -0.003 0.013 0.807 -0.009 0.008 0.276
Program X Depression
0.161 0.197 0.413 0.119 0.122 0.328 0.023 0.020 0.261 0.014 0.011 0.223
Hostility/family conflict
-0.147 0.288 0.609 0.204 0.188 0.278 -0.037 0.038 0.327 -0.029 0.016 0.058 +
Program X Hostility
0.007 0.385 0.985 -0.313 0.248 0.208 0.035 0.058 0.546 -0.056 0.038 0.142
Age
-0.025 0.056 0.651 0.022 0.039 0.579 0.005 0.010 0.627 0.001 0.005 0.898
Education Performance
-0.003 0.051 0.959 -0.049 0.027 0.069 + -0.014 0.013 0.257 -0.013 0.010 0.176
Peer Smoking
-0.005 0.038 0.890 0.035 0.019 0.067 + -0.008 0.005 0.104 -0.008 0.004 0.063 +
Pocket Money
0.001 0.003 0.602 0.000 0.001 0.856 0.000 0.000 0.763 0.000 0.000 0.789
Access to cigarettes
-0.105 0.096 0.271 -0.028 0.067 0.677 -0.006 0.015 0.680 -0.007 0.007 0.296
Father smoking
0.063 0.188 0.738 -0.012 0.072 0.863 0.021 0.015 0.178 -0.020 0.020 0.317
Mother smoking
0.196 0.222 0.378 -0.121 0.117 0.302 0.007 0.032 0.823 -0.016 0.016 0.318
Physical Activity
-0.006 0.101 0.953 0.018 0.062 0.764 0.005 0.015 0.755 0.005 0.011 0.658
Time to watch TV per day
0.005 0.037 0.892 0.018 0.019 0.343 -0.005 0.004 0.176 0.003 0.005 0.518
Effects on Intercept of ACU

Intervention Program
-0.088 0.138 0.522 -0.214 0.165 0.196 0.050 0.163 0.760 0.008 0.127 0.951
Depression
0.073 0.067 0.276 -0.010 0.066 0.882 0.091 0.063 0.145 0.122 0.051 0.017 *
Program X Depression
0.069 0.087 0.431 0.122 0.103 0.240 -0.033 0.089 0.713 -0.023 0.071 0.747
Hostility/family conflict
0.449 0.131 0.001 ** 0.402 0.163 0.014 * 0.097 0.170 0.567 0.113 0.110 0.306
Program X Hostility
-0.174 0.178 0.330 0.166 0.227 0.466 0.192 0.229 0.401 0.060 0.168 0.721
Age
-0.031 0.039 0.419 0.148 0.054 0.007 ** -0.026 0.050 0.603 0.022 0.042 0.597
Education Performance
-0.043 0.026 0.097 + -0.003 0.030 0.920 * -0.050 0.038 0.188 0.012 0.029 0.677
Peer Smoking
0.044 0.013 0.001 ** 0.033 0.012 0.006 ** -0.019 0.020 0.334 0.024 0.020 0.225
Pocket Money
0.003 0.001 0.010 * 0.000 0.001 0.830 0.000 0.001 0.848 0.003 0.001 0.070 +
Access to cigarettes
0.081 0.059 0.167 0.198 0.073 0.007 ** 0.057 0.072 0.427 0.119 0.061 0.050 *

138
Table 33: Continued

Boy Girl
Low (3-repeat) activity High (4-repeat) activity Genotype of 3/3 or 3/4 Genotype of 4/4
Estimate SE p Estimate SE p Estimate SE p Estimate SE p
Father smoking
-0.139 0.066 0.036 * 0.112 0.081 0.167 0.079 0.081 0.332 -0.026 0.072 0.720
Mother smoking
-0.075 0.097 0.440 0.084 0.144 0.559 -0.080 0.117 0.492 -0.055 0.094 0.558
Physical Activity
0.010 0.055 0.851 0.082 0.065 0.207 0.066 0.069 0.334 -0.052 0.051 0.305
Time to watch TV per day
0.033 0.016 0.046 * -0.009 0.022 0.692 0.035 0.020 0.079 + 0.024 0.018 0.179
Effects on Slope of ACU

