Citation
Sociocultural influences on tobacco use by Turkish youth

Material Information

Title:
Sociocultural influences on tobacco use by Turkish youth findings from quantitative and qualitative methods
Creator:
Yuksel, Hulya
Place of Publication:
Denver, CO
Publisher:
University of Colorado Denver
Publication Date:
Language:
English
Physical Description:
248 leaves : ; 28 cm

Subjects

Subjects / Keywords:
Youth -- Tobacco use -- Turkey -- Dursunbey ( lcsh )
Youth -- Social conditions -- Turkey -- Dursunbey ( lcsh )
Youth -- Social conditions ( fast )
Youth -- Tobacco use ( fast )
Turkey -- Dursunbey ( fast )
Genre:
bibliography ( marcgt )
theses ( marcgt )
non-fiction ( marcgt )

Notes

Bibliography:
Includes bibliographical references (leaves 229-248).
Thesis:
Health and behavioral sciences
General Note:
Department of Health and Behavioral Sciences
Statement of Responsibility:
by Hulya Yuksel.

Record Information

Source Institution:
|University of Colorado Denver
Holding Location:
|Auraria Library
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
55626158 ( OCLC )
ocm55626158
Classification:
LD1190.L566 2003d Y84 ( lcc )

Downloads

This item has the following downloads:


Full Text
SOCIOCULTURAL INFLUENCES ON TOBACCO USE BY TURKISH
YOUTH: FINDINGS FROM QUANTITATIVE AND QUALITATIVE
METHODS
by
Hulya Yuksel
B.A., Aegean University
M.A., Aegean University
M.A., University of Colorado at Denver
A thesis submitted to the
University of Colorado at Denver
in partial fulfillment
of the requirements for the degree of
Doctor of Philosophy
Health and Behavioral Sciences
2003
i \ :
i


This thesis for the Doctor of Philosophy
degree by
Hulya Yuksel
has been approved
by
hhlo 3
Date


Yuksel, Hulya (Ph.D., Health and Behavioral Sciences)
Sociocultural Influences On Tobacco Use By Turkish Youth: Findings From
Quantitative And Qualitative Methods
Thesis directed by Professor Kitty Corbett
ABSTRACT
This study identified social and cultural factors that influence smoking
behavior of rural/urban youth, age 15-17, who attend high school in
Dursunbey, a small town in Turkeys Aegean region. The methods
consisted of a cross-sectional, school-based survey of all 10th graders
(N=211) in 5 high schools in Dursunbey, augmented by qualitative semi-
structured interviews with 42 10th grade students, 10 students from other
grades, 14 adult educators, and 4 other adults.
The survey identified several risk factors for current smoking of
adolescents (any smoking within the last 30 days): the number of close
friends who smoke and teacher smoking, male gender, age, and academic
achievement. Urban/rural origin, pocket money, family smoking behavior,
and school type were not statistically significant. Exposure to teachers who
smoked was associated with increased smoking risk. Males were at 12
times the risk of smoking.
Both adolescents and adults say that everybody smokes. Interview
data revealed an adolescent need to fit in with a group, a sense of shared
excitement in transgressing the rules imposed by adult authority, and
pressure resulting from a strong association of smoking with the Turkish


male identity. Students identified with their teachers more than their parents;
anti-tobacco messages at school were diluted by examples of teachers
smoking. Females are discouraged from smoking, a protection that appears
to be waning with Westernization and dramatically rising ever smoking rates.
Academic achievement appears protective against smoking, but with few
recreational alternatives, students referred to smoking as something to
pass the time. Students often described themselves as having no future,
since few students are accepted into college and jobs are scarce in Turkeys
depressed economy. The social environment has highly prevalent pro-
tobacco messages in the media, in popular songs, and in social and
religious rituals. In a culture where objections to second-hand smoke are
considered rude, expensive foreign cigarette brands are considered status
symbols by the young.
The conclusions of this study, the first with Turkish youth to employ
mixed methods, should contribute to more effective tobacco control
interventions, to the benefit of Dursunbey and other communities in Turkey.
This abstract accurately represents the content of the candidates thesis. I
recommend its publication.
Signed
IV


DEDICATION
This dissertation is dedicated with love and appreciation to Timothy, my
ultimate supporter and biggest fan. Your words of encouragement meant the
world to me. Thank you for always being there during my hardest times.
I would also like to dedicate this work to my parents and my uncle. My
favorite uncle, a life-long heavy smoker, was recently diagnosed with
advanced cancer.


ACKNOWLEDGMENT
First and foremost, I would like to thank my dissertation committee members
for their guidance and continuous assistance throughout this process.
Kitty (Corbett), it has been a privilege to work with you and have you as my
mentor. Your insightful comments and recommendations about the project
and future career decisions have helped me enormously. You encouraged
me to use both qualitative and quantitative methods that have yielded a
work rich in both substance and meaning. Thank you so much for always
believing in me.
Sharry (Erzinger), you were there from the beginning of this long journey
and closely followed my progress as a researcher. You were always eager
to hear about my interviews and field experience. I always felt that you
understood the nuances of a bilingual research work its promise of a
simultaneously rewarding and challenging process. Thank you for all your
encouragements and a kindness very unique to you.
Debbi (Main), you shared your expertise in mixed methods. Your guidance
was invaluable in formulating my study design. You encouraged me to
create a manageable project. Despite your busy schedule, you stayed the
course and kept the process going. Thank you for staying on my committee.
Augie (Diana), you were there throughout my journey and provided me with
some the most insightful comments. Your words of encouragement
motivated me to continue. I appreciate all your help, reading my documents
multiple times, your genuine interest in my topic and the long distance
phone calls from D.C. You kept the process positive for me. Thank you!
Jim (Nimmer), your generous support and guidance of my anxious journey
through statistical analysis helped me going. You did not mind working with
me many hours on my data. Thank you for agreeing to be on my committee
despite your busy schedule.
Candan (Duran-Aydintug), you were part of my shadow committee. Thank
you for agreeing to be on my committee so late in the game. Being from the
same country of origin connected us in a very special way. We did not have


to explain many things to each other. So many times, your jokes made my
day. Thank you!
I would also like to acknowledge and thank the following people and
organizations for their support and contributions: Tim Ene, a member of my
shadow committee, for your many insightful and intelligent ideas, and your
editing every word; I owe the most gratitude to you; Mansoureh Tajik, my
wise guide, for your insightful comments about methods, statistical analysis,
and structure of the dissertation; Jeff Schweinfest, who understood the
hoops of academia, for your insightful comments on my qualitative data
during the interpretation and your edits throughout my education; the
principals, teachers, and counselors of the five high schools in Dursunbey
especially Ahmet Hayat, Dursunbey Regular High School Counselor, for
their support during my data collection; my friend, for past 17 years, Nevin
Sayili-Demircioglu, for your help in the initial part of my research with the
logistics in Izmir, Turkey and for your help with the data entry process. Yusuf
Yuksel, my brother, your help with the logistics in Dursunbey during my data
collection and being my unofficial research assistant in the field. Chris Pon,
program assistant at the Health and Behavioral Sciences Program, it has
been great pleasure to work with you. Finally but not last, I would like to
thank Research for International Tobacco Control for their financial support
of my qualitative research.


TABLE OF CONTENTS
Figures ..........................................................xiii
Tables............................................................xiv
CHAPTER
1. INTRODUCTION.............................................1
A. Specific Aims......................................3
B. The Tobacco Use Problem in Turkey..................3
Smoking Prevalence...............................6
Health Awareness and Knowledge of Health
Consequences of Smoking.........................10
C. Actions of Transnational Tobacco Companies........12
History.........................................13
Marketing Strategies of Transnational Tobacco
Companies.......................................16
D. Tobacco Control Efforts and Tobacco Companies
Response..........................................17
E. Significance of the Study.........................18
F. Overview of the Chapters..........................21
2. THEORETICAL APPROACHES FOR UNDERSTANDING
SOCIAL INFLUENCES ON SMOKING BY YOUTH ..................22
A. Overview of Relevant Theories.....................22
B. Social Cognitive Theory...........................31
Early Behaviorist Origins of Social Learning Theory.32
viii


Vicarious Learning.................................32
Steps in the Acquisition of an Abusive Habit.......33
Social Cognitive Theory Evolves from Social Learning
Theory.............................................34
Reciprocal Determinism.............................34
The Constructs of Social Cognitive Theory..........35
C. Family, Peer, Friend, and Teacher Influences.........38
Peer and Friend Influences.........................39
Family Influences..................................41
Teacher Influences.................................42
School Policies....................................42
D. Sociodemographic Factors.............................43
Age and Gender.....................................43
Academic Achievement...............................44
Urban/Rural Location and Economics.................45
E. Social, Cultural and Psychological Factors in Turkish
Tobacco Use...........................................46
F. Application of Social Cognitive Theory: Study Hypotheses 48
Hypotheses ........................................50
Sociodemographic Factors...........................51
Social Environment Factors.........................52
3. METHODS....................................................55
A. A Study Design ......................................55
B. Field Season One in Izmir and Dursunbey, 2001........58
Construction of the Survey for the Field Season One .. 59
Focus Groups.....................................61
ix


Semi-Structured Interviews.......................61
Field Season One Data Collection and Sampling....62
Analysis.........................................63
Field Season One Study Results...................63
C. Field Season Two in Dursunbey, 2002.................65
The Survey in Dursunbey, 2002.....................65
Measures..........................................70
Data Entry........................................74
Analysis..........................................75
Semi-Structured Interviews in Dursunbey, 2002.....75
Methodological Limitations........................78
4. DURSUNBEY, ITS HIGH SCHOOLS, AND
CHARACTERISTICS OF THE STUDENTS............................79
A. The Town of Dursunbey................................79
B. High Schools in Dursunbey............................83
Dursunbey Regular High School.....................84
Religious High School.............................84
Health Vocational High School.....................85
Girls Vocational High School......................86
Industrial Vocational High School.................86
C. Characteristics of Students Who Were Surveyed.......87
Age and Sex Distribution..........................89
Family Origin.....................................90
Place of Residence................................91
Educational Attainment of the Parents.............92
x


Economic Status...................................92
Academic Success..................................95
5. SURVEY FINDINGS...........................................96
A. Introduction .......................................96
B. Section One: Data Integrity.........................96
C. Section Two: Descriptive Statistics.................98
Description of Sample.............................98
Description of Variables..........................99
Bivariate Analysis to Describe the Relationships
between Variables................................100
The Study Population and Smoking Behavior........102
D. Section Three: Test of Hypotheses..................109
Test of Sociodemographic Factors ................110
Test of Social Environment Factors...............114
E. Section Four: Model Testing .......................118
Sociodemographic Factors and Smoking Behavior....118
Social Environment Factors and Smoking Behavior..122
Sociodemographic and Social Environment Factors ....124
Other Survey Findings............................126
6. DISCUSSION................................................129
A. Introduction ......................................129
B. A Review of Findings...............................130
Incidence of Ever-Smoking........................130
Incidence of Current Smoking.....................131
Influence of Sociodemographic Factors............131
Influence of Social Environmental Factors........140
XI


C. Theoretical Implications of the Findings........149
Limitations ..................................159
7. CONCLUSION............................................162
A. Summary of Findings................................162
Recommendations for Future Studies and For
Intervention...................................164
Policy Implications............................168
APPENDIX
A. SURVEY INSTRUMENTS AND SEMI-STRUCTURED
INTERVIEW GUIDE.......................................173
Preliminary Survey (English, June 2001)...............174
Preliminary Survey (Turkish, June 2001)...............180
Final Survey (English, April 2002)....................186
Final Survey (Turkish, April 2002)....................194
Semi-Structured Interview Guides......................201
B. FIELD SEASON ONE QUALITATIVE AND QUANTITATIVE
FINDINGS, SUMMER 2001.................................203
C. TURKISH SONGS REGARDING SMOKING AND POEMS
BY ONE OF THE STUDY SUBJECTS (English Translation)...216
D. CONSENTS AND HUMAN SUBJECTS APPROVALS.................220
BIBLIOGRAPHY................................................229
xii


FIGURES
Figure
3.1 Timelines for the Data Collection and Write-Up...............56
4.1 Map of Turkey................................................81
6.1 Predictors of Adolescent Current Smoking....................151
6.2 Reciprocal Interaction of the Predictors of Adolescent Current
Smoking....................................................152
6.3 Partial Model for Rural Environment Using Qualitative and
Quantitative Findings......................................154
6.4 Partial Model for School Environment Using Qualitative and
Quantitative Findings......................................157
xiii


TABLES
Table
2.1 Selected Factors Influencing Tobacco Use.....................49
3.1 Structure of the Questionnaire...............................60
3.2 Definitions of Adolescent Smoking.............................71
3.3 Explanatory (Independent) Sociodemographic Variables..........72
3.4 Explanatory (Independent) Social Environment Variables........74
4.1 School Type...................................................83
4.2 Age Categories................................................90
4.3 Family Origin.................................................91
4.4 Place of Residence............................................91
4.5 Parents Education............................................92
4.6 Monthly Household Income......................................93
4.7 Perceived Economic Situation..................................94
4.8 Weekly Student Allowance......................................94
4.9 Academic Achievement..........................................95
5.1 Variables with Missing Values.................................97
5.2 Description of the Study Population...........................99
5.3 Descriptive Statistics of Variables..........................100
5.4 Correlations Between Variables...............................101
XIV


5.4 Correlations Between Variables (Continued)...............102
5.5 Smoking Behavior in the Study Population.................104
5.6 Sociodemographic Factors and Smoking Behavior............106
5.7 Social Environment Factors and Smoking Behavior..........108
5.8 Hypothesis Testing for Sociodemographic Factors..........112
5.9 Hypothesis Testing for Social Environment Factors........115
5.10 Full Model for Sociodemographic Factors and Current
Smoking Behavior.........................................120
5.11 Final Model for Sociodemographic Factors and Current
Smoking Behavior.........................................122
5.12 Full Model for Social Environment Factors and Current
Smoking Behavior.........................................123
5.13 Final Model for Social Environment Factors and Current
Smoking Behavior.........................................124
5.14 Final Model for Predicting Adolescent Current Smoking
Behavior.................................................125
5.15 Companions When Adolescents Try Smoking..................126
5.16 Why do you think people of your age start to smoke?......127
5.17 Why did you start to smoke?..............................128
6.1 Influences in the Partial Model for Rural Environment....155
6.2 Influences in the Partial Model for School Environment...158
xv


CHAPTER 1
INTRODUCTION
the base of our business is the high school student Source: Lorillard
Tobacco Company, 8/30/78 Bates No.03537131/7132
todays teenager is tomorrows potential regular customer Source:
Philip Morris Tobacco Company, 3/31/1981, Bates No. 1000390803
Despite the massive literature on social influences on youth smoking
in the developed world (Conrad et al., 1992; Evans et al, 1988; Flay & Phil,
1985; Flay, 1993; Flayetal., 1994, 1997, 1999; Flay & Petraitis, 1994;
Petraitis et al., 1995; Shadel et al., 2000; USDHHS, 1994), findings about
factors contributing to uptake and addiction cannot be assumed as
applicable and useful across different social, political and cultural settings.
Patterns of tobacco uptake and maintenance must be understood and
addressed in specific contexts. The normative environments in developing
countries such as Turkey differ significantly from the environments of
Western nations. Countries differ widely in the political and economic
context of daily life. Developing countries have higher mortality rates and
greater uncertainty concerning the near-term future, as well as a lower
likelihood of a long life. Smoking and smoking-related behavior function as
key social symbols, and smoking control campaigns based on an inaccurate
understanding of the social context in which smoking occurs can reinforce
the very behavior they are designed to prevent (Eckert 1983).
1


Understanding of perceptions and attitudes about smoking, as a
social practice and a health risk must be situated within a broader set of
sociocultural concerns. Sound scientific findings specific to every nation are
needed in order to carry forward effective tobacco control policies and
programs. As was asserted by Flay & Petraitis (1991:84), social
environment variables predict tobacco use. Therefore the structure of the
economic, legal, social, and educational systems of a society are
determinants of behavior.
In an effort to understand tobacco use in Turkey, and ultimately to
design effective policies and interventions, I carried out research with 10th
grade students in a west-central Turkish town, Dursunbey, that includes
both urban and rural populations. The aim of the study was to explore the
social, environmental and sociodemographic factors that influence the
smoking patterns of high school youth, in the critical age group of 15 to 17
years. The research employed qualitative data from interviews with sets of
adult and adolescent informants, and a quantitative study of a cross-
sectional survey of all 211 tenth grade students at the communitys five high
schools. The research design allowed description of tobacco knowledge,
attitudes, and behaviors, tests of a series of hypotheses, and development
of a predictive model of smoking use by youth using logistic regression
techniques. Social Cognitive Theory, with emphasis on its environment,
observational learning, and situation constructs, served as the theoretical
foundation for the research (Akers, 1998; Bandura & Walters, 1963;
Bandura, 1977, 1986, 1991; Baranowski etal, 1997). The intent of this
research is to provide Turkey with its own base of data that is useful for
tobacco interventions.
2


A. Specific Aims
This research identifies social factors, as defined in the Social
Cognitive Theory (Bandura, 1986, 1991), that are associated with tobacco
use in Dursunbey, Turkey. Specific aims are:
Aim 1: To explore social environmental and situational factors that are
associated with smoking behavior of 10th grade students, including
associations of adolescent smoking behavior with:
smoking among family members
smoking among teachers
smoking among friends
friends approval of smoking behavior
perceived smoking prevalence among peers
Aim 2: To explore how patterns of smoking for 10th grade students are
associated with the following sociodemographic factors:
gender
age
place of origin (rural/urban)
school type
perceived economic situation
weekly allowances
perceived academic achievement
To accomplish these aims, the researcher collected the data in Turkish.
Translation occurred during the data analysis. Hypotheses based on Social
Cognitive Theory and the preliminary work are explained in Chapter 2.
B. The Tobacco Use Problem in Turkey
Tobacco-related diseases are the single largest cause of preventable
death and disability among adults in the world (Chandler, 1986; Stebbins,
1990; Trevathan et al.; 1999). Worldwide almost 47% of men and 12% of
3


women smoke (World Bank, 1999). Developing countries face a huge and
growing tobacco-related burden of human suffering, mortality, and economic
losses. By 2030, 7 million of the 10 million people who die from tobacco-
related disease will be in developing countries (Mackay & Crofton, 1996;
WHO, 2002). Adult cigarette consumption in low and middle-income
countries rose by 64% from 1971 to 1991. Almost five million people died
from smoking-related diseases across the world in 2000, researchers
estimate (WHO, 2002).
It is estimated that in high-income countries, 14,000-15,000 children
and young people take up smoking, versus 68,000 to 84,000 children and
young people in low and middle-income countries (World Bank, 1999). The
Global Youth Tobacco Survey, carried out under the auspices of the World
Health Organization (WHO) and Centers for Disease Control and Prevention
(CDC), tracks the prevalence of tobacco use and provides baseline data
about youth (students ages 13-15) across the world. This study revealed
that smoking ranged from a high of 33% to a low of 10% in fourteen
developing countries around the world (Warren et al., 2000). Individuals who
begin smoking at a young age are more likely to become heavy smokers
(World Bank, 1999; Zhang et al., 2000). They are also more likely to die
from smoking-related diseases later in life. To prevent smoking and other
tobacco use, programs logically target youth. Understanding youth uptake of
tobacco is necessary in order to design policies and programs that can
effectively counter the influences fostering initiation and regular use.
Although studies throughout the world suggest shared factors ranging
from gender differences to interpersonal and intrapersonal influences
(Chassin et al., 1990; Conrad et al., 1992; Flay & Petraitis, 1991, 1994; Flay
et al., 1999; Petraitis et al., 1995; Rissel et al., 2000), efforts to reduce
tobacco use in developing countries are characterized by over-reliance on
4


research carried out in, by, and for persons in wealthier Western nations.
Studies in high-income countries have found that most tobacco addiction
begins at a young age, as a result of a broad array of factors implicated in
tobacco initiation, and have developed interventions targeting youth
(Kessler, 1995; Smith et al., 1999; Tilly, 1991). Adolescents in western
cultures frequently report curiosity, social norms, and social pressure as
reasons for smoking uptake (Sarason et al., 1992). For much of the rest of
the world, relatively little is known about specific attitudes, knowledge,
behaviors and social influences that contribute to starting or avoiding
tobacco use. Culture-specific information to classify the reasons adolescents
engage in tobacco use is required if we are to design effective smoking
prevention programs (Sarason et al., 1992).
The larger context may also contribute to Turkeys tobacco problem.
Turkeys population of 68 million is geographically and culturally positioned
at the crossroads of Europe and Asia. With 40% of its population under 18
years of age and growing at a 1.2% annual rate, Turkey has a very young
population. According to 1999 data, 65% of the population lives in urban
areas. The overall literacy rate for the general population is 86.5% (94% for
males and 79% for females). Based on the World Fact Book (2002 est.),
Turkey is considered a low-income country with an estimated current GDP
of $7,000 per capita. Turkey has been experiencing a series of economic
crises for over a decade, the most severe hit Turkey in February 2001, after
the government was forced to let its currency float. Since that time, the
Turkish lira has lost 50% of its value against the US dollar, a fact that is a
continuing source of stress for most workers, businessmen and blue-collar
workers alike. This, combined with a 10.8% unemployment rate (plus
underemployment of 6.1%) (2002 estimate) and a near-hyperinflation rate of
5


45.2 % (The World Fact Book, 2002, est.), has created a very unstable
environment for Turkish citizens.
Smoking Prevalence
Overall Prevalence. Smoking in Turkey is a widely practiced, socially
accepted habit. With its consumption of 121 billion cigarettes a year in 2001,
Turkey has one of the largest markets for tobacco products in the world
(Karanci & Rustemli, 1987). Its 2001 annual cigarette consumption is a 21%
increase from its 1999 annual consumption of 100 billion cigarettes. Annual
per capita cigarette consumption in Turkey rose from 1,230 in 1985 to 1,495
in 1992. For persons above 15 years of age, it increased from 1,850 in 1965
to 2,600 in 1992 (Firat, 1996). Adult smoking rates vary in different studies.
Adolescent smoking rates also show a good deal of variation among various
studies. In the recent Tobacco Control Country Profiles (WHO, 2003)
smoking rates for 20 years and older males was 50.9% and 10.9% for
females. These most recent rates are lower than previous studies in Turkey.
In the 2000 nationwide study, smoking rates for 18 years and above ages
were reported as 58.7% for males and 26.4% for females (A&G Research
Reports, 2000). The problem is not limited to urban areas in Turkey.
Metintas et al. (2000) identified smoking as one of the most important public
health problems in rural areas within the governmental district of Eskisehir.
In their survey of adults ages 20 and older in 52 villages and 19 sub-centers,
the frequency of smoking was 44.3% in villages and 48.7% in sub-centers (a
terminology used by Metintas et al.).
Among certain Turkish occupational groups, smoking rates were
reported as much higher than the national average. In Dedeoglu et al.s
(1994) study in Antalya, 53% of male physicians and 41% female physicians
6


were smokers. Smoking rates are high among teachers also. In Saltik et al.
(1989) study, 62% of male teachers and 34% of the female teachers were
smokers. In Bilir et al.s (1997) study in Ankara, 64% of journalists, 51% of
teachers, 44% of physicians, 46% of artists, 27% of parliamentarians, and
35% of professional sportsmen were smokers. For some occupational
groups, smoking rates climbed even higher, reaching 74.3% among taxi and
truck drivers. In Bagci et al.s study (1996), 95% of drivers were smokers.
The high percentage of heavy smokers is also of concern in Turkey. In
earlier studies, 64% of the adult population (15 years in age and older)
indicated that they smoke one pack or more cigarettes a day (PIAR 1988, in
Asut, 1993).
Unlike high-income countries, health care professionals smoke at
high rates in Turkey. Individuals in Turkey who should best understand the
dangers of smoking, i.e., doctors and medical students, appear uninfluenced
by information. The smoking rates of male medical doctors (62.4%) matches
the smoking rates of the general population of adult males (62.5%). A
comparison of smoking patterns among medical school versus non-medical
school students revealed that the prevalence of smoking among first-year
medical students was 11.8% vs. 21.0% among non-medical students
(Kocabas et al., 1994). There was a significant jump in the smoking rates in
the last year of school for students in both groups, increasing to 30.2%
among medical students and 32.4% among non-medical students.
Gender and Tobacco Use. Gender differences are important in
tobacco use in Turkey. In Turkey many of the professional women and
women who live in urban areas are more likely to smoke than women in
rural areas. Female smoking rates are reported to vary by occupation in
several studies (Bilir and Onder, 2000; Firat, 1996). Saltik et al.s (1988)
7


study found a 11 % smoking rate for female doctors in Elazig. In a more
recent study, Bilir et al. (1999) found a 34.4% smoking rate for female
physicians nationwide. Kocabas et al. (1994) found that female doctors and
medical students were more likely to smoke than women in general: 44.1 %
of female doctors and 34.9% of female final year medical students were
smokers, while only 10.9% of the females in the general population smoke.
The high smoking rates among female professionals are illustrative of
recent trends in womens smoking in Turkey. Traditionally in the Middle East
and many Islamic countries, smoking has been seen as an inappropriate
and immoral behavior for females (Dagli, 1999). The figures above,
however, indicate that Turkish womens attitudes toward smoking are rapidly
changing. This was also evident from a 1996 study of 538 randomly
selected women in Istanbul, finding an overall smoking prevalence of 48%
(Dagli, 1999). This pattern is replicated elsewhere in the Middle East.
Maziak (2002) points out that smoking is associated with social liberalization
among women in Syria.
It is a well-known fact that transnational tobacco companies have
identified women as an important target (WHO, 1996), using the Western
symbol of the liberated woman to sell cigarettes to the Third World
(Headden, 1998). Especially in developing countries, smoking has been
promoted as a way of liberalization and Western modernism for women. Not
surprisingly, a 1996 industry review noted that the anticipated growth in
cigarette sales in Turkey is based on the increase in smoking amongst
women and the lack, at the present, of a strong anti-smoking lobby
(Hammond & Purcell, 1999:1).
Gender plays a role in recreational activities in Turkey, with potential
implications for tobacco use. Particularly in small towns, the sexes often
8


separate for evening relaxation. The men frequent all-male coffee houses,
where tobacco smoke is always densely present. The women gather in their
homes for social teas. To be able to speak freely, they usually send their
young adolescent males away. This usually results in these boys spending
more time with peers in the evenings, sometimes even in coffee houses.
Minors are not officially allowed into these places, where the association of
smoking with the male identity is almost unavoidable.
Youth and Tobacco Use. The smoking rate among Turkish youth
ages 15 to 18 was reported to be 20%, with 23.1% for males and 14.8% for
females (WHO, 2003). Data from 1986 on youth smoking reported higher
rates for males (31 %) and lower rates for females (5%) between the ages of
15 and19 (Hammond & Purcell, 1999). The prevalence of smoking among
youth has been provided by several other recent studies (Alikasifoglu et al.,
2002; Bilir et al., 1997, 2000; Celik et al., 2000; Derman et al., 1995;
Erbaydar et al., 2002; Emri et al., 1998; Karababa et al., 1997; Ozcan&
Ozcan, 2002). In Ankara, the capital of Turkey, 33% of high school students
were classified as ever smokers, and 28.3% (32 % boys and 20% girls,
respectively) were current smokers, with a mean number of 7 cigarettes per
day (Bilir et al., 1997). In Bilir et al.s (2000) nationwide study among 10th
graders, 32% of students were ever-smokers (64% for boys and 22% for
girls), and 18% of the boys and 11 % of the girls were current smokers. In
the same study, there were high numbers of ex-smokers, 18% of boys and
11% of girls.
The most current nationwide data on 10th graders reported that 58%
of the students had ever tried smoking, and 33% of boys and 18.3% of girls
between ages 15-17 reported they were current smokers (Erbaydar et al.,
2002). A marked increase in the rate of smoking by girls is clear: 1986 WHO
9


data showed that 31 % of boys and 5% of girls ages 15 to 19 were smokers
in Turkey, an increase particularly for girls.
Karababa et al. (1997) surveyed 990 Apprenticeship School students
in Izmir. Fifty-four percent of boys and 53% of girls had tried smoking, but
identified themselves as nonsmokers currently. Twenty seven percent of
boys and 37.1% of girls were found to be current smokers, with an average
age of 17 and reported age of uptake between 13-15; this was the only
study that demonstrates higher smoking rates for girls than boys.
A survey of 1066 high school students in Manisa, Turkey (Celik et al.,
2000) found that smoking prevalence among male students was 17.2%, and
7.1% among female students. The mean age for beginning smoking was 13.
Health Awareness and Knowledge of Health
Consequences of Smoking
According to Bilir and Onders recent report (2000), 70-100 thousand
Turkish people die due to smoking-related health problems each year, 10-15
times the number who die in motor vehicle accidents in Turkey. The most
frequent cause of death in Turkey is cardiovascular disease, followed by
cancer and then cerebrovascular diseases. In Turkey, cancer incidence has
tripled between 1983 and 1990, with bronchoalveolar cancer identified as
the primary cause of cancer related deaths (Bilir et al., 1997). Thus smoking
plays an increasingly important role in the increased incidence of cancer. It
is estimated that smoking may be responsible for 50% of cancers in Turkey.
A 1996 study shows that of the major of causes of cancer deaths in Turkish
men, carcinoma of the lung accounts for 30-40% of deaths and carcinoma of
larynx accounts for 7.5% of cancer deaths. Moreover, lung cancer is the
10


leading cause of cancer death in females, with a 16% relative mortality rate
(Firat, 1996).
The effects of smoking are not simply a matter of premature death.
Smoking is also an added burden in terms of health costs and misery,
predominantly in regions of the earth where misery is already higher than in
most developed part of the world. A portion of these diseases will not be
reflected in higher healthcare costs because, in many countries, these
diseases will go untreated. By the same token, it is difficult to estimate
direct and indirect healthcare cost of providing treatment to the Turkish
smokers, and the loss in Turkish productivity. Tobacco-related disease costs
US more than $123 billion annually with $50 billion in avoidable medical
expenses and $73 billion in lost productivity. Over 25 years, this could
amount to $3 trillion (Moskowitz, 1998). For developing countries struggling
to provide basic services and a subsistence-level existence for their citizens,
almost any loss in productivity is tragic. Over a period of a decade, 0.5%
annual loss in productivity in a country translates to a 5% relative loss of
wealth, potentially affecting the well being of thousands of people.
In addition to productivity losses, cigarette purchases divert badly
needed family income. As was noted above, an estimated 10 to 15% of
Turkish household budgets are spent on tobacco. It is far more efficient for
a family to simply put their income to better use.
In a survey of 1066 high school students in Manisa, Turkey (Celik et
al., 2000), 95% of the students stated that they were aware that cigarettes
were harmful and 95.8% knew about the close relationship between
smoking and lung cancer. Thus knowing the adverse health effects of
smoking does not stop young people in Turkey from becoming smokers.
Similarly, in another study in Istanbul, Turkey 100 medical students were
asked about the adverse health effects of smoking. One hundred percent of
11


students were aware of the adverse health effects of smoking but this
knowledge was not enough to stop them from smoking (Derman et al.,
1995).
C. Actions of Transnational Tobacco Companies
Turkey is a tobacco growing country; an estimated 500,000 farmers
grow tobacco (Dagli, 1999). Their families account for 3 million people.
Tekel, the state monopoly, is a huge state industry covering both tobacco
and alcohol production and employing more than 36,000 people (Tobacco
Reporter, 2001). The Turkish government has been following the
recommendations of the International Monetary Fund (IMF) to privatize the
state monopoly for many years, although this effort has been a very
controversial and is moving slowly. Philip Morris is especially keen to see
Tekel taken out of government hands. In 1997, Marco Terribilini, General
Manager of Philip Morris in Turkey, said:
Turkey is a significant market with major potential.... Only 18%
of the population is above 45 years of age....The Turkish
monopoly has to be privatized totally....Philip Morris will play a
role under any condition in this privatization (Dagli, 1999:117).
When the privileges of the Turkish state monopoly were rescinded in
1990, the economic consequences for the Turkish population were
significant. Despite the dramatic increase in Turkish tobacco consumption,
use of domestically produced tobacco actually decreased. With the rise of
tobacco imports, local farmers lost their market and their subsidies. In 1991,
329,000 tons of excess Turkish tobacco were burned. The percentage of
American tobacco in domestically consumed cigarettes rose from 0% in
1987 to 66% in 2000. Three million people were affected by these changes.
12


Many Turkish export companies went out of business. Aside from the
obvious economic impact, the battle against Turkish youths tobacco use
became a battle against formidable American tobacco companies and their
deep pockets. Currently Philip Morriss enjoys a quarter share of the Turkish
cigarette market.
History
Until the early 1980s, Tekel held exclusive rights to manufacture and
sell tobacco products. The monopolys run-down factories produced inferior
quality cigarettes that were harsh tasting (Country Case Study, 1999). In
1984 President Turgut Ozal, a former employee of the World Bank, led a
drive to privatize the Turkish markets and prepare Turkey for membership in
the European Union. As part of that drive, Tekel lost many of its monopoly
rights, opening up its market to foreign tobacco imports. However, Tekel
maintained control over pricing and distribution. Consequences of this
decision included:
An increase in tobacco imports as tobacco advertising and promotion
became more sophisticated and aggressive, targeting women and
the young for the first time.
Overall consumption per capita rose 23% from 1985 to 1991, in spite
of the skewing effects of a high birth rate. For the population over 15
years of age, per capita consumption rose 41% from 1965 to 1992.
As cigarette consumption increased, the amount of Turkish tobacco
used in the cigarettes decreased. In 1991, Tekel had to burn
329,000 tons of excess tobacco.
Many Turkish export companies went out business, dropping from 59
firms to 14.
In 1987, Turkish cigarettes contained no American tobacco. In 2000,
the amount was 66,000 tons.
13


In 1990, Philip Morris signed an agreement with Sabanci Holdings, a
powerful Turkish company exerting considerable influence on Turkish
tobacco policy. The parties agreed that the government regulations would
change, effectively exempting Philip Morris from any regulation of its prices,
distribution and sales practices (Dagli, 1997). Philip Morris paid Margaret
Thatcher, former prime minister of the United Kingdom, $1 million US for her
lobbying efforts in foreign countries (Cunningham, 1996). When she visited
Turkey in 1992, the press reported that the purpose of her visit was to help
Philip Morris gain greater access to Turkish markets. Neither Ms. Thatcher
nor Philip Morris denied the allegation (Dagli, 1997). In 1993, R.J. Reynolds
created its Turkish cigarette factory with $100 million in US capital. Thomas
Reibl, the CEO of R.J Reynolds, said at the time:
We have put the locomotive on the right tracks. Now our aim is
to speed it up. We will take advantage of this country. Turkey
is one of the worlds biggest markets. We have good
production, we are working with good people, have good
strategies. Turkey is promising with its young population
(Dagli, 1997:116).
In 1998, a partnership agreement was signed between Tekel and
British American Tobacco (BAT), the worlds third largest tobacco company.
These developments were followed by aggressive tobacco marketing by
Transnational Tobacco Companies. As the Wall Street Journal notes:
[T]he battle plan Philip Morris used to capture Turkey is the
same one it has perfected setting up operations in nearly 30
countries worldwide. First, it lobbied heavily to eliminate the
governments control of tobacco prices, enlisting the help of
one of Turkeys most influential businessmen. Then it poured
tens of millions of dollars into a state-of-the-art cigarette
factory, where, competitors say, it carefully engineered its
cigarettes to appeal to Turkish taste buds, but with a stronger
kick than local smokes. It dispatched salesmen to 130,000
stores across the country, successfully recruiting many of the
14


mom-and-pop shops to support a vast in-store marketing plan.
And it spent lavishly on an advertising blitz of cowboys and
panoramic vistas, selling a cellophane-wrapped piece of
America for $1.20 (Hwang, 1998; Hammond et a!., 1999).
Turkish smokers marvel at Marlboros smooth taste and
powerful buzz, as compared to harsher, local cigarettes made
with Turkish oriental tobacco, which is naturally lower in
nicotine content than the American leaf tobacco Philip Morris
imports here.
Turkey has seen financial and economic power used to employ
influential political personalities, skirt laws, and dominate foreign commercial
and governmental interests. Virtually unlimited funds fueled tobaccos legal
machine, such that it never lost a lawsuit until relatively recently. As
Cunningham states (1996:26):
One representative of an advertising agency working for a
tobacco company said that the client had more money than
God. Tobaccos enormous profit potential provides a
tremendous incentive for keeping the industrys sales high.
When it comes to efforts to combat tobacco control measures,
the industry can write a blank check at any time.
The entry of transnational tobacco companies in foreign markets is
correlated with increased consumption (Klebnikov, 1994; McKerrow, 1992;
Poole, 1998; Tilly, 1991; Yach, 1986). Transnational tobacco companies
promote increased consumption by out-marketing and out-spending the
domestic producers. These companies are often far more sophisticated in
their marketing strategies and promotional campaigns than the domestic
producers. Tekel, for instance, never advertised. This is generally true of
indigenous tobacco companies (Makary, 1998).
Current research shows that the entrance of multinational tobacco
companies into the Turkish market since 1984 has correlated with the
increase in smoking prevalence, and increase in relative mortality from lung
15


cancer for men from 25% to 40% and for women from 11 % to 16% between
1985 and 1996 (Firat, 1996; Kocabas et al., 1994).
Marketing Strategies of Transnational Tobacco
Companies
Turkeys population is becoming particularly young. Tobacco has
strong inducements for targeting a growing youth market (Emri, 1997),
although Cunninghams general assessment applies: There is as much
justification for promoting tobacco as there would have been for promoting
rats during the Plague (Cunningham, 1996:66). Transnational tobacco
companies marketing is backed by well-developed and time-tested
strategies. Naturally, these strategies are particularly directed at young
people (Headden, 1998; Nichter & Cartwright, 1991). The following is a
sample of these strategies:
Transnational tobacco companies distributed free cigarettes at
disco parties, advertised heavily in youth magazines, and
offered free tickets to a concert by a Taiwanese pop star in
return for five empty cigarette packages (Tilly, 1990:12).
One Turkish high school senior reported that he was handed
five free packs a day from sales representatives from Philip
Morris and competitor R.J. Reynolds Tobacco Company
(Hwang, 1998).
Marlboro organized a music disco show on 31 May 1991 -
World No-Smoking Day- at the Marmara University campus,
Istanbul, Turkey (Dagli, 1997:116).
During the passage of key anti-tobacco legislation in 1996 in
Turkey: ...tobacco control advocates started to hear that
newspaper bosses who were going to lose advertising
revenue, as well as advertisers, were upset about the bill.
16