Intervention Program
0.053 0.106 0.614 0.198 0.123 0.106 0.071 0.108 0.509 -0.049 0.086 0.564
Depression
0.006 0.053 0.916 0.076 0.053 0.152 -0.023 0.045 0.609 0.008 0.030 0.798
Program X Depression
-0.036 0.069 0.598 -0.136 0.078 0.082 + -0.027 0.059 0.653 0.029 0.046 0.526
Hostility/family conflict
-0.270 0.095 0.004 ** -0.113 0.107 0.293 0.043 0.121 0.722 -0.030 0.070 0.669
Program X Hostility
0.129 0.126 0.308 -0.132 0.166 0.427 -0.186 0.159 0.244 -0.002 0.114 0.988
Age
-0.011 0.027 0.685 -0.023 0.037 0.530 0.030 0.031 0.321 -0.011 0.030 0.708
Education Performance
0.034 0.018 0.067 + -0.014 0.021 0.511 -0.001 0.027 0.980 0.009 0.018 0.621
Peer Smoking
0.008 0.011 0.452 -0.002 0.009 0.810 0.013 0.014 0.323 -0.025 0.010 0.015 *
Pocket Money
-0.002 0.001 0.070 + -0.001 0.001 0.109 0.001 0.001 0.248 0.000 0.001 0.817
Access to cigarettes
-0.022 0.040 0.575 -0.128 0.053 0.016 * 0.048 0.047 0.304 0.005 0.041 0.901
Father smoking
0.121 0.043 0.005 ** -0.018 0.058 0.760 -0.028 0.057 0.626 0.049 0.048 0.309
Mother smoking
0.020 0.075 0.794 0.114 0.089 0.202 -0.007 0.079 0.933 0.084 0.059 0.156
Physical Activity
-0.053 0.039 0.169 -0.054 0.045 0.237 -0.005 0.045 0.917 0.014 0.033 0.677
Time to watch TV per day
-0.004 0.012 0.767 0.009 0.016 0.545 0.004 0.013 0.763 -0.013 0.012 0.274
+: p<0.10; *: p<0.05; **: p<0.01; ***: p<0.001

139
Table 34: Estimated Parameters from the Model with Refusal Self-efficacy as Mediator by
Gender
Boy Girl
Estimate SE p Estimate SE p
Means

Intercept of SMK
1.167 0.834 0.162 0.369 0.146 0.012 *
Slope of SMK
0.344 0.703 0.625 -0.008 0.149 0.955
Intercept of RSE
1.170 0.226 0.000 *** 0.885 0.177 0.000 ***
Slope of RSE
-0.112 0.163 0.492 -0.015 0.122 0.903
Intercept of SMK

Intercept of RSE
-1.575 0.388 0.000 *** -0.384 0.095 0.000 ***
Slope of RSE
1.183 0.607 0.051 + 0.543 0.425 0.201
Slope of SMK

Intercept of RSE
0.061 0.298 0.838 0.114 0.086 0.184
Slope of RSE
-2.185 0.637 0.001 ** -0.674 0.397 0.089 +
Effects on Intercept of SMK

Intervention Program
-0.003 0.261 0.991 0.030 0.054 0.578
Depression
0.368 0.255 0.149 0.012 0.036 0.746
Program X Depression
-0.130 0.157 0.408 -0.017 0.024 0.480
Hostility/family conflict
-0.189 0.504 0.707 0.023 0.074 0.759
Program X Hostility
-0.082 0.262 0.754 0.011 0.051 0.829
Age
0.024 0.040 0.545 0.002 0.009 0.837
Education Performance
-0.011 0.037 0.757 -0.008 0.009 0.364
Peer Smoking
-0.011 0.033 0.738 -0.006 0.010 0.518
Pocket Money
-0.002 0.001 0.087 + 0.000 0.000 0.306
Access to cigarettes
0.051 0.063 0.416 -0.003 0.015 0.823
Father smoking
-0.124 0.104 0.232 0.003 0.015 0.848
Mother smoking
0.102 0.190 0.593 -0.016 0.026 0.534
Physical Activity
-0.062 0.072 0.391 -0.024 0.016 0.132
Time to watch TV per day
0.019 0.022 0.379 0.007 0.005 0.180
Genetic Variant
0.116 0.193 0.549 0.001 0.043 0.988
Genetic V. X Depression
-0.197 0.137 0.150 0.004 0.026 0.871
Genetic V. X Hostility
0.089 0.276 0.748 0.003 0.051 0.956
Genetic V. X Program
0.029 0.146 0.842 -0.015 0.031 0.626
Effects on Slope of SMK