...eleven major newspapers ran the same headlines: smoke
and be more intelligent, referring to an unpublished study
presented at a meeting in the USA. Very kindly, Keith Ball who
was phoned offered help. He woke Prof Stanton Glantz (it was
4 am in the USA) to ask about the study. Prof Glantz found the
paper and faxed it in an hour. It had been misinterpreted by
the Turkish press. (Dagli, 1997:116).
That aggressive and successful marketing targets children is
reflected in Emri and Bagcis (1998) study examining the recognition rates of
cigarette brand names and logos by primary schoolchildren (age 8 or
younger) in Turkey. They found that the Camel logo and the Samsun and
Marlboro brand names were the most highly recognized of ail product logos
and brand names tested. Recognition of cigarette brands averaged 95.2%;
by comparison, the recognition of Chee-tos (a childrens snack food) was
58.1 %. It is ironic that worldwide mortality rates for children have dropped
dramatically, only to have those children become tobacco customers at risk
of premature death as adults (Nichter & Cartwright, 1991).
D. Tobacco Control Efforts and Tobacco Companies
Response
In 1988, the Minister of Health prepared Turkeys first anti-tobacco
legislation, which was unanimously accepted at the parliament on January
9, 1991. President Ozal vetoed the bill, claiming that the restriction of
advertising conflicted with freedom of trade. The Minister of Health resigned
in protest. In 1996, Turkey enacted antismoking legislation that banned
cigarette advertising from printed media, TV, and radio, and ends any
tobacco company sponsorship of sporting events. Moreover, the sale of
cigarettes to anyone under 18 was prohibited, and smoking was restricted in
17


all public areas (Emri, 1997). Various other measures have been
suggested, such as worldwide bans on tobacco promotion, banning sales to
minors, improved health warnings, setting limits to harmful cigarette
additives, and establishing smoke-free areas, and increasing taxes. The
estimated price elasticity of demand for cigarettes implies that a 10% price
increase will reduce consumption by 2% in the short run and 4% in the long
run (Tansel, 1993).
Turkish anti-tobacco legislation has been circumvented in various
ways. Philip Morris and RJ Reynolds, for example, retailored their
marketing strategies in which they print price announcements to
newspapers which often announce that prices have not changed (Dagli,
1999; Simpson, 1997). In response to the ban on advertising, Philip Morris
simply removed the name Marlboro" from its advertisements, but left its
distinctive chevron; the brands red and white chevron is seen everywhere.
Hundreds of small convenience stores are capped with lighted Marlboro
signs. In Istanbuls trendy Ortakoy neighborhood, sidewalk cafes use red
and white Marlboro umbrellas and trays, provided free by Philip Morris. A
19-year-old pointing to a red chevron in one store explains, This doesnt
say Marlboro, but we understand it. The law didnt change anything. Philip
Morris also hosts parties at nightclubs and sponsors contests. The former
head of Philip Morriss operations in Turkey said, Of course, the first thing
you do is paint the town red...we painted every building (Country Case
Report, 1999).
E. Significance of the Study
Crosscurrents of cultures, demographics, and economic stresses
make Turkey a compelling locale for studying tobacco use among youth.
18


The severity of Turkeys economic outlook has given todays Turkish
adolescents the bleakest future in several generations. Simultaneously,
Turkey continues to move westward in its cultural orientation, and the
population is increasingly exposed to media images and marketing of
tobacco products.
Serious gaps in tobacco control literature exist between Turkey and
the developed world, where most formal studies have focused on the
incidence of consumption rather than the causes of consumption (Bilir et al.,
2000). While the prevalence of smoking among Turkish youth has been
examined, there have been very few studies addressing adolescents
reasons and motivations for smoking. Bilir et al.s study (1997) of high
school students in a single municipal area, Ankara, found that 44% of high
school students stated that they smoke more with their friends. Thirty one
percent of the students think that smoking helps to build a relationship, while
14% think that they seem more mature while smoking. Some adolescents
smoke for relaxation, some for pleasure, and others to impress others. Sixty-
nine percent of the high school students smoke when they are angry and 62
% smoke at rest. Twelve percent do not know what to do with their hands if
they do not smoke. The reported reasons for not smoking included illness,
parental influence, and cost. In addition, parental modeling may be involved
in Turkish youth's development of smoking habits.
In Karababa et al.s (1997) study, no relation was found between
student smoking and parents' education or family structure, but parents and
colleagues smoking habits were positively correlated with student smoking.
Erbaydar et al. (2002) and Ozcan & Ozcan (2002) identified risk factors for
smoking among youth as follows: male gender, parent, relative and friend
smoking, problems with family and school, dysfunctional family life, low
grades, higher socioeconomic status, having more pocket money, spending
19


more time with friends after school and in the evenings, feeling lonely,
alcohol use, and neighborhood and school effects.
In summary, what is known about the social influences on youth
smoking in Turkey is the product of a handful of studies. More studies are
needed, especially studies which employ a clear theoretical framework.
Culture-specific information about the reasons adolescents give for
engaging in tobacco use is required in order to design effective smoking
prevention programs (Sarason et al., 1992). A better understanding of
smoking behaviors will contribute to more effectively targeted tobacco
control and health messages aimed at Turkish youth.
This study uses survey to measure the risk associated with the social
influences of gender, age, parents, siblings, relatives, friends, peers,
teachers and schools. It also assesses the sociodemographic influences of
gender, age, economic situation, academic achievement, and urban/rural
location. Ethnographic interviews analyzed the first-hand impressions of
students about their experience of smoking uptake and the influences that
affected their decisions to smoke, including their recreational activities,
ambitions, frustrations, and plans for the future. Moreover, they highlighted
findings concerning smoking influences related, for example, to custom and
gender. Students impressions are cross-validated with the reports of adults
such as teachers and school counselors. In this manner, issues of school
policy and the adult perspective of the changing social environment can be
taken into account. By combining the qualitative findings from interviews
with the quantitative findings from the survey, a richer, more culturally
specific description of the problem emerges, leading to a detailed model of
the social influences on youth smoking. This provides conclusions that
contribute to more effective tobacco control interventions, to benefit
Dursunbey and other communities in Turkey.
20


F. Overview of the Chapters
This thesis is organized in eight chapters. Chapter One has provided
an overview of the research aims, and background on tobacco use in
Turkey. Chapter Two discusses theoretical approaches relevant to the
study, with emphasis on Social Cognitive Theory, and presents the study
hypotheses. Chapter Three describes the research design, preliminary
studies, methods for the survey and its analysis. Chapter Four presents the
setting, describing Dursunbey and its high schools, and characteristics of
youth who participated in the survey are described. Chapter Five presents
findings from hypothesis testing, the predictive model, and other survey
results. Chapter Six is a discussion of survey findings, with elucidation
provided by findings from the studys qualitative data. Chapter Seven is a
conclusion, including implications for future efforts to understand and control
tobacco use in Turkey.
21


CHAPTER 2
THEORETICAL APPROACHES FOR
UNDERSTANDING SOCIAL INFLUENCES ON
SMOKING BY YOUTH
A. Overview of Relevant Theories
Although most of the literature on social influences of youth smoking
reflects studies carried out in high-income countries, those studies provide a
reasonable beginning point. A variety of theories, employing equivalent or
overlapping constructs but differing in emphasis and directionality, have
been used to explain factors relevant to smoking onset. Earlier theories are
broadly categorized as intrapersonal models or interpersonal models of
behavior. While the intrapersonal models used in tobacco research focus on
the personal traits, states and skills of individuals as determinants of
smoking behavior, the interpersonal models emphasize interpersonal
interactions and the social environment that provide reinforcements,
resources, and learning situations that influence smoking behavior (Lewis,
1997). Later models of health behavior combine these intrapersonal and
interpersonal elements into broader models of influences where personal
characteristics and the social environment interact.
This chapter reviews literature on Social Cognitive Theory, social
influences, and sociodemographic influences, covering the topical areas of
role modeling, the influences of family, friends, peers, and teachers who are
significant in the lives of adolescents. Socio-demographic factors are then
treated, such as gender, age, academic achievement, urban/rural location,
22


and disposable income. This is followed by an examination of social, cultural
and psychological factors that are special to Turkey and other developing
countries. The chapter concludes with a presentation of how Social
Cognitive Theory is applied in this study in the development of the study
hypotheses.
In general, the behavioral theories in tobacco control have grown in
sophistication and complexity as they attempt to adequately model the
influences documented in the literature. The next few pages review a variety
of commonly used theory in tobacco literature. Flay et al. (1999) emphasize
the key role of these multifaceted theories in producing effective anti-
smoking interventions, stressing the need for strongly theory-directed
studies that take into account the mutual influence of intrapersonal factors,
culture, and the social environment.
Intrapersonal models in an approximate chronological order of
development, include the Health Belief Model (Hochbaum, 1958, Janz &
Becker, 1984), the Theory of Planned Behavior and the Theory of Reasoned
Action (Ajzen & Fishbein 1975), and the Stages of Change Model (Flay et
al., 1983, Prochaska & DiClemente, 1982; Prochaska et al., 1987). The
Health Belief Model describes health behaviors as the outcomes of personal
cost-benefit decisions. It posits six determinants of behavior change:
perceived susceptibility; severity of a disease; perceived benefits and
barriers to action; internal or external cues to action; self-efficacy (ones
confidence in one's ability to take action) (Bandura, 1997, Strecher &
Rosenstock, 1997). Perceived barriers have been the most consistent
predictor of preventive health behavior, and perceived severity of
consequences of disease has been the weakest predictor (Zimmerman &
Vernberg, 1994). Health Belief Model attempts to understand the
determinants of health-related behaviors at the individual level in order to
23


stimulate positive behavioral changes. It sees the individual as the only unit
of intervention; groups, organizations, worksite and communities are almost
absent from Health Belief Model.
The Theory of Reasoned Action and the Theory of Planned Behavior
(an extension of Theory of Reasoned Action) are the most important
conceptualizations of the attitude-behavior relationship (Zimmerman &
Vernberg, 1994). Generally, the strongest relationship has been found
between attitudes and intentions, rather than between intentions and
behavior (Zimmerman & Vernberg, 1994). Behavioral intentions to act and
the role of normative beliefs in mediating between behavior and action are
emphasized in both theories. The attitude toward the act is the product of
outcome beliefs and outcome evaluations, and the subjective norm is the
product of beliefs about how significant others feel about the behavior and
motivations to comply with those other significant others (Zimmerman &
Vernberg, 1994: 48). Thus, these two models include perceptions of others
in their decision-making framework. Studies in the US show that nationwide
use of specific substance among high school students usually declines with
the increased perceived risk and social disapproval associated with the
substance (Petraitis et al., 1995:69). In addition to attitudes and normative
beliefs, the Theory of Planned Behavior posits that self-efficacy (i.e.,
perceived control over ones behavior) will affect the intentions and
behaviors. Self-efficacy has two important forms: use self-efficacy and
refusal self-efficacy. The first form represents adolescents beliefs in their
abilities to obtain and use substances. The second form represents
adolescents beliefs in their abilities to resist social pressure to begin
substance use (Petraitis et al., 1995). This theory has been used to predict
smoking behavior of youth by various researchers (Carvajal et al., 2000;
Flay et al., 1998; Norman and Tedeschi, 1989). In several studies, intention
24


to smoke predicts the smoking behavior several years later (Flay et al.,
1998).
In a limited sense, the Theory of Reasoned Action/Theory of Planned
Behaviors inclusion of normative beliefs makes them attractive theories for
studying social influences. They also include intentions and self-efficacy in
their intrapersonal model. One problem, however, is that they do not
explicitly include important forms of social learning such as observational
learning. In spite of their use of norms, their greatest problem is that their
treatment of the environment is limited to that aspect of the social
environment, such that other features like social policy and physical
surroundings are largely excluded.
The Stages of Change model sees behavior as a process involving a
several stages (Prohaska et al., 1997). The stage theory has important
implications for understanding the different stages of youth smoking, which
are usually identified as never users, triers, experimenters, and regular
users (Flay et al., 1998). The focus of this model is plainly intrapersonal to
the virtual exclusion of interpersonal interactions.
The preferred theories used in this research are Social Learning
Theory and Social Cognitive Theory. Social Learning Theory (Miller and
Dollard, 1941; Akers et al., 1977) is an interpersonal theory that forms the
basis for the more extensive Social Cognitive Theory (Bandura, 1986). The
central principle of Social Learning Theory is observational learning, where
individuals modify their behavior based on observing the behavior, rewards,
and punishments of others. Social Learning Theory allows for role modeling,
modeling of behavior, and social normative beliefs. (Flay et al., 1991). Like
the Theory of Reasoned Action, the individual learns to adapt his/her
behavior to the normative beliefs and expectations of others. Social
Cognitive Theory expands on the principles of Social Learning Theory,
25


incorporating Reciprocal Determinism as its central concept. In Social
Cognitive Theory, interactions may occur between a person and his
environment (either social or physical), a person and his behavior, and his
behavior and his environment. Any of these three interactions may flow in
either direction; for example, a person may influence his social environment,
as well as be influenced by that environment. Social Cognitive Theory not
only includes interpersonal constructs, but intrapersonal constructs as well:
emotional coping responses, self-efficacy, values that a person places on
outcomes (expectancies), and anticipated outcomes of behavior
(expectations) (Bandura, 1997). Bandura (1997) extends the concept of
reciprocal determinism to self-influence, to include self-control and self-
reward. With these constructs, Social Cognitive Theory is more than an
interpersonal theory, offering a group of interactions among person,
environment, and behavior. The framework of Social Cognitive Theory
provides a good basis for describing social influences on smoking behavior,
and was selected as the theoretical framework for my study. Social
Cognitive Theory will be reviewed in greater detail later in this chapter.
A model of social influences is incomplete without accounting for
differences among individuals; a given social influence on tobacco use will
not yield the same behavior for all individuals. This variation in response to
exogenous influences is partially explained by mediating processes and
moderating processes (Flay et al., 1997). In a mediation process, a variable
of interest has influenced one or more intermediate variables, which in turn
influences the behavior. In a moderating process, the influence of a variable
is altered (strengthened, reduced, delayed) by another variable, typically a
sociodemographic variable such as gender, SES, age, or ethnicity (Flay et
al., 1997). To accurately gauge the effect of a social influence, it is therefore
important to include sociodemographic measures as control variables. This
26


applies an adjustment for the mediating influence of potentially skewed
sociodemographics in the study population, and also furnishes valuable data
concerning subsets of the study population that may be at higher risk than
the general population. Given these concerns, my study includes a number
of sociodemographic variables in its measures.
Comprehensive understanding of tobacco use in Turkey requires
multiple diverse investigations into various dimensions of the problem. This
section presents a broad view of the literature addressing tobacco onset but
focuses on the interpersonal or social influences and sociodemographic
influences. The topics on adolescent smoking introduced are smoking
prevalence, smoking uptake, addiction, gender, age, SES, academic
achievement, peer influence, family smoking, and media influences.
Smoking rates vary by gender, as indicated by a 47% smoking rate
for males and a 12% smoking rate for females (the worldwide prevalence is
29%) (World Bank, 2000). These rates differ by stage of national
development; high-income countries have a 39% male smoking rate, versus
a 49% male smoking rate for low and middle-income countries. For women,
the difference is reversed; 22% of women in high-income countries smoke,
as opposed to 9% in low and middle-income countries. In the United States,
rates of smoking uptake declined for male adolescents from 45% in 1974 to
33% in 1985 (Fiore et al., 1989). During the same period, rates of uptake for
female adolescents remained unchanged at 34%.
Young people have reasons for initiation and maintenance of
smoking. In the literature related to initiation and maintenance of smoking
the following factors are widely listed for beginning smoking: curiosity, social
norms, and social pressure (Sarason et al., 1992); peer pressure (Barber et
al., 1999; Brynin, 1999; Smith et al., 1999; West et al., 1999); and signs of
independence and Western modernism (Headden, 1998; Zhang et al.,
27


2000). Relaxation, pleasure, addiction, and an effort to impress others
(Sarason et al., 1992) have been proffered as reasons for the maintenance
of smoking. As Denscombe (2001: 161) states:
[Tjhere is a need to investigate the meaning which smoking
holds for young people. There is a need to understand how
they view smoking and what significance it holds in terms of
their perceptions of themselves. And there is a need to link
the findings to the social context within which young people
find themselves at the start of the new millennium.
His method focuses on the hermeneutic understanding of human
action in relation to some wider whole that gives it meaning. Several studies
discuss smoking in terms of adolescent needs for identity. And desirability
and in the context of tobacco consumption. Nichter and Cartwright (1991:
430) nicely state this as follows:
[R]ather than being regarded simply as objects to consume,
cigarettes become indices of social membership for
adolescents who are searching for their identities or who want
to escape the immediate reality into which they have been
born.
The impact of consumption on the construction of identities is
emerging as a key concern of social theory (Miles et al., 1998). Since
cigarettes are one of many consumer products, teenagers might perceive
some meaning in them that goes beyond their face value. Simmels (1971)
discussion about the role of fashion as a means of constructing a
relationship between individual and social concerns supports this idea.
Given that most tobacco addiction begins at a young age,
interventions must be targeted at the youth in the developing nations
(Kessler, 1995; Smith et al., 1999; Tilly, 1991). Nicotine addiction starts with
tobacco consumption. Addiction is sustained by its psychopharmacological
28


relation with dopamine and the psychological state of satisfaction, and is
characterized by compulsive drug-seeking and use-even in the face of
negative health consequencesand also by neurochemical and molecular
changes in the brain. Tobacco use certainly fits both of these criteria (NIDA,
1999). Most smokers accept that tobacco is harmful (perceived severity of a
disease) and express a desire to reduce or stop using it, and nearly 35
million of them in the US make a serious attempt to quit each year.
Unfortunately, less than seven percent of those who try to quit on their own
achieve more than one year of abstinence; most relapse within a few days
of attempting to quit (1990 National Household Survey, in Heishman et al.,
1997).
Genes involved in the brains dopaminergic reward mechanisms are
one possible source for genetic factors that influence smoking practices. To
date, the strongest evidence involves polymorphism (variation) in the
dopamine transporter (SLC6A3) gene. This gene codes for the dopamine
transporter protein (DAT) which transports released dopamine, thereby
limiting the level and duration of dopamine receptor activation. Recent
studies suggest that variation in SLC6A3 is associated with the likelihood of
being a smoker, age of smoking onset, and length of periods of smoking
abstinence (Lerman, 2000).
The prevalence statistics above show that males are at greater risk of
smoking than females (Malcon et al., 2003; Flay et al., 1999). Of all the
sociodemographic variables, gender is a frequent moderating variable in
other influences on smoking. Some studies show that gender moderates
social influences such as population migration (Brunswick and Messeri,
1984), level of parent education (Ary et al., 1988), number of friends who
smoke (Brunswick and Messeri, 1984), mothers smoking status on male
childrens smoking status (Skinner et al., 1985), parents smoking status on
29


female childrens smoking status (Charlton and Blair, 1989; Chassin et al.,
1986) , mothers smoking status on male childrens uptake (Skinner et al.,
1985), sibling smoking status on female childrens uptake (Mittelmark et al.,
1987) , friends smoking status on female adolescents (Charlton and Blair,
1989). Although gender is covered in a separate section in the literature
review, some gender-specific findings are reported in other sections, for
example that females are susceptible to peer pressure starting at ages 14 or
15 (Barber et al, 1999) and that teachers intervene more often for male
students (de Moor et al., 1992). Smoking has been linked to greater self-
confidence and social skills in girls (Clayton, 1991; Michell and Amos, 1996),
but less self-confidence in boys (Clayton, 1991). Other studies find that: girls
are more susceptible to parent and teachers smoking influences than boys
at ages 12-13 (Lloyd-Richardson et al., 2002); concerns about body weight
leads to uptake in females (Gritz and Crane, 1991; Camp et al., 1993); boys
initiate smoking earlier than girls (Lucas and Lloyd, 1999); and smoking is
linked with female eating disorders (Crisp et al., 1999) and issues of power
(Fisher, 1976). Smoking has been linked to other sociodemographics
factors: increasing age (Alexander et al., 1983), low SES (Maziak, 2002),
and pocket money (Alexander et al., 1983). It has also been linked to low
academic ambition (Bauman et al., 1984), low grade point averages
(Newcomb et al., 1989), and low academic achievement (Bailey and
Hubbard, 1990; Choi et al., 1997).
Social influences have been documented in relation to smoking.
Among these have been peer pressure (Barber et al., 1999; Brynin, 1999;
Smith et al., 1999; West et al., 1999) and social norms supporting tobacco
use (Sarason et al., 1992). Smoking uptake has been linked to parents
smoking (Chassin et al., 1984; Jackson and Henriksen, 1997), although
some studies have found that only the fathers smoking was significant
30