Intervention Program
0.258 0.262 0.324 -0.054 0.044 0.219
Depression
-0.389 0.199 0.050 * -0.003 0.028 0.918
Program X Depression
0.103 0.129 0.424 0.032 0.020 0.118
Hostility/family conflict
0.079 0.442 0.859 -0.040 0.064 0.533
Program X Hostility
-0.176 0.251 0.484 -0.031 0.043 0.476
Age
-0.018 0.039 0.638 -0.001 0.008 0.866
Education Performance
-0.019 0.030 0.531 -0.017 0.009 0.057 +
Peer Smoking
0.042 0.029 0.148 0.003 0.009 0.685
Pocket Money
0.001 0.001 0.365 0.000 0.000 0.264

140
Table 34: Continued

Boy Girl
Estimate SE p Estimate SE p
Access to cigarettes
-0.062 0.058 0.285 -0.002 0.013 0.867
Father smoking
0.075 0.091 0.412 0.003 0.013 0.796
Mother smoking
0.093 0.155 0.549 -0.001 0.025 0.978
Physical Activity
0.080 0.065 0.216 0.008 0.014 0.557
Time to watch TV per day
0.003 0.022 0.897 -0.001 0.006 0.885
Genetic Variant
-0.153 0.180 0.393 -0.003 0.035 0.927
Genetic V. X Depression
0.220 0.118 0.061 + -0.010 0.020 0.621
Genetic V. X Hostility
-0.010 0.243 0.967 0.012 0.047 0.793
Genetic V. X Program
-0.190 0.132 0.150 0.020 0.030 0.495
Effects on Intercept of RSE

Intervention Program
0.060 0.077 0.440 -0.049 0.063 0.434
Depression
-0.082 0.056 0.144 -0.026 0.040 0.505
Program X Depression
-0.029 0.036 0.426 -0.028 0.025 0.253
Hostility/family conflict
-0.174 0.118 0.138 -0.188 0.102 0.064 +
Program X Hostility
0.103 0.074 0.161 0.023 0.063 0.720
Age
-0.018 0.016 0.261 0.011 0.012 0.372
Education Performance
0.040 0.010 0.000 *** 0.019 0.010 0.056 +
Peer Smoking
-0.053 0.006 0.000 *** -0.031 0.008 0.000 ***
Pocket Money
0.000 0.000 0.237 0.000 0.000 0.526
Access to cigarettes
-0.041 0.024 0.081 + -0.046 0.019 0.013 *
Father smoking
-0.032 0.024 0.180 -0.003 0.020 0.871
Mother smoking
-0.010 0.050 0.834 -0.034 0.036 0.335
Physical Activity
-0.025 0.022 0.238 -0.010 0.017 0.548
Time to watch TV per day
-0.017 0.007 0.016 * -0.008 0.006 0.172
Genetic Variant
-0.022 0.055 0.683 -0.054 0.059 0.357
Genetic V. X Depression
0.030 0.036 0.404 0.012 0.031 0.684
Genetic V. X Hostility
0.001 0.075 0.991 0.094 0.077 0.222
Genetic V. X Program
-0.045 0.043 0.295 0.049 0.042 0.250
Effects on Slope of RSE

Intervention Program
-0.077 0.062 0.214 -0.020 0.043 0.643
Depression
0.028 0.045 0.537 0.012 0.028 0.668
Program X Depression
0.009 0.028 0.758 -0.002 0.017 0.907
Hostility/family conflict
0.123 0.089 0.170 -0.008 0.066 0.903
Program X Hostility
-0.037 0.056 0.503 0.019 0.041 0.644
Age
-0.005 0.011 0.652 0.000 0.009 0.954
Education Performance
0.002 0.007 0.744 0.007 0.007 0.296
Peer Smoking
0.010 0.004 0.005 ** 0.016 0.005 0.001 **
Pocket Money
0.000 0.000 0.416 0.000 0.000 0.060 +
Access to cigarettes
0.010 0.017 0.540 -0.010 0.013 0.475
Father smoking
0.013 0.018 0.481 0.007 0.015 0.628
Mother smoking
0.014 0.033 0.672 0.006 0.021 0.769