(Ahlgren et al.f 1982) or only the mothers smoking (Skinner et al., 1985).
Although Alexander et al. (1983) found that sibling smoking was a significant
influence, West et al. (1999) found this effect in younger adolescents ages
15 and 16. The incidence of ever-trying has been associated with exposure
to positive portrayals of smoking in film (Sargent et al., 2001). Gray et al.
(1996) found that 12-13 year olds have very concrete and literal
interpretations of images of cigarette use in magazines that lead to negative
impressions of smoking, but 15-16 year olds interpretations were more
abstract, leading to either negative or neutral impressions of smoking.
B. Social Cognitive Theory
Social Cognitive Theory emphasizes observational learning and, role
modeling, as well as the mutual influences of individual, behavior, and
society. It provides a theoretical structure for parental modeling, differential
peer association, positive or negative attitudes toward smoking, perceived
positive and negative consequences of smoking, the integration of smoking
into lifestyle, and the interaction of the social environment with individual
characteristics. In sum, Social Cognitive Theory furnishes a large variety of
modeling tools for studies involving social influences on individual behaviors
such as smoking, providing a solid theoretical foundation for studies such as
this.
Social Learning Theory (Akers et al., 1977) and Social Cognitive
Theory (Bandura, 1977) are two theories categorized under interpersonal
models. Social Learning Theory (SLT) presents a commonly used
perspective on smoking. Social Learning Theory can be categorized under
socialization theories in sociology. Socialization is the process by which
31


individuals learn to become members of society, both by internalizing the
norms and values of society, and also by learning to perform their social
roles (Marshall, 1998).
Early Behaviorist Origins of Social Learning Theory
Early learning theories explain behavior is mechanistically, in which
human behavior explained based on positive and negative reinforcements
that elicit the desired behavior (Baranowski et al., 1997:154). B.F. Skinners
behaviorist theory is essential to the early definition of social learning model.
He describes the conditions under which behavior is modified, not explaining
the mechanism by which this process occurs.
In attempting to apply Skinners methods to smoking cessation, the
proper reinforcements for the desired behaviors would have to be identified.
Here addiction provides an interesting challenge to the behaviorist model,
since addiction provides powerful motivations for continuing smoking that
are not visible to the conditioner. After all, for the addict, the effects of
addiction carry with it some perceived benefits such as relaxation, stress
relief, and stimulation. This demonstrates that although the behaviorist
approach may be useful in a limited sense, it does not furnish all the tools
necessary for modifying behaviors under all circumstances such as nicotine
addiction.
Vicarious Learning
On the other hand, Bandura and Walters (1963) asserted that rewards
do not have to be directly applied to the individual for learning to occur.
Instead the individual can learn from observing the behavior and the
consequent reward of others. In the case of the perceived rewards of
32


smoking in youth, social acceptance and social facilitation is a common
experience of adolescents. Applying the Social learning theory to smoking, it
includes parental modeling, differential peer association, positive or negative
attitudes toward smoking, perceived positive and negative consequences of
smoking, and integration of smoking into lifestyle (Spear & Akers, 1988).
People also learn norms, attitudes, and orientations by interacting with their
social groups. Social groups control individuals major sources of
reinforcement and punishment and expose them to behavioral models and
normative definitions (Akers et al., 1979:838). Under SLT/SCT, learning
may even occur without direct observation of others (Petraitis et al., 1995).
Spoken endorsements by role models may be sufficient for youth to adopt
positive beliefs about substances like tobacco.
Steps in the Acquisition of an Abusive Habit
Social Learning Theory identifies a three-step sequential process in the
abuse of substances by youth, starting with influences external to the
individual and ending with altered internal characteristics (Petraitis et al.,
1995). Key to this process is the adoption of cognitive dispositions, called
definitions in Social Learning Theory, about the substance. The first step is
imitation, in which youths model their own behaviors or attitudes on others
behaviors by simply observing and replicating the behaviors or, transposing
them into forms more acceptable to the youth lifestyle. The second step is
social reinforcement in which adolescents internalize definitions and exhibit
behaviors and values approved by significant others. In the final step, the
individuals own expectation of physical and social rewards perpetuates the
use of the substance.
33


Social Cognitive Theory Evolves from Social Learning
Theory
Social Cognitive Theory (SCT) is an extension of social learning
theory, and has its roots in reinforcement theory (Baranowski et al., 1997).
It begins with the assumption that individuals are more likely to engage in
behavior they find enjoyable or makes them feel good about themselves, or
that they know significant others find enjoyable. It integrates concepts of
self-efficacy, emotional coping responses, and cognitive world-modeling into
an evolved form of Social Learning Theory (Petraitis et al., 1995). Bandura
may have renamed his enhanced SLT as the Social Cognitive Theory in
order to clearly distinguish SCT from its early behaviorist roots (Stone,
1998).
Reciprocal Determinism
Social Cognitive Theory explains human behavior by the reciprocal
determinism among behavior, environmental determinants, and personal
factors (Bandura, 1977, 1997). This is the central organizing principle of
SCT:
In social cognitive theory, human agency operates within an
interdependent causal structure involving triadic reciprocal
causation (Bandura, 1986a). In this transactional view of self
and society, internal personal factors in the form of cognitive,
affective, and biological events; behavior; and environmental
events all operate as interacting determinants that influence
one another bidrectionally (Bandura, 1997:5).
SCT identifies three types of interaction under reciprocal determinism
(Bandura, 1977). Person-behavior interactions specify that ones behavior is
influenced by ones thoughts and feelings, and that ones behavior will
34


conversely influence ones mental life. While the environment may influence
ones behavior, ones behavior may also change the physical and social
environment. Finally, ones personal characteristics, either mental
characteristics such as intentions, beliefs, perceptions, and expectations, or
physical characteristics such as gender and behavioral pre-dispositions
(Stone, 1998), may be influenced by the environment. Likewise, the
environment may change personal characteristics. These interactions
among may be two-way or three-way, simultaneous or delayed. The
behavioral, environmental, and personal phenomena modeled under SCT
may therefore be very complex.
The Constructs of Social Cognitive Theory.
Bandura posits ten constructs within the SCT model: Environment,
Situation, Observational Learning, Behavioral Capability, Reinforcements,
Outcome Expectations, Outcome Expectancies, Self-Efficacy, Self-Control
of Performance, and Emotional Coping Responses (Baranowski et al.,
1997). The Environment construct is meant to represent those factors that
are external to the individual, consisting of the physical environment and the
social environment (Baranowski et al., 1997). Just as the Environment
construct describes factors external to the individual, the Situation construct
represents the individuals perception of elements in the environment. This
not only includes an awareness of the objects and actions in the
environment, but also the roles of those objects and even cues about the
social acceptability of social behaviors (Baranowski et al., 1997). Since this
construct involves perceptions, those mental representations of the
environment may be mistaken or distorted (Baranowski et al., 1997). Unlike
factors in the Situation construct, factors in the Environment construct can
35


influence behavior without any awareness of that factor (Baranowski et al.,
1997).
Humans often learn new behaviors in a social setting by observing
the behaviors of others and discerning the desirable/undesirable outcomes
for those others. The advantage of this vicarious learning is that outcomes
are more predicable, often involving less time and risk than learning based
on direct experience (Baranowski et al., 1997). This learning activity falls
under the Observational Learning SCT construct. Individuals engaged in
observational learning form a model of the environment and the functioning
of the learned behavior in the environment (Bandura, 1986). This is the
basis for family members having similar behavior patterns (Baranowski et
al.,1997). Not all observationally learned behaviors are effectively put into
practice. The Behavior Capability construct specifies that behavior has two
prerequisites: knowledge of the behavior and knowledge of how to perform
the behavior (Baranowski et al., 1997). This construct reflects the intellectual
capacity, training, and skill needed to engage successfully in the behavior.
Social Learning Theory and Social Cognitive Theory suggest that the most
effective way to prevent tobacco uptake is to decrease the appeal of
tobacco advocates and users as role models, while simultaneously
increasing the appeal of role models who abstain from tobacco use (Petraitis
et al., 1995). In sum, role models are important factors in the initiation and
maintenance of tobacco use. Individuals with friends who smoke are more
likely to feel peer pressure.
The Reinforcements construct consists of three kinds of
reinforcement (Baranowski et al., 1997). Direct reinforcement is operant
conditioning, having positive and negative rewards. Vicarious reinforcement
is synonymous with observational learning. Self-reinforcement refers to self-
control. Reinforcements may be either external or internal under Social
36


CognitiveTheory, where external reinforcements have predictable value as
seen by an independent observer of an event, whereas internal
reinforcements are based on the individuals own perception of the
reinforcement value of the event.
In the Outcome Expectations construct (originally called antecedent
determinants), the individual projects an outcome of a behavior, anticipating
a result (Baranowski et al., 1997). Smoking may be seen as pleasurable, for
example, because the individual has learned to associate that outcome with
the behavior. This is not to be confused with the Outcome Expectancies
(originally incentives) construct, in which the outcome of a behavior is
valued either positively or negatively. According to Baranowski et al. (1997),
the timing of the outcome falls under this construct, where the actualization
of a value may be immediate or delayed. This is particularly significant for
smoking behavior, where immediate gratification is often chosen over long-
term health benefits (McAlister, 1980).
Self-Efficacy is a construct closely associated with self-confidence
that one can perform an activity (Baranowski et al., 1997). Self-confidence
is an important aspect of behavioral change, since it influences the level of
investment an individual makes (Bandura, 1977). It also plays a role in the
refusal of youths to take up smoking (DeVries et al., 1990). Bandura (1991)
specifies several components in the Self-Control construct: setting goals,
monitoring of ones behavior, assessing its effects against goals, and self-
reward and self-reprimand for ones behavior. In the Emotional Coping
Responses construct, excessive emotions such as anxiety and hostility are
recognized as an inhibitor of useful learning (Baranowski et al., 1997).
Several techniques for overcoming these responses have been identified.
These include stress management, effective problem solving, and problem
restructuring. These techniques have particular relevance for smoking
37


cessation interventions, and work well with the Social CognitiveTheory
(Baranowski etal., 1997).
Social Cognitive Theory is complex, having disparate constructs
addressing a broad range of determinants of behavior, ranging from external
environment factors to learning paradigms, to internal mental dispositions
and beliefs, to disciplines of self-regulation. The number of choices of
construct combinations is large, and the choice of interactions adds yet
another layer of complexity. For this reason, studies based on Social
Cognitive Theory usually limit themselves to utilizing one or two constructs
(Stone, 1993).
Having completed the review of Social Cognitive Theory, this chapter
continues with a review of the literature concerning the influences that are
relevant to my study. These include several aspects of the social
environment: peer, friend, parent, sibling and teacher influences, as well as
school policy. Sociodemographic and personal factors are also reviewed:
age and gender, academic achievement and urban/rural location. A review
of the literature specific to Turkey follows.
C. Family, Peer. Friend and Teacher Influences
Role models have a powerful influence on youths use of tobacco and
other experimental substances, supporting both the Social Learning and
Social Cognitive theories (Petraitis et al., 1995). The influence of parents
and friends are issue specific in which parental influences are stronger for
future roles; peer influences are stronger for issues related to immediate
adolescent life style. Tobacco use is one of many adolescent behaviors that
result from interaction between individual characteristics and the competing
38


influences of multiple social groups (Kandel, 1980: 256).
Peer and Friend Influences
Peer influence has two aspects: First, the idea that peers have a
distinct culture in which they use sanctions to enforce group norms. Second,
the peer group is comprised of individuals who share certain key attributes,
such as age, gender, and ethnic origin (Denscombe, 2001). Peer pressure
and appearance-consciousness play a role as well. In Brynins (1999)
survey of adolescents aged 11-15 in 5000 households in the US smoking is
believed to make young people look more fashionable, older, and sexier.
Kandel et al. (1978) found positive correlations between marijuana
use and friendships where her subjects were exposed to drug, alcohol and
tobacco use, were offered those substances, or were exposed to positive
attitudes towards those substances. Huba et al. (1980) and Bailey &
Hubbard (1990) found similar evidence in the case of friends who used
these substances. Huba et al. (1980) also found evidence that marijuana
use is correlated with friendships where these substances were offered for
use.
Flay et al. (1983) and Fisher et al. (1988) have found evidence
contrary to the idea that role modeling of friends causes experimental
substance use. Friends were found to select each other from their common
interests and activities involving experimental substances, rather than those
friendships causing substance use. This birds of a feather syndrome is a
point of significant controversy among researchers. Kandel (1985) finds that
both friendship selection and friendship influence are factors in substance
abuse. Nevertheless, an overwhelming amount of research shows the
significance of peer and friend influence in the case of youth smoking
39


(Poulsen et al., 2002; Alexander et al., 2001; Wang et al., 1999; Barber et
al., 1999; Rose et al., 1999; Smith and Stutts, 1999; West et al., 1999; Flay
et al., 1998; Michell, 1997; Flay et al., 1993, 1994; Norman and Tedeschi,
1989).
An alternative view is that smoking is a response to immediate need,
often psychological. Peer pressure provides additional forces in favor of the
adoption of tobacco, driven by fear of rejection. Adolescents who are at the
extremes in the popularity spectrum are more likely to be regular smokers
(Spear and Akers, 1988). Adolescent cigarette use can be predicted from
the influence of best friends and friendship groups. These influences include
peer smoking, peers tolerance of smoking, and approval of smoking among
best friends (Kandel, 1980; Rose et al., 1999). Alexander et al. (2001) found
that current smoking status was significantly associated with peer networks
in which at least half of the members smoked, also finding significantly
increased risk where best friends smoked. Barber et al. (1999) describe
peer pressure as an important predictor of smoking for males in all age
groups and for females starting from age 14-15. Furthermore, West et al.
(1999) shows that uptake of smoking between ages 15-23 was associated
significantly with friends smoking behavior, and Flay et al. (1998) found that
the percentages of experimental and regular smokers rises precipitously
with the increasing number of friends who smoke.
Denscombe (2001), in his recent study, discussed peer pressure from
a different perspective. His subjects did not acknowledge the impact of peer
pressure on their smoking initiation because this conflicted with their sense
of personal autonomy, independence and self-determination. They might be
heavily influenced by their friends but they might perceive their smoking as a
free choice. On the other hand, Turkish youth might not have strong desires
to express their personal autonomy since they grow up in a more traditional
40


society in which family and friends are very influential and used as reference
points for defining self in other areas (Guneri et at., 1999). Thus, I prefer to
use the term peer influence instead of peer pressure for my study.
Family Influences
Some studies have found family to be a significant influence on youth
smoking (Distefan et al1998; Farkas et al1999). Jackson and Henriksen
(1997) found a significant risk of uptake when one or both parents smoke.
Children of current smokers were found to be at greater risk of intending to
smoke and perceived easy access to cigarettes. Jackson and Henriksen
also found that children of former smokers were still at risk of modeling
smoking after their parents. Flay et al. (1998) found that parents smoking
and family conflict were primary predictors of youth transitioning from
experimental smoking to regular smoking, a finding confirmed in other
studies (Distefan et al., 1998; Biglan et al., 1995). Similarly, Kandels (1985)
study revealed that parents play a central role in the transition from
marijuana use to other substances. In another closely related finding,
Fearnow et al. (1998) determined that parental actions to prevent their
childrens smoking lessened under environmental stress, while parental
permissiveness was associated with their smoking status. Adolescents are
more likely to smoke when their parents verbal anti-smoking messages are
diluted by their own smoking status (Henriksen and Jakson, 1998; Baer and
Katkin, 1981). Other studies, however, have found that parental and family
influence has a minimal or limited role in adolescent smoking when
compared to peer and friend influences (West et al., 1999; Rowe et al.,
1994; McNeill et al., 1988; Skinner et al., 1985).
West et al. (1999) determined that sibling smoking had a significant
41


effect on smoking uptake for subjects only between ages 15 and 16.
According to Rowe et al. (1994), siblings usually do not have friends in
common, and therefore have different peer influences in regards to smoking;
the differences in friend selection may be attributable to differences in
genetic makeup and differences in the way siblings experience their shared
family and social environments (Rowe and Plomin, 1981).
Teacher Influences
Teachers have also been found to be an influence on youth smoking
(Zhang et al., 2000; McNeill et al., 1988). In a Korean study, 35% of
students indicated that they were affected by their teachers smoking (Lee,
1997). Poulsen et al. (2002) found that Danish students subjected to
teachers smoking outdoors on school premises had a significantly
increased risk of being daily smokers or heavy smokers. Poulsen speculates
that teachers smoking outdoors is related to an increased frequency of
exposure on playgrounds, since student exposure to teachers smoking
indoors was found to be mostly in staff rooms, where the duration of
exposure was presumed to be briefer and less frequent. Poulsen notes that
school policies aimed at limiting second-hand smoke may be partially
misguided, and that total bans on school premises may be necessary to
prevent students role modeling of teachers smoking. Griesbach et al.
(2002), however, found that total bans on teachers smoking actually
increased their outdoor smoking on school premises.
School Policies
The topic of school policies was introduced in the above review of the
literature on teacher influences. Furthering this exposition, Griesbach et al.
42


(2002) found that lower levels of smoking perceived by students was
associated with the consistent enforcement of school policies against
student smoking. This study also found that teachers who smoke were less
likely to intervene to curtail student smoking, a finding echoed by de Moor et
al. (1992). In a classic demonstration of reciprocal determinism, de Moor
also found that teacher intentions to intervene were strongly associated with
the smoking prevalence of male students, but not female students. An
Australian (Clarke et al., 1994) study disputes these results, concluding that
neither smoking policies nor staff smoking had a significant impact on
student smoking. A Welsh study (Moore et al., 2001), however, contradicts
the Clarke et al. conclusion regarding student anti-smoking policy, finding a
strong correlation between the degree of enforcement of student smoking
restrictions and the incidence of student smoking; no correlation was found
between student smoking and enforcement of teacher smoking restrictions.
D. Sociodemoqraphic Factors
Age and Gender
Gender differences in smoking behavior arise early in adolescence,
although in the US many of these differences disappear by the later teenage
years. Across the globe, boys age of initiation into smoking is significantly
earlier than girls (Lucas & Lloyd, 1999). Girls are particularly more
susceptible to peer pressure at the 14-15 age level (Barber et al., 1999), and
Lloyd-Richardson et al. (2002) found that girls are more susceptible to
parental and teachers smoking influences at ages 12-13 than boys. Lloyd-
Richardson, however, finds virtually no difference in risk between US
43


adolescents of both sexes in all categories of smoking. This outcome is very
different from that of Turkey and most of the world, where male smoking
rates are much greater than female smoking rates (Firat, 1996). Some
studies show that gender moderates social influences such as population
migration (Brunswick and Messeri, 1984), level of parent education (Ary et
al., 1988), number of friends who smoke (Brunswick and Messeri, 1984),
mothers smoking status on male childrens smoking status (Skinner et al.,
1985), parents smoking status on female childrens smoking status
(Charlton and Blair, 1989; Chassin et al., 1986), mothers smoking status on
male childrens uptake (Skinner et al., 1985), sibling smoking status on
female childrens uptake (Mittelmark et al., 1987), and friends smoking
status on female adolescents (Charlton and Blair, 1989). Smoking has been
linked to greater self-confidence and social skills in girls (Clayton, 1991;
Mitchell and Amos, 1996), but less self-confidence in boys (Clayton, 1991).
Two explanations are usually given for female smoking, as opposed
to male smoking. For girls, smoking is often associated with concerns about
body-weight and shapeliness (Gritz and Crane, 1991; Camp et al., 1993), a
factor that is particularly pronounced in females with eating disorders (Crisp
et Al., 1999). For women, this pattern remains throughout life. Fisher (1976)
found that womens smoking is linked to a preoccupation with issues of
power, an association frequently exploited by tobacco companies worldwide
(WHO, 1996; Headden, 1998).
Academic Achievement
Although academic achievement is not purely a demographic
variable, I categorized it under this section for organizational purposes. Low
academic performance has also been associated with the use of tobacco
44


and other experimental substances (Choi et al., 1997; Bailey & Hubbard,
1990; Conrad etal., 1992; Jessoret al., 1991; Kaplan et al. 1984; Kandel et
al., 1978). Brynin (1999) claims that there is an association between
adolescent smoking and difficulties at home and school, but also with
greater self-confidence and presumably greater interest in appearance.
Nutbeam et al. (1993) found a strong link between adolescent smoking and
their perception of their school, their academic capabilities, and their future
academic plans. Additional data support that the onset of tobacco use is
related to poor academic achievement: 21% of below-average students
were heavy smokers compared to 7% of above-average students (Elder,
1994; Surgeon Generals Report, 1994). Flay et al. (1998) found an
increased risk of experimental smoking and ever-smoking among students
with poorer grades, but found an increased risk of regular smoking among
the top grade level (A students), versus B students, an unusual finding
not discussed in his study. Lloyd-Richardson et al. (2002), however,
confirms the Flay results, showing that students with grades in the top
quartile are more likely to become regular smokers than their counterparts in
the second quartile.
Urban/Rural Location and Economics
Various studies report contradictory findings on urban-rural
differences, and economic level. Children in rural areas have been found at
greater risk of smoking uptake than their urban counterparts (Harrell et al.,
1998; OConnell et al., 1981), although this is disputed by Stanton et al.
(1994).
Some studies also have found an association between spending
money and tobacco use in adolescents (Rissel et al., 2000; Alexander et al.,
1983), although lower socioeconomic status has also been associated with
45


youth smoking (Harrell et al., 1998; Stanton et al., 1994; Conrad et at.,
1992).
E. Social. Cultural and Psychological Factors in
Turkish Tobacco Use
Although Turkish youth obviously share numerous psychological
characteristics with their Western counterparts, the mediation of those
characteristics by society and culture can produce surprising differences and
surprising similarities in smoking-related behavior. In his recent study
Denscombe (2001) discussed the reluctance of his subjects to acknowledge
the impact of peer pressure on their smoking initiation because this
conflicted with their sense of personal autonomy, independence and self-
determination. Guneri et al. (1999) stressed that Turkish youth might not
have strong desires to express their personal autonomy since they grow up
in a more traditional society in which family and friends are very influential
and used as reference points for defining self in other areas.
Erbaydar et al. (2002) found that students who work are exposed to
smoking workplaces during their working internships, and earn pocket
money that gives them more opportunity than their non-vocational peers to
buy cigarettes. Ozcan and Ozcan (2002) noted that most of the vocational
school student population comes from lower-class families. Some of the
Turkish patterns are replicated in other Middle Eastern countries. Maziak
(2002) concluded that smoking is associated with poor academic
performance and low SES in Syria, whereas Jarallah et al. (1999) found
smoking to be associated with a lack of education in Saudi Arabia.
Many of the factors influencing smoking in the developed world
appear to be universal. Cunningham (1996) argues that the same tools
46


which exploit adolescent needs in the US work effectively across the globe,
operating on common principles of human nature everywhere. These
principles are sometimes called the 5 Ss: sophistication, slimness, social
acceptability, sexual attractiveness, and status. Similarly, Headden (1998)
states that feminine independence is a draw for women in developing
countries. In Western cultures, adolescents frequently report curiosity, social
norms, and social pressure as reasons for smoking uptake (Sarason et al.,
1992). Maziak & Mzayek (2000b) found that peer influence is the greatest
influence on smoking uptake among high school students in Syria, although
parent and sibling influences are also significant factors. In Indonesia, the
smoking status of best friends is the greatest determinant of smoking (Smet
et al., 1999). In China, tobacco use is associated with peers, teachers, and
mothers who smoke (Zhang et al., 2000).
The data that do exist indicate that the development of the smoking
habit is delayed, at least in Syria among the Middle Eastern countries, until
adult supervision is relaxed in the post-secondary school years of young
adults, corresponding to increases in smoking rates after age 18 (Maziak &
Mzayek. 2000). Rissel et al. (2000) had similar findings for Vietnamese,
Asian, and Arabic youth living in Sydney, Australia. Rissel cites closer family
ties, a stronger commitment to education, and lack of pocket money as
reasons for the delay in smoking onset.
The brief review above leaves several factors unexplained in terms of
why and how Turkish youth pick up and maintain the smoking habit. In
Turkey, the adult smoking rates are 63% for men, and 25% for women yet
adolescent smoking rates (ages between 15-19) have been found to be
(only) 30% in Turkey. Taking the average of these adult rates, this indicates
that the adult smoking rate is approximately 44%. This is highly suggestive
that many Turkish youth are becoming regular smokers as young adults.
47


Since individuals in the US rarely take up smoking after age nineteen (Tilly,
1991; Kessler, 1995; Corbett, 2001), clearly some factors are at work in
Turkey that are not evident in the United States.
F. Application of Social Cognitive Theory; Study
Hypotheses
This studys primary concern is with the social environments
influence on smoking status. Sociodemographic variables are also useful
and necessary as control variables when modeling the influences of the
social environment. Social Cognitive Theory emphasizes observational
learning, role modeling, as well as the mutual influences of individual, her
behavior, and society. The literature review provides a theoretical structure
for parental modeling, differential peer association, positive or negative
attitudes toward smoking, perceived positive and negative consequences of
smoking, the integration of smoking into lifestyle, and the interaction of the
social environment with individual characteristics. In sum, Social Cognitive
Theory furnishes a large variety of modeling tools for studies involving social
influences on individual behaviors such as smoking, providing a solid
theoretical foundation for studies such as this.
Three types of information were specified under Social Cognitive
Theory: person-environment; person-behavior; behavior-environment. The
first two interactions are the main focus of the study but the third interaction
will be also included in the discussions where it is needed.
This study incorporates constructs that are directly applicable to the
social environment and socio-demographics, namely the Environment and
Situation constructs. The observational learning construct is categorized
under environment. Each hypothesis and measure in this study is
48


identifiable with these three constructs (or with sociodemographics). As is
typical for studies using Social Cognitive Theory, only one or two constructs
are necessary to study the interactions of interest (Stone, 1993). These
constructs and the potential influences examined are illustrated in Table 2.1.
Table 2.1 Selected Factors Influencing Tobacco Use
Previously Identified Risk Factors Social Cognitive Theory Constructs
SOCIODEMOGRAPHIC
FACTORS o School type o Male gender o Age o Origin (urban/rural) o Allowances (SES) o Economic situation (SES) o Academic Achievement Environment and Person
SOCIAL ENVIRONMENT FACTORS Interpersonal factors o Parent smoking Environment
o Sibling smoking o Relative smoking o Teacher Smoking o Friend smoking Modeling
Perceived Environmental Factors o Close friends approval of smoking (social support) Situation
o Perceived prevalence Situation
among peers (norms)
49


Hypotheses
Various sources suggested potential hypotheses for the study. The
existing literature indicates that family, friend, peer, teacher and school
influences are globally at work on the youth of many cultures. Likewise, the
influences of socio-demographic factors are documented across the globe:
gender, age, place of origin, economic circumstances, and academic
success. These, together with the preliminary focus groups, interviews and
surveys, as well as the researchers own experience as a citizen of Turkey,
informed the creation and elimination of the potential hypotheses. Each
hypothesis furnishes an opportunity to ask a series of questions in the
survey that will reveal the mechanisms of smoking influences in greater
detail. The hypotheses of the study are presented below, organized by
major theoretical construct and topical area of influence.
Sociodemoaraphic Factors
School Type. In the preliminary interviews, many perceived the
industrial high school as the smoking school. In Turkey, high school
students are targeted by levels of academic achievement into different types
of high schools: regular (academic), specialized vocational, and religious
schools. Ozcan and Ozcan (2002) found a greater risk of smoking for
students of some types of vocational schools in Ankara. In addition, testing
a hypothesis related to type of school introduces a potentially significant
control variable. This leads to the hypothesis:
(H1): Adolescents smoking differs by school type.
Gender. The literature strongly supports gender as a significant factor
in smoking uptake and maintenance. The preliminary survey indicated that
50


male adolescents in Dursunbey are at much greater risk of smoking than
girls. In addition, gender is an important factor that will serve as a control
when analyzing other factors related to smoking. A new hypothesis is
added:
(H2): Adolescent boys will be more likely smokers than girls.
Age. The literature supports age as a significant factor in smoking
status. The preliminary survey supported this association. Like gender, age
is an important factor that will serve as a control when analyzing other
factors related to smoking. The hypothesis is:
(H3): The older the adolescent, the higher the probability of being a
smoker.
Place of Origin, Dursunbey is a mixed urban/rural community, making
it an excellent area for studying potential differences between urban and
rural smoking patterns. Indeed, this opportunity was one of the motivations
in selecting of Dursunbey for this study. Like gender and age, place of
origin also serves as an important control when analyzing other factors
related to smoking. The hypothesis:
(H4): Adolescent smoking differs by place of origin (urban/rural).
Weekly Allowance. The literature reviewed indicates that there is a
significant relationship between smoking status and spending money. The
hypothesis is included:
(H5): Adolescent weekly allowances will be correlated with smoking
status.
Economic Situation. Turkey has been undergoing severe economic
51


stress over the last several years. With an unemployment rate of 10%, a
40% annual inflation rate, and currency devaluations, families in Turkey are
under considerable economic stress. In Dursunbey itself, the slumping
forest and coal industries are adding to the problems in the local economy.
The preliminary survey indicated that neither family income nor weekly
allowance was related to adolescent smoking. However, the general
economic conditions in Dursunbey prompts one to check on youths
perception of their economic situation in relation to smoking. A hypothesis
relating to adolescents perception of their economic situation is added as
insurance against missing a potentially important factor in youth smoking:
(H6): Adolescents perception of their economic status will be
associated with their smoking status.
Academic Success. There is a substantial body of evidence in the
literature linking an increased risk of smoking uptake and experimental
smoking to low academic achievement. To explore this possibility, a new
hypothesis related to academic achievement is added:
(H7): Adolescents perception of their academic achievement will be
associated with their smoking status.
Social Environment Factors
Family Influence. Families are also an important part of ones social
environment from which one learns through observation. The literature is
mixed in its findings about the influence of family on ones smoking status,
much of it finding that family is a lesser influence than friends and peers.
Nevertheless, some studies suggest adolescents model the smoking
behavior of their parents, and that children of smoking parents are more
likely to shift from experimental smoking to regular smoking. This suggests
52


that this topic is worth investigating in Dursunbey, where the relative
strength of parental influence may differ somewhat from studies performed
in the West.
The literature on siblings is also mixed, indicating a vulnerability to
the influence of siblings between ages 15 and 16. This is, however, within
the age range in my studys target population, which is a prime age for
smoking uptake. Interviews from the preliminary studies indicate that
relatives exert some role modeling influence. For these reasons, three
hypotheses related to family are tested in this study:
(H8): Adolescents whose parents smoke are more likely to smoke
themselves.
(H9): Adolescents whose siblings smoke are more likely to smoke
themselves.
(H10): Adolescents whose relatives smoke are more likely to smoke
themselves.
Teacher Influence. The reviewed literature from countries as
dissimilar as Denmark and China show that teachers are a significant
influence on adolescent smoking status. The role modeling influence of
teachers was strongly suggested in the preliminary interviews and focus
groups. Teachers have traditionally been revered in Turkey, although the
interviews indicated that this special regard might be waning. Earlier Turkish
studies have shown that 44% of teachers are smokers, potentially leading to
the frequent exposure of their students to their smoking. It is hypothesized
that:
(H11): Adolescents who are exposed to teacher smoking are more
likely to smoke.
Friend Influence. The reviewed literature documents the influence of
smoking friends on ones smoking status, not only in Western cultures, but
53


also in countries as diverse as Syria, Indonesia, and China. Friends usually
play an important part in social environment of individuals, and from the
researchers personal experience this may be particularly true of Turkey. It is
therefore reasonable to suspect under SCT that observational learning and
role modeling takes place among friends, and with this an association
between friends smoking status and ones own smoking status. The
preliminary focus groups and interviews also indicate that smoking in Turkey
often receives approval from adolescent friends. This leads to two
hypotheses:
(H12): Adolescents whose close friends smoke are more likely to
smoke themselves.
(H13): Adolescents who receive positive reactions about their
smoking from their close friends are more likely to smoke.
Perceived Smoking Prevalence Among Peers. The literature strongly
supports the link between peer smoking and the risk of smoking. To the
adolescent, the perception of smoking prevalence among peers involves a
judgment on the sum of his experiences, which may involve both objective
and subjective factors. Subjective perceptions of prevalence, however, could
well color a youths judgment about whether or not smoking is a normal and
acceptable activity. A new hypothesis is added:
(H14): Adolescents who perceive higher smoking prevalence among
peers are more likely to smoke themselves.
54


CHAPTER 3
METHODS
A. Study Design
A multi-method research design was employed for this study,
although the primary method was a survey. Focus groups, interviews, and
observations allowed insight into the context of tobacco use by youth in
Turkey, and contributed to the development of the survey. These qualitative
methods aided in forming the final set of hypotheses and creating
understandable, culturally appropriate, and comprehensive items for use in
the cross-sectional survey of all 10th graders in the five high schools of
Dursunbey. The initial work was useful in planning data collection
procedures, eliminating potential problems before the actual survey
instrument was administered.
Data gathering took place over two seasons of fieldwork in Turkey
(see Figure 3.1). Preliminary work in the field season one took place during
June and July of 2001. After passive informed consent was obtained, a
preliminary survey of 247 students in Dursunbey and Izmir aided the
development of the field season twos final survey, supplemented by focus
groups and semi-structured interviews of 32 students in Izmir. Lessons from
those research exercises (see field season one study below) benefited the
construction of the survey instrument used in the second field season.
55


Figure 3.1. Timelines for the Data Collection and Write-Up
June-Juiy 2001 Field season one study completed in Dursunbey and Izmir,
August 2001- March, 2002 Turkey (247 surveys, 32 in-depth semi-structured interviews and 3 focus group) Data entry and preliminary analysis of field season one study. Designed the final questionnaire, translated questionnaire into Turkish. Finalized the dissertation proposal.
April 2-11,2002 Obtained permission from Education Ministry; shared the
questionnaire with school counselors, modified the questionnaire
after counselors feedback.
Talked with relatives and friends about adolescents and adult
smoking.
April 8-12, 2002 Pilot tested the survey; scheduled survey administration for each
school; distributed passive parental consent forms. Interviewed
April 15-26, 2002 school principals. Administered the survey to 10th grade students in 5 schools in Dursunbey (N=211). Recruited students for semi-structured interviews; began encoding the survey results into SPSS statistical software; reviewed survey data; refined semi- structured interview guide.
April 12-May 25, 2002 In-depth semi-structured interviews with students and interviews with school counselors, teachers, and principals (N=70). In addition, three focus groups with students.
June 2002- September 2003 Transcribed interviews; analyzed interviews using Microsoft Word; translated the supporting quotes. Completed inputting survey results; analyzed the survey data; wrote up survey results.
The second field seasons data collection in April and May 2002. The
purpose of the final survey was to collect baseline data regarding the 10th
graders' smoking behavior. Like the first season, the final survey was
accompanied by focus groups and semi-structured interviews. This
investigation allowed me to identify social factors and smoking patterns
56


among youth. The 57-item questionnaire was designed to obtain data on
demographic variables, exposure to smokers in their family environment, in
peer and friend groups, and in school environment. Demographic variables
of school type, age, gender, family income, perceived economic situation,
weekly allowances, place of origin, family, and friend influences constitute
the independent variables. The dependent variable is the individual's
smoking status categorized according to respondents self-identification
using WHO's question of smoking in the past thirty days.
All interviews, focus groups and surveys were conducted and
analyzed in Turkish. Some survey questions that did not translate well were
discussed with Turkish speaking people from the local community and
school counsellors to develop appropriate vocabulary.
Passive parental consent was again obtained, followed by the
administration of the questionnaire in all five of Dursunbeys high schools,
consisting of industrial vocational, girls vocational, religious, and regular
schools. This cross-sectional survey included all 10th graders in the town of
Dursunbey (n=211). The study inclusion criteria were that students must be
enrolled in a high school, be enrolled in 10th grade, attend school on the
day(s) of survey administration, and consented to the study. Of the 57
survey questions, sixteen questions were developed based on input from
previous focus groups. The focus group study also aided in the selection of
forty-six questions drawn from various previously validated surveys.
I conducted the qualitative portion of study using semi-structured in-
depth interviews with a grant received from Research for International
Tobacco Control, International Development Research Center in Ottawa,
Canada. A convenience sample of the survey population (n=42) was
selected, together with students from other grades (n=10). These interviews
helped me to interpret the survey results and are incorporated into the
57


Dursunbey the following year. The qualitative part of the study was used to
fill the gaps identified in the quantitative part.
The initial survey was designed to be exploratory rather than
hypothesis-testing. Its 62 questions asked about individual attributes such
as age, gender, family income, monetary allowances, knowledge, attitudes,
behavior, stress, decision-making and family, friend influences, social
networks and norms, as independent variables. An individuals smoking
status was categorized based on the respondents self-identification,
according to the World Health Organization's classifications of daily smoker,
occasional smoker, ex-smokers and nonsmokers (WHO, 1997).
The qualitative components of the Field Season One allowed the
researcher to explore meanings, definitions, and other cultural aspects of
tobacco use by Turkish youth in diverse settings. The semi-structured
interviews allowed the teenage subjects to describe their smoking behavior
in their own words, giving the researcher another view of social contexts in
which smoking behaviors occurs. The combination of survey and semi-
structured interviews provided two perspectives for understanding youth
knowledge, attitudes, and behaviors about smoking uptake and
maintenance of smoking.
Construction of the Survey for the Field Season One
The survey questionnaire consisted of multiple choice and open-
ended questions. The survey questionnaire (See Appendix A) included
fifteen open-ended questions to allow teenagers to express their ideas in
their own categorizations, rather than the researchers categorization. Forty-
six questions were drawn from the WHO Global Youth Tobacco Survey, the
COMMIT Youth Survey (1993), the AMC Arresting Smoking Uptake Using
59