141
Table 34: Continued

Boy Girl
Estimate SE p Estimate SE p
Physical Activity
0.044 0.016 0.006 ** 0.008 0.012 0.506
Time to watch TV per day
-0.005 0.005 0.316 -0.003 0.004 0.554
Genetic Variant
0.015 0.044 0.728 -0.019 0.039 0.628
Genetic V. X Depression
-0.011 0.029 0.705 -0.013 0.022 0.563
Genetic V. X Hostility
-0.055 0.058 0.342 -0.010 0.051 0.849
Genetic V. X Program
0.065 0.032 0.047 * 0.022 0.030 0.478
+: p<0.10; *: p<0.05; **: p<0.01; ***: p<0.001

142

Table 35: Estimated Parameters from the Model with Social Norm as Mediator by Gender
Boy Girl
Estimate SE p Estimate SE p
Means

Intercept of SMK
-0.52 0.782 0.507 0.042 0.104 0.688
Slope of SMK
-0.098 0.739 0.895 -0.037 0.145 0.796
Intercept of SN
-0.459 1.534 0.765 0.150 1.313 0.909
Slope of SN
2.119 1.155 0.067 + 1.380 0.909 0.129
Intercept of SMK

Intercept of SN
0.091 0.072 0.203 0.010 0.012 0.393
Slope of SN
-0.100 0.150 0.505 -0.016 0.037 0.674
Slope of SMK

Intercept of SN
0.076 0.068 0.264 -0.022 0.010 0.040 *
Slope of SN
0.370 0.173 0.033 * 0.102 0.058 0.076 +
Effects on Intercept of SMK

Intervention Program
-0.102 0.234 0.662 0.033 0.040 0.407
Depression
0.512 0.260 0.049 * 0.019 0.033 0.563
Program X Depression
-0.082 0.156 0.600 -0.009 0.022 0.684
Hostility/family conflict
0.091 0.468 0.846 0.089 0.055 0.102
Program X Hostility
-0.333 0.252 0.187 0.011 0.040 0.778
Age
0.020 0.037 0.589 -0.003 0.006 0.644
Education Performance
-0.081 0.044 0.068 + -0.012 0.009 0.189
Peer Smoking
0.068 0.023 0.002 ** 0.012 0.007 0.109
Pocket Money
-0.002 0.001 0.170 0.000 0.000 0.587
Access to cigarettes
0.082 0.060 0.170 0.005 0.012 0.648
Father smoking
-0.092 0.086 0.286 0.007 0.009 0.456
Mother smoking
0.081 0.204 0.692 0.000 0.024 0.992
Physical Activity
0.035 0.067 0.602 -0.016 0.013 0.203
Time to watch TV per day
0.033 0.022 0.129 0.006 0.005 0.222
Genetic Variant
0.219 0.182 0.228 -0.003 0.043 0.951
Genetic V. X Depression
-0.276 0.133 0.038 * -0.001 0.022 0.953
Genetic V. X Hostility
0.083 0.243 0.732 -0.046 0.039 0.238
Genetic V. X Program
0.148 0.130 0.254 -0.013 0.022 0.549
Effects on Slope of SMK

Intervention Program
0.308 0.269 0.252 -0.045 0.053 0.399
Depression
-0.331 0.219 0.132 0.021 0.037 0.572
Program X Depression
0.065 0.137 0.635 0.040 0.023 0.081 +
Hostility/family conflict
-0.087 0.425 0.838 -0.042 0.075 0.571
Program X Hostility
0.036 0.245 0.883 -0.031 0.046 0.497
Age
0.004 0.042 0.932 0.002 0.009 0.818
Education Performance
0.029 0.043 0.498 -0.017 0.009 0.070 +
Peer Smoking
0.001 0.022 0.963 -0.004 0.006 0.561
Pocket Money
0.002 0.002 0.296 0.000 0.000 0.439