Interactive Multimedia (2000) survey, and the Marmara University (MARUN)
survey (2001). Table 3.1 shows the structure of the questionnaire that was
used during the field season one.
Table 3.1. Structure of the Questionnaire
Area of measurement Sample of Questions j
Demographics -Age, gender, household income, parents education
Knowledge, attitudes and Beliefs -Perceived health dangers of smoking -Perceived list of diseases linked to smoking -Beliefs and expectations about future prospects and health -Pro-tobacco messages -Anti-tobacco messages
Social and symbolic influences -Brand preferences -Parents and teachers who smoke -Perceived prevalence of smoking -Perceived parental approval of smoking -How many of their friends smoke -Offers of cigarettes from friends
Locus of control -Relationships with family -Do you think you can give up smoking? -Do you think not smoking is a way to express your independence?
Self-efficacy -Self-reported refusals of offers of cigarettes -Performance in school
Motivations/reasons for smoking or not smoking -Who encouraged you to try your first cigarette? -Who were you with when you tried your first cigarette and where were you? -Why do you smoke now?
Intentions to smoke -Do you think you will smoke a cigarette anytime during the next year?
Access -How do they obtain their cigarettes -Have they ever purchased single cigarettes
Cessation -Do you want to completely stop smoking? -How many times have you tired to quit smoking? [
In addition, during the survey questionnaire development, students
were given small pieces of paper and asked to list the reasons why they
started to smoke or why their peers started to smoke in four high schools in
60


Dursunbey from grade 9 to 11.
Focus Groups
Focus groups were conducted to orient the researcher to current
Turkish adolescent attitudes and beliefs about smoking, as well as their
smoking/nonsmoking behaviors.
During the summer of 2001, a focus group was carried out in
Dursunbey, consisting of 7 girls and 2 boys from low to middle income
levels, with ages between 15 and 17 (10th and 11th graders). One individual
was a nonsmoker and the rest were smokers. In addition, two focus groups
were conducted in the urban city of Izmir. These focus groups enabled the
researcher to explore smoking issues from the perspective of the teenagers
own meaning and experiences. Among the questions asked were: When
and how did they start smoking? Why do they smoke now? In which
environment do they smoke the most? What are their reasons for smoking
and what role do their peers play in the decision to smoke? What are their
reasons not to smoke? The focus group study also aided in the selection of
46 questions drawn from the WHO Global Youth Tobacco Survey.
Semi-Structured Interviews
During the Field Season One 32 students were interviewed in Izmir,
Turkey. Students who were interviewed were asked to tell their stories: e.g.,
how did they start smoking, and when and where did they try their first
cigarette? Why do they smoke now? Or for nonsmokers, why do they think
young people smoke? What does smoking mean to them in their social
environment? In addition, questions were asked to identify the peer pressure
and family influences in smoking (see Appendix A)
61


Field Season One Data Collection and Sampling
Convenient sampling procedures were used in Dursunbey and Izmir.
After an initial focus group in Dursunbey provided insights into adolescent
attitudes about smoking, a questionnaire was given to 247 high school
students of mixed grade levels in Izmir and Dursunbey. This was followed
by 32 semi-structured interviews given to subsample of the survey subjects
in Izmir. The study was completed with two more focus groups in Izmir, the
third largest city in Turkey.
A total of 174 students from all five schools were surveyed in
Dursunbey: the religious high school (n=37), the health vocational school
(n=21), the regular high school (n=43), and the industrial vocational high
school (n=70). In four high schools, all the 10th graders who were available
during the administration of the survey were surveyed. In addition, one
class of 9th graders from the industrial vocational high school class, one
class of 9th graders from the regular high school, and one class of 9th and
11th graders from Religious high school were surveyed.
For the survey administered in Izmir, students (n=43) from two 10th
grade classes of the Izmir Karsiyaka Regular High School, and 32 students
from various EVKA-4 high schools were surveyed. The administration
procedures were identical to those followed in Dursunbey.
All the semi-structured interviews were conducted in Izmir. The set of
subjects for the interviews was a subset of the subjects who completed the
Field Season One survey, where the parental consent forms covered the
contingencies for both methods of data collection. Students participating in
the survey were chosen by convenience sampling for participation in the
interviews. Thirty-two teenagers from various grades (9th, 10th, 11th) were
interviewed, with the majority of them 10th graders. These teenagers were
62


from the EVKA-4 Regular High School (n=16), the Industrial Vocational High
School (n=6), the Karsiyaka Regular High School (n=2), and high school
dropouts (n=8). Smokers were represented by higher numbers than
nonsmokers in the semi-structured interviews. Female and male segments
of the sample were at rough parity (14 girls and 18 boys).
Analysis
The open-ended survey questions were coded first, then the
questionnaire results were entered into a SPSS database. Chi-square tests
were used to look for associations between smoking status and predictor
variables.
Semi-structured interviews were analyzed by using content analysis.
Descriptive data regarding key factors and conditions influencing smoking
attitudes and behaviors of Turkish youth were developed through in-depth,
systematic examination and interpretation of the interview data and
identification of key patterns.
Field Season One Study Results
Study results showed high smoking prevalence in the study
population. Seventy-four percent (N=246) of the subjects ever tried smoking,
with 60% of girls and 82% of boys. Twenty-one percent of the students
identified themselves as occasional smokers, and 24% as regular smokers.
Boys were 3.1 times (95% Cl, .18, .58; p< .000) more likely than girls to
have ever tried smoking, and boys were 3.14 (95% Cl, .19, .54; p< .000)
times more likely to be a current smoker than girls. Boys were also 2.1 times
(95% Cl, 1.05, 4.1) more likely than girls to be regular smokers and 1.2
times more likely to be occasional smokers (95% Cl, .44, 3.4; p<.000).
63


Family influences were mentioned during the Field Season One
interviews, but in the survey, father and sibling smoking were not significant,
although the mothers smoking was. The number of close friends who
smoke was significantly associated with adolescent smoking status, both in
the field season one survey results and interviews.
Grades were significantly correlated with smoking behavior.
Household income was not significantly correlated with smoking behavior,
and neither was weekly allowance.
From the interviews, it was clear that the subjects thought that
smoking was relatively benign, and less serious than drinking, gambling, or
drugs. Most interviewees knew about the harmful effects of smoking, but
were undeterred by this knowledge. Smoking was a circumstantial and
unplanned activity. They observed that others find that smoking is
enjoyable, it helps to forget worries, or is even an activity associated with
celebration. Often the decision to try smoking is spontaneous, the reaction
to an upsetting event where friends offer them a cigarette, saying, Take
one. It will help you relax.
Often in focus groups and interviews, students described practicing
smoking with their friends. They practiced inhaling in groups, some believing
that if you dont inhale, you are a fake and are not a real smoker. Some
believe that the smoke must be inhaled to be harmful; others believe that it
is harmful if not inhaled, since this might result in mouth cancer. These are
some of the myths about tobacco use found in the qualitative portion of my
field season one study (Appendix B, includes more detailed information on
preliminary results from quantitative and qualitative fieldwork, Summer
2001).
Although the field season one study offered a good background for
understanding youth smoking behavior in Turkey, It did not include a
64


random sample of the students within these schools and I wanted to get
more detailed information about the interpersonal influences on youth
smoking. As a result I conducted another survey during the second field
season (the survey that is the subject of this dissertation).
C. Field Season Two in Dursunbev. 2002
Field season two has two components survey and semi-structured
interviews. Both the survey (N=211) and semi-structured interviews (N=70)
were conducted in Dursunbey, Turkey.
The Survey in Dursunbev. 2002
Study Sample. This is a cross-sectional survey of all 10th graders in
five high school of Dursunbey, Turkey (N=211). Criteria for inclusion in the
study included: being enrolled in the 10th grade in a Dursunbey high school,
school attendance on the day(s) of survey administration, and informed
consent. The researcher returned to two high schools where 4 students
were absent on the initial day of administration. The mean age for the 10th
grade sample population was 16 years, and is significant because prior
studies indicate that 39% of Turkish smokers start uptake in an age range of
15 to 17 years, and 20% of them between 11 to 14 years (Asut, 1993). The
detailed information about the study area and population can be found in
Chapter three.
Data Collection Procedures. IRB approval from the University of
Colorado at Denver was obtained in June 2001. The same approval was
extended for another year in 2002. In addition, permission to conduct the
65


survey was granted by the local bureau representative of the Ministry of
Education in Dursunbey. The principal of each school was then consulted
about the schedule for administering the survey.
In April 2002, permission for a high school tobacco study was
obtained from the Dursunbey office of the Ministry of Education. Informed
consent forms were distributed to the classes in the weeks prior to the
administration of the survey instrument in each school. Parent(s) were to
sign if they refused their childrens participation. Since Turkish parents place
a high degree of trust in the authority and integrity of school administrations,
parental permission was obtained without exception. After passive parental
consent was obtained, the questionnaire was given in all five of Dursunbeys
high schools, consisting of industrial vocational, girls vocational, health
vocational, religious, and regular schools, in April and May of 2002.
The first two weeks of the data collection involved meeting school
counselors and principals, refining the survey questionnaire. One of the
school counselors was consulted to finalize the questionnaire. The survey
items were discussed with one of the Dursunbey school counselors in terms
of clarity and relevance of the questions, in order to create a student friendly
questionnaire. This resulted in several recommended changes. After the
counselors suggestions were incorporated into the questionnaire, a pilot
test of the questionnaire was conducted, first administered in a face-to-face
interview to two students (one girl and one boy). The researcher solicited
feedback from the students about their understanding of the survey
questions and their interpretation of the concepts and phrasing used, as well
as the criteria the students used in selecting their responses. This test
uncovered a number of difficulties in the questionnaire. The students had
problems, for example, when some of the skip patterns were not clear to
them, or when the questions were formatted using the Likert-scale in table
66


form. Using this feedback, the researcher reduced the size of the tables,
dropped some of the question tables from the questionnaire, and eliminated
some of the survey questions. As a next step developing the Field Season
Two survey, the refined survey was administered to one class of twenty-five
9th graders. During this phase, it became clear that the students were not
carefully reading the instructions, which were now seen as extensive in
length. This meshed well with an observation by one of the school
counselors, that many of these students are not accustomed to taking
surveys. Several refinements in the instructions and administration
procedures were made; the researcher shortened instructions and prepared
a fifteen minute detailed verbal explanation to preface administration of the
survey. The researcher planned to be present while students were taking
the survey, so that the students could seek clarification about the survey
questions. This exercise also helped to estimate the actual administration
time, which in practice consumed the entirety of a 45-minute class period.
The questionnaire was administered by the researcher during class hours
designated by the teacher and school administration.
Parental Informed Consent and Student Assent Form. For the survey,
students in selected schools were given a consent form for the parent(s) to
sign if they chose to refuse their students participation. In the student assent
form, students were told that the participation was strictly voluntary and they
could stop any time they wanted. Students were told that the information
they were to give would be used to develop better education programs for
young people like themselves. Students were assured that the questions
that asked about their background would only be used to describe the types
of students completing this survey. The information would not be used to
find out their name, and no names were ever to be reported.
67


Field Season Two Study Instrument. The questionnaire was designed
and modified with the knowledge gained from field season one study and
drawing from various validated questionnaires from prior studies. The final
survey questionnaire had 57-items (See Appendix A), including 54 multiple-
choice questions and 3 open-ended questions. Eight questions on the
questionnaire were created with knowledge gained from the prior focus
group study.
Knowledge gained in the field season one was invaluable in
improving the survey instrument and strengthening the study as a whole.
The number of answers to some multiple-choice questions was reduced to
cover potential low-frequency responses. Most importantly, it became clear
that the Field Season One survey was too long, ranging in topics from
demographics, to locus of control, social influences, health awareness, and
perceptions of tobacco companies. Field Season One study helped to cut
down the length of the survey to focus on socio-demographics and social
influences, so that virtually all students could complete it in one class period.
During the focus group, the researcher realized that teenagers
definitions of a smoker differed from the definition in tobacco control
literature. The latter defines a smoker as someone who smokes one or
more times within thirty days prior to the survey. Teenagers in the focus
group did not consider themselves smokers if they only smoke occasionally,
and if they do not carry a pack or inhale. Inhaling cigarette smoke was an
important issue for teenagers because most of them indicated that they
initially practiced inhaling with the assistance of their smoker friends. In
addition, the researcher realized that brand issues surfaced frequently
during the field season one. This led to the creation of eight survey
questions about the reasons for brand preferences. At the end of the
68


preliminary study, three of those questions were dropped to streamline the
survey.
Various other questions were deleted from the survey. Questions
about locus of control were removed as the study focused on social
influences. Similarly, questions were dropped concerning intra-personal
variables such as beliefs, rebelliousness and risk-taking behavior. It became
clear in interviews that virtually no negative publicity about tobacco
companies was reaching these youth; I removed the question, do you think
cigarette companies target teens to replace smokers who die?
Some questions were added to the survey. In the field season one
survey, many students were unsure about their family income, but in the
interviews they certainly knew their weekly allowance. After interviewing a
school counselor, the income levels in the family income question were
adjusted and a question about weekly allowance was kept. The interviews
revealed that adolescents were buying single cigarettes, rather than just full
packs. As a result, a question was added about single-cigarette sales.
During the data analysis, I will mainly focus on the variables related to social
environment, interpersonal influences, and sociodemographic influences.
Translation of the Questionnaire. The construction of the final survey
began in the fall of 2001. The questionnaire blended 49 of questions
adopted from prior studies conducted in Colorado, Turkey, and worldwide
with 8 new questions from the focus group and interview results. The
researcher then translated the questionnaire into Turkish; the translation
was reviewed for accuracy and clarity by Candan Duran-Aydintug, Ph.D., a
University of Colorado at Denver sociologist of Turkish national origin.
69


Measures
In this section study measures dependent and independent variables
will be described in details.
Dependent Variable. The World Health Organization (WHO, 1998:76)
provides precise definitions for smoking status, categorized as daily smoker,
occasional smoker, and non-smoker. Each category is discussed below.
A Smoker is someone who, at the time of the survey, smokes any
tobacco products either daily or occasionally, as defined under those two
composing categories. In this study there was a total of 88 students (42%
out of 211) who met this definition. Smokers may be further divided into two
categories of daily smoker and occasional smoker. A daily smoker is
someone who smokes any tobacco product at least once a day. In this
study 54 (61%) students out of 88 were daily smokers. An occasional
smoker is someone who smokes, but not everyday. Thirty- four (39%)
people out of 88 smokers were occasional smokers.
Occasional smokers can be further divided into two groups:
continuing occasional and experimenter. Continuing occasional smokers
are people who have never smoked daily, but who have smoked 100 or
more cigarettes (or the equivalent amount of tobacco) and now smoke
occasionally. In this study a total of 13 students met this definition.
Experimenters are people who have smoked less than 100 cigarettes (or the
equivalent amount of tobacco) and now smokes occasionally. In this study
21 students met this definition (Table 3.2 presents study results based on
these definitions).
70


Table 3.2. Definitions of Adolescent Smoking (N=211)
Smoking Status ever smoked, even a puff? smoked in last 30 days ffimllfllm has smoked, not a whole cigarette iMi has smoked 1-99 cigarettes in lifetime has smoked 100 or more cigarettes smoked on 20 or more days in last 30 days
never Smoked No 35 (17%) flRHHNfli
lifetime use yes 176 (83%) maybe maybe maybe maybe maybe
current yes yes 88 (42%) maybe maybe maybe maybe
S puffs yes maybe yes 37 (18%) no no no
O ' ^ level experiments yes maybe no yes 21 (10%) no maybe
use established yes maybe no no yes 74 (35%) maybe
frequent yes yes no no yes yes 54 (26%)
A nonsmoker is someone who, at time of the survey, does not smoke
at all. In this study123 (58.3%) students indicated that they did not smoke in
the past 30 days. Non-smokers may be further divided into two categories:
Never-smokers and ex-occasional smokers. Never smokers are those who
either have never smoked at ail (n=35) or have never been daily smokers
and smoked less than 100 cigarettes (or the equivalent amount of tobacco)
in their lifetime (n=78).
Ex-occasional smokers are those who were formerly occasional, but
never daily, smokers and who smoked 100 or more cigarettes (or the
equivalent amount) in their lifetime (n=10).
Smoking status as an outcome variable was measured by the
question (Q17): During the past 30 days, on how many days did you smoke
cigarettes? This question was answered by seven categories: (1=0 days;
2=1 to 2 days; 3=3 to 5 days; 4=6 to 9 days; 5=10 to 19 days; 6=20 to 29
days; 7=AII 30 days). To form a dichotomous dependent variable for logistic
71


regression analysis, subjects answering anything but 0 days were
classified as smokers.
Independent Variables. Table 3.3 and Table 3.4 present the
independent (explanatory) variables and their descriptions.
Table 3.3. Explanatory (Independent) Sociodemographic Variables
Variable Variable Description Variable Coding
Demographic Factors
School type The type of school students enrolled 1=religious high school 2=health vocational school 3=regular high school 4=girls vocational school 5=industrial vocational school
Gender What is your gender? 1=male 0=female
Age What is your age? 15 to19
Family origin Where did you originally 0=urban
come from? 1=rural
Perceived economic How do you perceive 1 =very poor
situation your economic status? 2=poor 2=average 3=good 4=very good
Weekly Allowances (in How much money do 1=None
USD) you receive from your parents each week that you can spend anyway you want? 2=< 0.38 3=0.38-0.77 4=0.78-2.30 5=2.31-3.83 6=>3.83
Perceived academic How do you perceive 1 =very good
achievement your academic 2=average
success? 3=poor
The review of literature helped to determine the choice of
independent variables. Banduras Social Cognitive Theory provided the
72


theoretical foundation. As a result the independent variables were
categorized using this framework. Under the demographic factors domain
the following questions were asked: school type, gender, age, family origin
(urban/rural), levels of weekly allowance, perceived economic situation, and
academic performance.
Table 3.4 presents the social environmental factors and perception of
the social environment. Social environmental variables measured whether
the adolescents parent, siblings, and other household members smoke
(0=no, 1=yes). Teacher influence was measured exposure to teachers
smoking on school premises (0=rarely/never, 1=everyday). Friend
influences were measured by two questions: close friends smoking status
(0=none, 1=1 friend, 2=2 friends, 3=3 friends, 4=4 friends, 5=not sure), and
close friends approval of smoking (1=approve, 2=not sure, 3= dont
approve). Perceived smoking prevalence among peers is an ordinal variable
but for the purpose of the analysis is treated as continuous (1 =less than
10%, 2=10-25%, 26-50%, 51-75%, 76-90%, 6=more than 90%).
73


Table 3.4. Explanatory (Independent) Social Environmental factors
Variables
I Variable Variable Description Variable Coding I
Social Environment | Factors
Parent smoking Does your mother or father smoke? 0=neither smokes 1 =one smokes 2=both smoke
Sibling smoking Does any of your siblings smoke? 0= no 1= yes
Relative smoking Does any one else in your household smoke? 0= no 1= yes
Teacher smoking How often do you see teachers smoking? 0=rarely/never 1 =everyday
Friend smoking Out of four close friends how many of them are smokers? 0=none 1=1 friend 2=2 friends 3=3 friends 4=4 friends 5=not sure
Perceived Environmental Factors
Close friends approval of smoking What would your close friends thing about your smoking? 1 =approve 2=not sure 3=dont approve
Perceived prevalence of smoking among peers How many percentage of your peers do you think smoke at least once a week? 1=less than 10% 2=10 to 25% 3=26-50% 4=51-75% 5=76-90% 6=more than 90%
Data Entry
After the data collection, 211 questionnaires were entered into the
SPSS statistical software by the researcher. The data entry sheet was
created with all 57 questions and their sub-items. Open-ended questions
were categorized where it was needed. The data entry and cleaning process
took two months.
74


Analysis
Analytic techniques included descriptive statistics and logistic
regression analysis using SAS 8.02 and SPSS 11.0, under guidelines
described in Tabachnick & Fidell (2001). Each hypothesis was tested using
logistic regression, followed by separate tests for socio-demographic and
social environment factors. Logistic regression was used again to produce a
final model based on these constructs. Models were built and tested using
classification of cases, coefficient of determination via the Nagelkerke
adjusted R2 statistic, and chi-square for goodness-of-fit tests. Models using
constructs were tested for goodness-of-fit using the Hosmer & Lemeshow
statistic.
Semi-Structured Interviews in Dursunbev, 2002
As originally conceived, this study combined a cross-sectional survey
with semi-structured interviews, with the interview subjects being a
convenience sample of the survey subjects. Sharing the purpose of the field
season one interviews, the qualitative portion of the study was designed to
further explore the cultural influences on youth smoking. Semi-structured
interviews in the field season two asked many of the same questions as the
field season one study interviews. For smokers, it asked how the subjects
started smoking, when and where they tried their first cigarette, what
influenced them to start smoking, and why do they smoke now? A typical
subjects day related to smoking? What do they see as benefits and
disadvantages of smoking? What people around them smoke? How does
that influence them? For nonsmokers, why do they think their peers smoke?
75


For both smokers and nonsmokers, what does smoking mean in their social
environment (See Appendix A).
Data Collection and Sampling. During the conduct of the survey, a
subset of the survey subjects in each school was interviewed. The 42
student interviews averaged 45 to 60 minutes in length. Some of the
students were interviewed more than once, since they were able to
articulate their smoking very well. Some of the principals, school counsellors
and teachers were interviewed more than once. There were total of 15
secondary interviews. Secondary interviews allowed researcher to discuss
the emerging themes in greater detail.
Students were assured that everything they said was confidential
and would not be revealed to school administrators and parents. While the
researcher followed an interview script, she also pursued relevant topics
raised by the interviewees. This opened the interview process to explore
new cultural and social factors from the perspective of the informants.
Additional interviews were conducted with 3 teachers and 2 school
counsellors, 5 principals and 4 vice-principals. Due to the emerging themes,
one middle school principal and vice principal were informally interviewed,
along with 4 adults and 10 students from other grades. Thirty interviews
were taped-recorded.
Analysis. As the transcription and analysis of the qualitative data
proceeds, it is clear that its value lies in its description of the links between
adolescent smoking and cultural/institutional factors such as Turkish
education policy, long-standing tradition, and economic hardship. All records
of the interviews have been examined. The repetition of themes occurring in
the interviews suggests that they can be confidently used to interpret the
76


results in the quantitative portion of the study. Further analysis for sub codes
will take place in the future.
All the notes from these interviews were read for the emerging
themes. All of the taped interviews were listened and transcribed. Seventy
interviews have been formally transcribed and coded under the Open
Coding method. The Social Cognitive Theory informed the coding process
and helped in organizing the themes and categories. Codes were
constructed based on a preliminary scanning of the text as recommended by
Crabtree & Miller (1999), the relevant text in the transcripts were highlighted.
Emerging themes and categories under these themes were created. The
coding consists of tables associating direct quotes with themes identified by
the participants, such as motivation to lose weight, hospitality ritual",
mixed messages from teachers and family members, cigarette brands as
status symbol, relaxation, boredom, or passing the time. The codes
and emerging themes were discussed with Candan Duran-Aydintug, Ph.D.,
a Turkish sociology professor at the University of Colorado. In the process of
analyzing the transcribed data, the main themes, categories and
subcategories under these themes have been identified. The adult
interviews particularly shed light on the role of tobacco in the lives of the
youth. These adult interviews raised concerns about the structural level of
societal change and social issues on tobacco use. All of these data was
triangulated to have the full understanding of adolescent smoking behaviour
from different point of views. The quotes related to major themes were
translated into English.
77


Methodological Limitations
For the most part, the findings of this study can be generalized to the
population of Dursunbey. The study gives useful insights in understanding
rural towns with conservative populations. It should be noted that smoking
among females is much more common in larger cities. The social norms and
high acceptability of smoking in Dursunbey is common in Turkey.
Although the smoker is defined as smoking even one cigarette within
thirty days of the survey, teenagers often did not see themselves as
smokers. Various reasons were given: occasionally smoking did not
constitute smoking in their view, or if they did no buy the packs themselves.
They also do not see themselves as smokers if they do not inhale. Many
say that the practice of inhaling qualifies as real smoker, if you dont inhale
you are faking it. Some of the teenagers said that if you dont inhale you
will get throat cancer. The researcher relied on the youth's self-report in
determining smoking status. It is possible that some of the students did not
report their smoking or some of them underreported the frequency of
smoking. Also many other measures relied on adolescents self-report rather
than direct measure.
78


CHAPTER 4
DURSUNBEY, ITS HIGH SCHOOLS, AND CHARACTERISTICS
OF THE STUDENTS
A. The Town of Dursunbev
This chapter describes the setting of the study at two levels: the small
city of Dursunbey, and the five high schools in the city. In addition, the life
style and daily activities of youth in Dursunbey will be described.
Ethnographic observations will be utilized in describing the daily lives of
these students.
The study area for this research, Dursunbey, Turkey, is a small city or
district town in the Aegean region that is one of the most populated regions
of Turkey. The town of Dursunbey is on the border of the Aegean and
Marmara regions of Turkey and is situated within the province of Balikesir.
Located about 200 kilometers south of Istanbul, it is one of nineteen towns
of its size or larger in Balikesir. Figures 4. 1 show the maps of Turkey and
the province of Balikesir with the town of Dursunbey respectively. In this
chapter, I first introduce the study area, Dursunbey, Turkey. Then, I provide
an explanation of the data population. Other health scientists have
investigated tobacco use by youth in cities and larger towns; no other
researcher is focusing on the communities of smaller size that represent
important segments of the Turkish population.
79


The majority of Dursunbeys land is covered with forest and the town
itself has an altitude of 639 meters1 above sea level. Its climate is one of
extremes, with cold winters and hot summers. Dursunbey covers an area of
1,952 square kilometers. It has a total of 643 kilometers of roads; twenty-
two percent is asphalt and seventy-eight percent is gravel (Kaymakam
Office, 2002). While seventy nine percent of the villages has sufficient
drinking water, twenty one percent of them lacks sufficient drinking water
supplies. An adequate sewage system exists in twenty villages, however,
there are inadequate sewage systems in six villages. As of 2002, the
construction of new sewage systems has been in progress for fourteen
villages.
The town of Dursunbey is the center of health care and education for
its 103 villages. There is one hospital with 75 beds, plus one health center
in the town and four in the villages. There is one district library with 15,347
books. Many village students come to Dursunbey for their education. In my
survey sample (N=211), 50. 2% (105) of the tenth grade students come from
the areas villages. In 1994, a branch of Balikesir Universitys Vocational
College opened in Dursunbey. This school offers education in forest
industry, construction, and fashion design. Currently, it hosts about 200
students.
1 Equal to 2,096 feet.
80


Figure 4.1. Map of Turkey
BUlGAHiA
i
MtKff'":. Istanbul 1
/ KpMCl<
\ .GerrfiL
Bursa
Bahkesr^ Eski^eln
Samsun*
6ANKARA
.Sivas
,Mamsa
Izmir
Anatolia
,Konya
Kayaori
;rw
l^fcndsrjn S
f
160 fcvi
'h liiDm,
Trabzon
A7T
''' F*i- . ""i
W ^
\
K*\
*?
^ 'r%-
\
Erzurum
,V3i5
Kabramanniafa├é┬ž
.Anfaiys !qe? Adana "a'lSn^*
Diyarbakir*
.Sami Uria
* i .A

rstRfA-.

* y\
RAO
CVPftUS
Town of Dursunbey in Balikesir
Province, Turkey
Dursunbey has both urban and rural characteristics. Based on 2000
census data, the urban population of 14,774 was mainly concentrated in
Dursunbey; the rural population was 32,928, scattered among one hundred
three villages. Twenty-eight percent of these villages have a population size
of below two hundred people. There are sixteen villages with a population
of five hundred or more people, and only three villages have populations of
above one thousand people.
Dursunbey has a very young population, reflecting Turkeys general
age demographic. According to local Government health organizations, 23.
1 % of the population is under 15 years of age; 38. 6% is between the ages
81


of 15 to 40. The greatest proportion of the towns population is young
(approximately 61.7%). Due to the lack of employment opportunities, many
young people leave Dursunbey for the bigger cities, the latest exodus
occurring after the economic crisis of February 2001.
Two of the municipalitys government banks closed down in 2002.
The largest remaining employer of local residents is the struggling logging
industry, which provides jobs for around 1000 people. Dursunbey hosts the
second largest forest in Turkey. Sixty percent of its land is covered with
forest, however, over-cutting has led local businessmen to start bringing
timber from Russia. Another Dursunbey industry, coal mining, is on the
wane in Odakoy, Cakirca and Hamzacik, three of the districts villages. The
agriculture in the region consists mostly of subsistence farming with a minor
cash crop component.
Dursunbey has a religiously conservative population with a low level
of education. The literacy rate is 97% by 1997 figures (Dursunbey
Governors office, 2002). The town of Dursunbey has six primary schools2
and the remaining 45 schools are in the 103 villages. The total number of
pupils in primary schools is 6,320 in the greater Dursunbey area. Girls and
boys constitute 3,094 and 3,226 of the students respectively. In addition,
there are five high schools with 885 students (278 girls). The total number
of students in all educational institutions is 7,205 (i. e., primary and high
schools combined). The five high schools have a total of 103 teachers; 226
teachers are working in Dursunbeys middle schools. There is one teacher
for every 25 students.
During April and May 2002, I conducted a survey with the census of
all of Dursunbeys 211 tenth graders, ages 15 to 17, using a 57-item, self-
2 Primary school in Turkey consists of 8 years compulsory education that includes 5 years
of elementary school and 3 years of middle school.
82


administered questionnaire. Nine variables in the survey described the
study population: school type, gender, age, mother's and fathers
educational attainment, household income, perceived economic status,
student's weekly allowances, place of origin, and place of residence. Each
variable is discussed separately below.
B. High Schools in Dursunbev
There are five high schools in Dursunbey. The high schools
represent a wide range of school types that are common in Turkey. They
are regular high school with multiple programs, religious high school, health
vocational high school, female vocational high school, and industrial
vocational high school. There are a total of 885 students in these five high
schools, in grades 9 through 12. Table 4.1 represents the population
distribution among five high schools.
Table 4.1. School Type
School Type Girls Boys Total
Number Percent Number Percent Number Percent
Religious - - 14 100. 0 14 6. 6
Health 16 72. 7 6 27.3 22 10.4
Regular 31 42. 0 43 58. 1 74 35. 1
Girls 7 100. 0 - - 7 3. 3
Industrial 9 9. 6 85 90.4 94 44. 5
Total 63 29. 9 148 70. 1 211 100. 0
Although all these schools lack facilities for physical education,
organized soccer teams are very popular among the male students. For a
83


limited number of students, these schools offer intensive classes for college
preparation.
Dursunbev Regular High School
Dursunbey High School is the largest regular high school with 304
students (190 boys and 114 girls). Usually the students of this school have
high academic aptitude and plan to get into college after graduation.
However, the rate for college acceptance is as low as 5 or 6 students per
year due to an intense competition for college entrance. Annually, 1.5
million students participate in the nation-wide placement examination for
college; of these, approximately 100,000 students get accepted to enter.
The curriculum of this school is geared toward the college preparation exam.
In addition to hard science classes, psychology, sociology, logic, philosophy,
and literature classes are taught. This school also has a training division for
accounting. The students who enter into this program take a nationwide
placement exam. These students work outside the school as interns for two
semesters. Some of these students stay at the dormitories.
This high school draws nearly 45% to 50% of its population from
surrounding villages. Due to the distance from the school, some of these
students live in a private dormitory. There are 74 10th graders in this school
in four classes. Forty-two percent of these students were females.
Religious High School
There are 102 students (70 boys and 32 girls) in this school. More
than 90% of its population comes from villages. Usually these students are
from low-income farm families and live in a school dormitory. These
students have a very slim chance of getting into college, due to lack of
84


college capacity in their chosen specialty. The curriculum of this school is
similar to regular high schools in terms of science classes and literature. In
addition to these classes, they study the Koran and prepare for the entry-
level religious cleric positions that would available to them after graduation,
especially in the villages. There was only one tenth-grade class in religious
high school, an all-male class of 14 students.
Health Vocational High School
There are a total of 123 (61 boy and 62 girl) students in this school
unlike the other schools, this school is managed under the Ministry of
Health. It is one of 382 of its type of school in Turkey. The students of this
school are placed based on a nationwide exam. As a result, this school
enrolls students from various other cities and towns in the region. Fifteen
percent of them come from Dursunbey, 20 percent from cities and towns
other than Dursunbey, and 65% of them come from villages of Dursunbey
and villages of adjacent towns (e. g., the towns of Sindirgi, Bigadic, Kepsut).
Since this school does not provide a dormitory for its out-of-town students,
most out-of-town students rent apartments or stay with relatives. There was
only one tenth-grade class in this school with 22 students; 73 percent of
them were females.
Until 1999, the students of this school received additional
consideration for college acceptance. Graduates could expect employment
both in local hospitals and in other cities. However, the prospects for
employment for these graduates collapsed with the Turkish economy and
shrinking government expenditures. The last alumnus to find a job
graduated from this high school in 1994.
85


Girls Vocational High School
This school has 57 students, with nearly all of them coming from
Dursunbey. Girls who go to this school learn home development, childcare,
and other household-related subjects. In addition to this regular curriculum,
classes are taught in subjects such as literature, geography, and
mathematics. Most of these students do not plan to attend college, and
would have virtually no chance if they did. There was one tenth-grade class
in this school, having only 7 students.
Industrial Vocational High School
This school hosts 299 students (286 boys and 13 girls). Fifty to sixty
percent of its students come from villages and stay in the school dormitory.
There were seven 10th grade classes in this school with 94 students. The
school curriculum prepares students for work in the forest industry, the metal
industry, and in the mechanics of various engines. In addition, regular
science, health, literature, and foreign language courses are taught. The
students usually take internships in local industries during the school year
and summer break. They are usually not very academically motivated and
do not plan to pursue a college degree, something which is beyond the
financial means of many of these students families. This school also
houses the Anatolian technical high school, which accepts students through
a nationwide exam. Unlike their counterparts in the rest of the school,
students in the Anatolian technical program have four years of coursework,
which includes English and computer instruction. As a result, these
students have an academically different profile than the rest of the high
school.
86


Full Text

PAGE 1

SOCIOCULTURAL INFLUENCES ON TOBACCO USE BY TURKISH YOUTH: FINDINGS FROM QUANTITATIVE AND QUALITATIVE METHODS by Hulya Yuksel B.A., Aegean University M.A., Aegean University M.A., University of Colorado at Denver A thesis submitted to the University of Colorado at Denver in partial fulfillment of the requirements for the degree of Doctor of Philosophy Health and Behavioral Sciences 2003 t _;; .. t i '"

PAGE 2

This thesis for the Doctor of Philosophy degree by Hulya Yuksel has been approved by Deborah S. Main James Nimmer Date

PAGE 3

Yuksel, Hulya (Ph.D., Health and Behavioral Sciences) Sociocultural Influences On Tobacco Use By Turkish Youth: Findings From Quantitative And Qualitative Methods Thesis directed by Professor Kitty Corbett ABSTRACT This study identified social and cultural factors that influence smoking behavior of rural/urban youth, age 15-17, who attend high school in Dursunbey, a small town in Turkey's Aegean region. The methods consisted of a cross-sectional, school-based survey of all 1Oth graders (N=211) in 5 high schools in Dursunbey, augmented by qualitative semi structured interviews with 42 1Oth grade students, 10 students from other grades, 14 adult educators, and 4 other adults. The survey identified several risk factors for current smoking of adolescents (any smoking within the last 30 days): the number of close friends who smoke and teacher smoking, male gender, age, and academic achievement. Urban/rural origin, pocket money, family smoking behavior, and school type were not statistically significant. Exposure to teachers who smoked was associated with increased smoking risk. Males were at 12 times the risk of smoking. Both adolescents and adults say that "everybody smokes." Interview data revealed an adolescent need to fit in with a group, a sense of shared excitement in transgressing the rules imposed by adult authority, and pressure resulting from a strong association of smoking with the Turkish iii

PAGE 4

male identity. Students identified with their teachers more than their parents; anti-tobacco messages at school were diluted by examples of teachers smoking. Females are discouraged from smoking, a protection that appears to be waning with Westernization and dramatically rising ever smoking rates. Academic achievement appears protective against smoking, but with few recreational alternatives, students referred to smoking as something to "pass the time." Students often described themselves as having no future, since few students are accepted into college and jobs are scarce in Turkey's depressed economy. The social environment has highly prevalent pro tobacco messages in the media, in popular songs, and in social and religious rituals. In a culture where objections to second-hand smoke are considered rude, expensive foreign cigarette brands are considered status symbols by the young. The conclusions of this study, the first with Turkish youth to employ mixed methods, should contribute to more effective tobacco control interventions, to the benefit of Dursunbey and other communities in Turkey. This abstract accurately represents the content of the candidate's thesis. recommend its publication. Signed iv

PAGE 5

DEDICATION This dissertation is dedicated with love and appreciation to Timothy, my ultimate supporter and biggest fan. Your words of encouragement meant the world to me. Thank you for always being there during my hardest times. I would also like to dedicate this work to my parents and my uncle. My favorite uncle, a life-long heavy smoker, was recently diagnosed with advanced cancer.