143
Table 35: Continued

Boy Girl
Estimate SE p Estimate SE p
Access to cigarettes
-0.067 0.066 0.313 0.014 0.016 0.368
Father smoking
-0.008 0.080 0.923 -0.001 0.015 0.943
Mother smoking
0.135 0.172 0.435 -0.021 0.032 0.514
Physical Activity
-0.031 0.064 0.628 0.000 0.014 0.996
Time to watch TV per day
0.006 0.024 0.811 0.006 0.007 0.401
Genetic Variant
-0.109 0.187 0.563 0.065 0.055 0.241
Genetic V. X Depression
0.183 0.124 0.139 -0.027 0.028 0.333
Genetic V. X Hostility
0.051 0.236 0.828 0.051 0.059 0.392
Genetic V. X Program
-0.286 0.123 0.020 * -0.013 0.039 0.729
Effects on Intercept of SN

Intervention Program
-0.465 0.525 0.376 0.621 0.490 0.205
Depression
-0.192 0.380 0.614 0.520 0.271 0.055 +
Program X Depression
0.093 0.246 0.705 0.102 0.167 0.539
Hostility/family conflict
0.908 0.769 0.237 -0.086 0.677 0.900
Program X Hostility
0.158 0.473 0.738 -0.155 0.413 0.708
Age
0.182 0.106 0.087 + 0.018 0.093 0.846
Education Performance
-0.035 0.063 0.572 -0.104 0.075 0.165
Peer Smoking
0.179 0.032 0.000 *** 0.155 0.040 0.000 ***
Pocket Money
0.006 0.003 0.028 * 0.004 0.002 0.119
Access to cigarettes
0.346 0.151 0.022 * 0.180 0.136 0.187
Father smoking
0.466 0.153 0.002 ** 0.173 0.141 0.218
Mother smoking
0.272 0.287 0.344 0.242 0.258 0.349
Physical Activity
0.002 0.139 0.987 0.117 0.125 0.350
Time to watch TV per day
0.081 0.046 0.082 + 0.171 0.045 0.000 ***
Genetic Variant
-0.672 0.385 0.081 + 0.612 0.408 0.134
Genetic V. X Depression
0.348 0.251 0.166 -0.290 0.198 0.142
Genetic V. X Hostility
-0.395 0.484 0.415 0.649 0.485 0.181
Genetic V. X Program
0.137 0.266 0.609 -0.673 0.319 0.035 *
Effects on Slope of SN

Intervention Program
0.406 0.398 0.308 0.126 0.340 0.711
Depression
-0.270 0.267 0.312 -0.232 0.210 0.267
Program X Depression
0.025 0.179 0.890 -0.080 0.121 0.509
Hostility/family conflict
-0.461 0.529 0.383 -0.148 0.509 0.771
Program X Hostility
-0.405 0.333 0.224 -0.117 0.293 0.688
Age
-0.071 0.078 0.361 -0.013 0.063 0.837
Education Performance
-0.129 0.050 0.009 ** -0.053 0.051 0.297
Peer Smoking
0.005 0.026 0.845 -0.035 0.027 0.202
Pocket Money
-0.004 0.002 0.047 * -0.001 0.002 0.484
Access to cigarettes
-0.129 0.109 0.238 -0.112 0.096 0.242
Father smoking
0.049 0.117 0.677 0.035 0.103 0.732
Mother smoking
-0.255 0.203 0.209 0.173 0.189 0.360

144
Table 35: Continued

Boy Girl
Estimate SE p Estimate SE p
Physical Activity
0.030 0.101 0.765 0.037 0.087 0.667
Time to watch TV per day
0.000 0.034 0.991 -0.022 0.030 0.461
Genetic Variant
-0.055 0.289 0.849 -0.467 0.299 0.119
Genetic V. X Depression
0.083 0.177 0.639 0.202 0.155 0.193
Genetic V. X Hostility
0.239 0.337 0.478 -0.075 0.381 0.845
Genetic V. X Program
-0.143 0.200 0.475 0.102 0.243 0.674
+: p<0.10; *: p<0.05; **: p<0.01; ***: p<0.001

145
Table 36: Estimated Parameters from the Model with Attitude to Cigarette Use as Mediator by
Gender
Boy Girl
Estimate SE p Estimate SE p
Means

Intercept of SMK
-1.048 0.663 0.114 -0.047 0.108 0.666
Slope of SMK
0.450 0.576 0.434 0.088 0.084 0.296
Intercept of ACU
1.076 0.474 0.023 * 1.459 0.445 0.001 **
Slope of ACU
0.178 0.327 0.585 -0.078 0.312 0.803
Intercept of SMK

Intercept of ACU
0.264 0.146 0.071 + 0.044 0.022 0.044 *
Slope of ACU
-0.100 0.261 0.702 -0.046 0.067 0.496
Slope of SMK

Intercept of ACU
0.111 0.153 0.465 0.013 0.026 0.608
Slope of ACU
0.492 0.284 0.083 + 0.068 0.056 0.231
Effects on Intercept of SMK

Intervention Program
-0.095 0.267 0.722 0.026 0.041 0.518
Depression
0.499 0.237 0.035 * 0.026 0.026 0.328
Program X Depression
-0.108 0.168 0.520 -0.005 0.020 0.798
Hostility/family conflict
0.142 0.431 0.742 0.091 0.055 0.102
Program X Hostility
-0.257 0.253 0.309 0.006 0.043 0.884
Age
0.033 0.031 0.285 -0.002 0.007 0.749
Education Performance
-0.064 0.033 0.051 + -0.011 0.009 0.236
Peer Smoking
0.074 0.019 0.000 *** 0.013 0.007 0.047 *
Pocket Money
-0.002 0.001 0.173 0.000 0.000 0.740
Access to cigarettes
0.086 0.071 0.223 0.006 0.009 0.485
Father smoking
-0.039 0.103 0.709 0.008 0.009 0.375
Mother smoking
0.149 0.172 0.385 0.003 0.024 0.887
Physical Activity
0.017 0.066 0.798 -0.016 0.013 0.204
Time to watch TV per day
0.035 0.020 0.075 + 0.007 0.003 0.037 *
Genetic Variant
0.178 0.177 0.316 0.017 0.031 0.587
Genetic V. X Depression
-0.240 0.125 0.055 + -0.010 0.017 0.560
Genetic V. X Hostility
-0.019 0.239 0.936 -0.044 0.038 0.243
Genetic V. X Program
0.141 0.119 0.233 -0.015 0.021 0.497
Effects on Slope of SMK

Intervention Program
0.352 0.243 0.148 -0.047 0.033 0.154
Depression
-0.474 0.196 0.016 * -0.013 0.020 0.504
Program X Depression
0.105 0.137 0.445 0.029 0.017 0.092 +
Hostility/family conflict
-0.108 0.380 0.776 -0.055 0.046 0.232
Program X Hostility
-0.107 0.249 0.669 -0.038 0.035 0.281
Age
-0.006 0.037 0.878 0.000 0.005 0.932
Education Performance
-0.024 0.028 0.385 -0.020 0.009 0.027 *
Peer Smoking
0.011 0.020 0.576 -0.010 0.005 0.049 *
Pocket Money
0.001 0.001 0.365 0.000 0.000 0.703
Access to cigarettes
-0.067 0.064 0.298 -0.004 0.007 0.557

146
Table 36: Continued

Boy Girl
Estimate SE p Estimate SE p
Father smoking
0.022 0.094 0.817 -0.003 0.009 0.744
Mother smoking
0.041 0.144 0.774 -0.011 0.021 0.616
Physical Activity
0.001 0.063 0.988 0.002 0.010 0.874
Time to watch TV per day
0.009 0.022 0.700 0.000 0.005 0.990
Genetic Variant
-0.200 0.175 0.252 0.006 0.027 0.836
Genetic V. X Depression
0.245 0.117 0.036 * -0.001 0.014 0.912
Genetic V. X Hostility
0.091 0.232 0.694 0.028 0.032 0.383
Genetic V. X Program
-0.298 0.106 0.005 ** 0.013 0.022 0.540
Effects on Intercept of ACU

Intervention Program
-0.290 0.160 0.070 + 0.239 0.166 0.151
Depression
0.103 0.112 0.359 0.042 0.093 0.652
Program X Depression
0.100 0.067 0.133 -0.032 0.056 0.565
Hostility/family conflict
0.252 0.225 0.261 -0.009 0.229 0.969
Program X Hostility
-0.080 0.141 0.569 0.082 0.134 0.539
Age
0.039 0.033 0.230 0.000 0.032 0.995
Education Performance
-0.028 0.020 0.156 -0.011 0.023 0.621
Peer Smoking
0.038 0.009 0.000 *** 0.010 0.015 0.511
Pocket Money
0.001 0.001 0.124 0.002 0.001 0.101
Access to cigarettes
0.128 0.046 0.006 ** 0.093 0.047 0.045 *
Father smoking
-0.046 0.052 0.380 0.008 0.054 0.876
Mother smoking
-0.033 0.083 0.689 -0.046 0.073 0.531
Physical Activity
0.035 0.042 0.408 0.012 0.041 0.765
Time to watch TV per day
0.019 0.013 0.144 0.025 0.014 0.071 +
Genetic Variant
-0.038 0.116 0.741 -0.133 0.152 0.383
Genetic V. X Depression
-0.049 0.068 0.473 0.063 0.071 0.373
Genetic V. X Hostility
0.156 0.143 0.274 0.122 0.184 0.509
Genetic V. X Program
0.111 0.085 0.190 -0.168 0.115 0.145
Effects on Slope of ACU

Intervention Program
0.241 0.116 0.037 * -0.022 0.110 0.841
Depression
0.030 0.087 0.734 -0.018 0.062 0.774
Program X Depression
-0.081 0.052 0.118 0.012 0.037 0.754
Hostility/family conflict
-0.236 0.165 0.153 -0.012 0.156 0.938
Program X Hostility
0.042 0.101 0.676 -0.044 0.092 0.635
Age
-0.017 0.022 0.446 0.010 0.023 0.661
Education Performance
0.014 0.014 0.310 0.007 0.015 0.646
Peer Smoking
0.003 0.007 0.714 -0.012 0.008 0.152
Pocket Money
-0.001 0.001 0.015 * 0.000 0.001 0.836
Access to cigarettes
-0.070 0.032 0.028 * 0.028 0.031 0.372
Father smoking
0.067 0.035 0.052 + 0.015 0.037 0.682
Mother smoking
0.067 0.058 0.251 0.046 0.047 0.336

147
Table 36: Continued

Boy Girl
Estimate SE p Estimate SE p
Physical Activity
-0.047 0.029 0.109 -0.001 0.027 0.965
Time to watch TV per day
0.002 0.010 0.855 -0.004 0.009 0.651
Genetic Variant
0.041 0.086 0.630 -0.005 0.109 0.964
Genetic V. X Depression
0.006 0.053 0.913 0.011 0.048 0.820
Genetic V. X Hostility
0.009 0.104 0.927 -0.013 0.129 0.918
Genetic V. X Program
-0.097 0.060 0.105 0.001 0.079 0.994
+: p<0.10; *: p<0.05; **: p<0.01; ***: p<0.001

148

Figure 18: # of Cigarette Use/Day during Last Month at Baseline by Allele Size (Male)
0
0.1
0.2
0.3
0.4
0.5
3 repeats 4 repeats
# of Cigarette Use/Day Druing Last Month

Figure 19: # of Cigarette Use/Day during Last Month at Baseline by Allele Size and Program
(Male)
0
0.1
0.2
0.3
0.4
0.5
3 repeats 4 repeats
# of Cigarette Use/Day Druing Last Month
Control Program

149

Figure 20: # of Cigarette Use/Day during Last Month at Year 1 Follow-up by Allele Size (Male)
0
0.1
0.2
0.3
0.4
0.5
3 repeats 4 repeats
# of Cigarette Use/Day Druing Last Month

Figure 21: # of Cigarette Use/Day during Last Month at Year 1 Follow-up by Allele Size and
Program (Male)
0
0.1
0.2
0.3
0.4
0.5
3 repeats 4 repeats
# of Cigarette Use/Day Druing Last Month
Control Program

150

Figure 22: Change of Cigarette Use/Day during Last Month from Baseline to Year 1 Follow-up by
Allele Size (Male)
-0.2
-0.1
0
0.1
0.2
3 repeats 4 repeats
Change of Cigarette Use/Day Druing Last Month

Figure 23: Change of Cigarette Use/Day during Last Month from Baseline to Year 1 Follow-up by
Allele Size and Program (Male)
-0.2
-0.1
0
0.1
0.2
3 repeats 4 repeats
Change of Cigarette Use/Day Druing Last Month
Control Program

151

Figure 24: # of Cigarette Use/Day during Last Month at Baseline by Allele Size, Program, and
Depression Score (Male)
0
0.4
0.8
1.2
1.6
2
Low Depression & 3
repeats
Low Depression & 4
repeats
High Depression & 3
repeats
High Depression & 4
repeats
# of Cigarette Use/Day During Last Month
Control Program

Figure 25: # of Cigarette Use/Day during Last Month in Year 1 Follow-up by Allele Size,
Program, and Depression Score (Male)
0
0.4
0.8
1.2
1.6
2
Low Depression & 3
repeats
Low Depression & 4
repeats
High Depression & 3
repeats
High Depression & 4
repeats
# of Cigarette Use/Day During Last Month
Control Program

152
Figure 26: Change of Cigarette Use/Day during Last Month from Baseline to Year 1 Follow-up by
Allele Size, Program, and Depression Score (Male)
-1.5
-1
-0.5
0
0.5
Low Depression & 3
repeats
Low Depression & 4
repeats
High Depression & 3
repeats
High Depression & 4
repeats
Change of Cigarette Use/Day During Last Month
Control Program

Abstract (if available)

Abstract For adolescent smoking progression, understanding the potential risk factors, including genetic, psychosocial and other environmental factors, is important for identifying adolescents with high risk, and designing efficient prevention curriculum for deterring smoking development and nicotine dependence. The three studies of this dissertation, utilizing the sample from a longitudinal, randomized, school-based smoking prevention trial with 2661 urban 7th grade students at baseline, explore the patterns and risk factors for progressions of adolescent smoking and the potential moderating and mediating mechanisms among genetic and environmental risk factors.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

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

PDF

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

PDF

Psychosocial and cultural factors in the primary prevention of melanoma targeted to multiethnic children

PDF

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

PDF

Social self-control and adolescent substance use

PDF

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

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

Anxiety symptoms and nicotine use among adolescents and young adults

PDF

Adolescent social networks, smoking, and loneliness

PDF

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

PDF

Intrapersonal and environmental factors associated with Chinese youth alcohol use experimentation and binge drinking behaviors

PDF

Psychosocial and behavioral ractors associated with emotional eating in adolescents

PDF

Smoke-free housing policies and secondhand smoke exposure in low income multiunit housing in Los Angeles County

PDF

Competing risks: the role of the perceived consequences of refusing to share injection equipment among injection drug users

PDF

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

PDF

The role of acculturation in the sun-safe behaviors of US Latino adults in the United States

PDF

Using a structural model of psychopathology to distinguish relations between shared and specific features of psychopathology, smoking, and underlying mechanisms

PDF

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

PDF

Social network influences on depressive symptoms among Chinese adolescents

Asset Metadata

Creator Sun, Wei (author)

Core Title Genetic variants and smoking progression in Chinese adolescents

School Keck School of Medicine

Degree Doctor of Philosophy

Degree Program Preventive Medicine (Health Behavior)

Publication Date 06/18/2007

Defense Date 04/27/2007

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

Tag Adolescents,genetic,longitudinal,OAI-PMH Harvest,random-effects model,smoking

Language English

Advisor Johnson, Andy (committee chair), Conti, David V. (committee member), Richardson, Jean L. (committee member), Shih, Jean C. (committee member), Unger, Jennifer B. (committee member)

Creator Email wsun@usc.edu

Permanent Link (DOI) https://doi.org/10.25549/usctheses-m533

Unique identifier UC168650

Identifier etd-Sun-20070618 (filename),usctheses-m40 (legacy collection record id),usctheses-c127-508939 (legacy record id),usctheses-m533 (legacy record id)

Legacy Identifier etd-Sun-20070618.pdf

Dmrecord 508939

Document Type Dissertation

Rights Sun, Wei

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