PAGE 6

ACKNOWLEDGMENT First and foremost, I would like to thank my dissertation committee members for their guidance and continuous assistance throughout this process. Kitty (Corbett), it has been a privilege to work with you and have you as my mentor. Your insightful comments and recommendations about the project and future career decisions have helped me enormously. You encouraged me to use both qualitative and quantitative methods that have yielded a work rich in both substance and meaning. Thank you so much for always believing in me. Sharry (Erzinger), you were there from the beginning of this long journey and closely followed my progress as a researcher. You were always eager to hear about my interviews and field experience. I always felt that you understood the nuances of a bilingual research work its promise of a simultaneously rewarding and challenging process. Thank you for all your encouragements and a kindness very unique to you. Debbi (Main), you shared your expertise in mixed methods. Your guidance was invaluable in formulating my study design. You encouraged me to create a manageable project. Despite your busy schedule, you stayed the course and kept the process going. Thank you for staying on my committee. Augie (Diana), you were there throughout my journey and provided me with some the most insightful comments. Your words of encouragement motivated me to continue. I appreciate all your help, reading my documents multiple times, your genuine interest in my topic and the long distance phone calls from D.C. You kept the process positive for me. Thank you! Jim (Nimmer), your generous support and guidance of my "anxious" journey through statistical analysis helped me going. You did not mind working with me many hours on my data. Thank you for agreeing to be on my committee despite your busy schedule. Candan (Duran-Aydintug), you were part of my shadow committee. Thank you for agreeing to be on my committee so late in the game. Being from the same country of origin connected us in a very special way. We did not have

PAGE 7

to explain many things to each other. So many times, your jokes made my day. Thank you! I would also like to acknowledge and thank the following people and organizations for their support and contributions: Tim Ene, a member of my shadow committee, for your many insightful and intelligent ideas, and your editing every word; I owe the most gratitude to you; Mansoureh Tajik, my wise guide, for your insightful comments about methods, statistical analysis, and structure of the dissertation; Jeff Schweinfest, who understood the hoops of academia, for your insightful comments on my qualitative data during the interpretation and your edits throughout my education; the principals, teachers, and counselors of the five high schools in Dursunbey especially Ahmet Hayat, Dursunbey Regular High School Counselor, for their support during my data collection; my friend, for past 17 years, Nevin Sayili-Demircioglu, for your help in the initial part of my research with the logistics in lzmir, Turkey and for your help with the data entry process. Yusuf Yuksel, my brother, your help with the logistics in Dursunbey during my data collection and being my unofficial research assistant in the field. Chris Pon, program assistant at the Health and Behavioral Sciences Program, it has been great pleasure to work with you. Finally but not last, I would like to thank Research for International Tobacco Control for their financial support of my qualitative research.

PAGE 8

TABLE OF CONTENTS Fg r I U es ................................................................................................... XIII Tables ..................................................................................................... xiv CHAPTER 1. INTRODUCTION ............................................................................. 1 A. Specific Aims ......................................................................... 3 B. The Tobacco Use Problem in Turkey ..................................... 3 Smoking Prevalence ......................................................... 6 Health Awareness and Knowledge of Health Consequences of Smoking ............................................. 1 0 C. Actions of Transnational Tobacco Companies .................... 12 History ............................................................................ 13 Marketing Strategies of Transnational Tobacco Companies ..................................................................... 16 D. Tobacco Control Efforts and Tobacco Companies' Response ............................................................................ 17 E. Significance of the Study ...................................................... 18 F. Overview of the Chapters ..................................................... 21 2. THEORETICAL APPROACHES FOR UNDERSTANDING SOCIAL INFLUENCES ON SMOKING BY YOUTH ..................... 22 A. Overview of Relevant Theories ........................................... 22 B. Social Cognitive Theory ....................................................... 31 Early Behaviorist Origins of Social Learning Theory ....... 32 viii

PAGE 9

Vicarious Learning ......................................................... 32 Steps in the Acquisition of an Abusive Habit ..... ............ 33 Social Cognitive Theory Evolves from Social Learning Theory ...................................................................... ..... 34 Reciprocal Determinism .............................................. ... 34 The Constructs of Social Cognitive Theory .................... 35 C. Family, Peer, Friend, and Teacher Influences ..................... 38 Peer and Friend Influences ............................................ 39 Family Influences ........................................................... 41 Teacher Influences ........ ................................................ 42 School Policies ............................................................ ... 42 D. Sociodemographic Factors ................................................. 43 Age and Gender ... ......................................................... 43 Academic Achievement ................................................. 44 Urban/Rural Location and Economics ............................ 45 E. Social, Cultural and Psychological Factors in Turkish Tobacco Use ... ................................................................... 46 F Application of Social Cognitive Theory: Study Hypotheses 48 Hypotheses ........ ...................................................... ... 50 Sociodemographic Factors ............................................. 51 Social Environment Factors ........................................... 52 3. METHODS .................................................................................... 55 A A Study Design .................................................................. 55 B. Field Season One in lzmir and Dursunbey, 2001 ................ 58 Construction of the Survey for the Field Season One .. 59 Focus Groups ............................................................... 61 ix

PAGE 10

Semi-Structured lnterviews ........................................... 61 Field Season One Data Collection and Sampling ........ 62 Analysis ........................................................................ 63 Field Season One Study Results .................................. 63 C. Field Season Two in Dursunbey, 2002 ................................ 65 The Survey in Dursunbey, 2002 ..................................... 65 Measures ........................................................................ 70 Data Entry ....................................................................... 7 4 Analysis .......................................................................... 75 Semi-Structured Interviews in Dursunbey, 2002 ............. 75 Methodological Limitations ............................................. 78 4. DURSUNBEY, ITS HIGH SCHOOLS, AND CHARACTERISTICS OF THE STUDENTS .................................. 79 A. The Town of Dursunbey ....................................................... 79 B. High Schools in Dursunbey .................................................. 83 Dursunbey Regular High School ..................................... 84 Religious High School ..................................................... 84 Health Vocational High School ....................................... 85 Girls Vocational High School .......................................... 86 Industrial Vocational High School ................................... 86 C. Characteristics of Students Who Were Surveyed ................ 87 Age and Sex Distribution ................................................ 89 Family Origin ................................................................... 90 Place of Residence ......................................................... 91 Educational Attainment of the Parents ............................ 92 X

PAGE 11

Economic Status ............................................................. 92 Academic Success ......................................................... 95 5. SURVEY FINDINGS ...................................................................... 96 A. Introduction ................................................................. ....... 96 B. Section One: Data Integrity .................................................. 96 C. Section Two: Descriptive Statistics ...................................... 98 Description of Sample .................................................... 98 Description of Variables ................................................. 99 Bivariate Analysis to Describe the Relationships between Variables ........................................................ 1 00 The Study Population and Smoking Behavior .............. 1 02 D. Section Three: Test of Hypotheses ................................... 1 09 Test of Sociodemographic Factors .............................. 110 Test of Social Environment Factors ............................. 114 E. Section Four: Model Testing ............................................. 118 Sociodemographic Factors and Smoking Behavior ..... 118 Social Environment Factors and Smoking Behavior ..... 122 Sociodemographic and Social Environment Factors .... 124 Other Survey Findings .................................................. 126 6. DISCUSSION .............................................................................. 129 A. Introduction ....................................................................... 129 B. A Review of Findings ......................................................... 130 Incidence of Ever-Smoking ........................................... 130 Incidence of Current Smoking ...................................... 131 Influence of Sociodemographic Factors ....................... 131 Influence of Social Environmental Factors ................... 140 xi

PAGE 12

C. Theoretical Implications of the Findings ............................ 149 Limitations ................................................................... 159 7. CONCLUSION ............................................. .................. ............ 162 A. Summary of Findings ............................................................. 162 Recommendations for Future Studies and For Intervention ..................................................... ............... 164 Policy Implications .......................................................... 168 APPENDIX A. SURVEY INSTRUMENTS AND SEMI-STRUCTURED INTERVIEW GUIDE .................................................................... 173 Preliminary Survey (English, June 2001) .................................... 174 Preliminary Survey (Turkish, June 2001 ) ..................................... 180 Final Survey (English, April 2002) ............................................... 186 Final Survey (Turkish, April 2002) ............................................... 194 Semi-Structured Interview Guides ............................................... 201 B. FIELD SEASON ONE QUALITATIVE AND QUANTITATIVE FINDINGS, SUMMER 2001 ......................................................... 203 C. TURKISH SONGS REGARDING SMOKING AND POEMS BY ONE OF THE STUDY SUBJECTS (English Translation) ...... 216 D. CONSENTS AND HUMAN SUBJECTS APPROVALS ................ 220 BIBLIOGRAPHY ................................ ...... ............................................ 229 xii

PAGE 13

Figure 3.1 4.1 6.1 6.2 6.3 6.4 FIGURES Timelines for the Data Collection and Write-Up ............................. 56 Map of Turkey ...................................................... ............... . ....... 81 Predictors of Adolescent Current Smoking ... .............. ............. ... 151 Reciprocal Interaction of the Predictors of Adolescent Current Smoking ....... .............. .... ................................................ ..... ..... 152 Partial Model for Rural Environment Using Qualitative and Quantitative Findings ....... .................................... ................ ..... 154 Partial Model for School Environment Using Qualitative and Quantitative Findings ...... ........... ............................... ....... ........ 157 xiii

PAGE 14

TABLES Table 2.1 Selected Factors Influencing Tobacco Use .................................. 49 3.1 Structure of the Questionnaire ...................................................... 60 3.2 Definitions of Adolescent Smoking ................................................ 71 3.3 Explanatory (Independent) Sociodemographic Variables ............. 72 3.4 Explanatory (Independent) Social Environment Variables ............ 74 4.1 School Type ........................... ...................................................... 83 4.2 Age Categories ............................................................................. 90 4.3 Family Origin .................................................................. . ............ 91 4.4 Place of Residence ...................... .................................... ............ 91 4.5 Parents' Education ........................................................................ 92 4.6 Monthly Household Income ........................................................... 93 4.7 Perceived Economic Situation ....................................................... 94 4.8 Weekly Student Allowance ............................................................ 94 4.9 Academic Achievement.. ............................................................... 95 5.1 Variables with Missing Values ...................................................... 97 5.2 Description of the Study Population .............................................. 99 5.3 Descriptive Statistics of Variables ............................................... 100 5.4 Correlations Between Variables .................................................. 1 01 xiv

PAGE 15

5.4 Correlations Between Variables (Continued) ............................... 1 02 5.5 Smoking Behavior in the Study Population .................................. 104 5.6 Sociodemographic Factors and Smoking Behavior ..................... 1 06 5.7 Social Environment Factors and Smoking Behavior .................... 108 5.8 Hypothesis Testing for Sociodemographic Factors ..................... 112 5.9 Hypothesis Testing for Social Environment Factors ............... .... 115 5.10 Full Model for Sociodemographic Factors and Current Smoking Behavior ........................................................................ 120 5.11 Final Model for Sociodemographic Factors and Current Smoking Behavior. ....................................................................... 122 5.12 Full Model for Social Environment Factors and Current Smoking Behavior ....................................................................... 123 5.13 Final Model for Social Environment Factors and Current Smoking Behavior ....................................................................... 124 5.14 Final Model for Predicting Adolescent Current Smoking Behavior ...................................................................................... 125 5.15 Companions When Adolescents Try Smoking ............................ 126 5.16 Why do you think people of your age start to smoke? ................. 127 5.17 Why did you start to smoke? ....................................................... 128 6.1 Influences in the Partial Model for Rural Environment ................ 155 6.2 Influences in the Partial Model for School Environment .............. 158 XV

PAGE 16

CHAPTER 1 INTRODUCTION "the base of our business is the high school student" Source: Lorillard Tobacco Company, 8/30/78 Bates No.03537131/7132 "today's teenager is tomorrow's potential regular customer'' Source: Philip Morris Tobacco Company, 3/31/1981, Bates No. 1000390803 Despite the massive literature on social influences on youth smoking in the developed world (Conrad et al., 1992; Evans et al, 1988; Flay & Phil, 1985; Flay, 1993; Flay et al., 1994, 1997, 1999; Flay & Petraitis, 1994; Petraitis et al., 1995; Shadel et al., 2000; USDHHS, 1994 ), findings about factors contributing to uptake and addiction cannot be assumed as applicable and useful across different social, political and cultural settings. Patterns of tobacco uptake and maintenance must be understood and addressed in specific contexts. The normative environments in developing countries such as Turkey differ significantly from the environments of Western nations. Countries differ widely in the political and economic context of daily life. Developing countries have higher mortality rates and greater uncertainty concerning the near-term future, as well as a lower likelihood of a long life. Smoking and smoking-related behavior function as key social symbols, and smoking control campaigns based on an inaccurate understanding of the social context in which smoking occurs can reinforce the very behavior they are designed to prevent (Eckert 1983). 1

PAGE 17

Understanding of perceptions and attitudes about smoking, as a social practice and a health risk must be situated within a broader set of sociocultural concerns. Sound scientific findings specific to every nation are needed in order to carry forward effective tobacco control policies and programs. As was asserted by Flay & Petraitis (1991 :84 ), social environment variables predict tobacco use. Therefore the structure of the economic, legal, social, and educational systems of a society are determinants of behavior. In an effort to understand tobacco use in Turkey, and ultimately to design effective policies and interventions, I carried out research with 1Oth grade students in a west-central Turkish town, Dursunbey, that includes both urban and rural populations. The aim of the study was to explore the social, environmental and sociodemographic factors that influence the smoking patterns of high school youth, in the critical age group of 15 to 17 years. The research employed qualitative data from interviews with sets of adult and adolescent informants, and a quantitative study of a cross sectional survey of all 211 tenth grade students at the community's five high schools. The research design allowed description of tobacco knowledge, attitudes, and behaviors, tests of a series of hypotheses, and development of a predictive model of smoking use by youth using logistic regression techniques. Social Cognitive Theory, with emphasis on its environment, observational learning, and situation constructs, served as the theoretical foundation for the research (Akers, 1998; Bandura & Walters, 1963; Bandura, 1977,1986, 1991; Baranowski etal, 1997). The intent of this research is to provide Turkey with its own base of data that is useful for tobacco interventions. 2

PAGE 18

A. Specific Aims This research identifies social factors, as defined in the Social Cognitive Theory (Bandura, 1986, 1991 ), that are associated with tobacco use in Dursunbey, Turkey. Specific aims are: Aim 1: To explore social environmental and situational factors that are associated with smoking behavior of 1Oth grade students, including associations of adolescent smoking behavior with: smoking among family members smoking among teachers smoking among friends friends' approval of smoking behavior perceived smoking prevalence among peers Aim 2: To explore how patterns of smoking for 10th grade students are associated with the following sociodemographic factors: gender age place of origin (rural/urban) school type perceived economic situation weekly allowances perceived academic achievement To accomplish these aims, the researcher collected the data in Turkish. Translation occurred during the data analysis. Hypotheses based on Social Cognitive Theory and the preliminary work are explained in Chapter 2. B. The Tobacco Use Problem in Turkey Tobacco-related diseases are the single largest cause of preventable death and disability among adults in the world (Chandler, 1986; Stebbins, 1990; Trevathan et al.; 1999). Worldwide almost 47% of men and 12% of 3

PAGE 19

women smoke (World Bank, 1999). Developing countries face a huge and growing tobacco-related burden of human suffering, mortality, and economic losses. By 2030, 7 million of the 10 million people who die from tobacco related disease will be in developing countries (Mackay & Crofton, 1996; WHO, 2002). Adult cigarette consumption in low and middle-income countries rose by 64% from 1971 to 1991. Almost five million people died from smoking-related diseases across the world in 2000, researchers estimate (WHO, 2002). It is estimated that in high-income countries, 14,000-15,000 children and young people take up smoking, versus 68,000 to 84,000 children and young people in low and middle-income countries (World Bank, 1999). The Global Youth Tobacco Survey, carried out under the auspices of the World Health Organization (WHO) and Centers for Disease Control and Prevention (CDC), tracks the prevalence of tobacco use and provides baseline data about youth (students ages 13-15) across the world This study revealed that smoking ranged from a high of 33% to a low of 10% in fourteen developing countries around the world (Warren et al., 2000). Individuals who begin smoking at a young age are more likely to become heavy smokers (World Bank, 1999; Zhang et al., 2000). They are also more likely to die from smoking-related diseases later in life. To prevent smoking and other tobacco use, programs logically target youth. Understanding youth uptake of tobacco is necessary in order to design policies and programs that can effectively counter the influences fostering initiation and regular use. Although studies throughout the world suggest shared factors ranging from gender differences to interpersonal and intrapersonal influences (Chassin et al., 1990; Conrad et al., 1992; Flay & Petraitis, 1991, 1994; Flay et al., 1999; Petraitis et al., 1995; Rissel et al., 2000), efforts to reduce tobacco use in developing countries are characterized by over-reliance on 4

PAGE 20

research carried out in, by, and for persons in wealthier Western nations. Studies in high-income countries have found that most tobacco addiction begins at a young age, as a result of a broad array of factors implicated in tobacco initiation, and have developed interventions targeting youth (Kessler, 1995; Smith et al., 1999; Tilly, 1991 ). Adolescents in western cultures frequently report curiosity, social norms, and social pressure as reasons for smoking uptake (Sarason et al., 1992). For much of the rest of the world, relatively little is known about specific attitudes, knowledge, behaviors and social influences that contribute to starting or avoiding tobacco use. Culture-specific information to classify the reasons adolescents engage in tobacco use is required if we are to design effective smoking prevention programs (Sarason et al., 1992). The larger context may also contribute to Turkey's tobacco problem. Turkey's population of 68 million is geographically and culturally positioned at the crossroads of Europe and Asia. With 40% of its population under 18 years of age and growing at a 1.2% annual rate, Turkey has a very young population. According to 1999 data, 65% of the population lives in urban areas. The overall literacy rate for the general population is 86.5% (94% for males and 79% for females). Based on the World Fact Book (2002 est.), Turkey is considered a low-income country with an estimated current GOP of $7,000 per capita. Turkey has been experiencing a series of economic crises for over a decade, the most severe hit Turkey in February 2001, after the government was forced to let its currency float. Since that time, the Turkish lira has lost 50% of its value against the US dollar, a fact that is a continuing source of stress for most workers, businessmen and blue-collar workers alike. This, combined with a 10.8% unemployment rate (plus underemployment of 6.1 %) (2002 estimate) and a near-hyperinflation rate of 5

PAGE 21

45.2% (The World Fact Book, 2002, est.), has created a very unstable environment for Turkish citizens. Smoking Prevalence Overall Prevalence. Smoking in Turkey is a widely practiced, socially accepted habit. With its consumption of 121 billion cigarettes a year in 2001, Turkey has one of the largest markets for tobacco products in the world (Karanci & Rustemli, 1987). Its 2001 annual cigarette consumption is a 21% increase from its 1999 annual consumption of 100 billion cigarettes. Annual per capita cigarette consumption in Turkey rose from 1 ,230 in 1985 to 1 ,495 in 1992. For persons above 15 years of age, it increased from 1,850 in 1965 to 2,600 in 1992 (Firat, 1996). Adult smoking rates vary in different studies. Adolescent smoking rates also show a good deal of variation among various studies. In the recent Tobacco Control Country Profiles (WHO, 2003) smoking rates for 20 years and older males was 50.9% and 10.9% for females. These most recent rates are lower than previous studies in Turkey. In the 2000 nationwide study, smoking rates for 18 years and above ages were reported as 58.7% for males and 26.4% for females (A&G Research Reports, 2000). The problem is not limited to urban areas in Turkey. Metintas et al. (2000) identified smoking as one of the most important public health problems in rural areas within the governmental district of Eskisehir. In their survey of adults ages 20 and older in 52 villages and 19 sub-centers, the frequency of smoking was 44.3% in villages and 48.7% in sub-centers (a terminology used by Metintas et al.). Among certain Turkish occupational groups, smoking rates were reported as much higher than the national average. In Dedeoglu et al.'s (1994) study in Antalya, 53% of male physicians and 41% female physicians 6

PAGE 22

were smokers. Smoking rates are high among teachers also. In Saltik et al. (1989) study, 62% of male teachers and 34% of the female teachers were smokers. In Bilir et al.'s (1997) study in Ankara, 64% of journalists, 51% of teachers, 44% of physicians, 46% of artists, 27% of parliamentarians, and 35% of professional sportsmen were smokers. For some occupational groups, smoking rates climbed even higher, reaching 7 4.3% among taxi and truck drivers. In Bagci et al.'s study (1996), 95% of drivers were smokers. The high percentage of heavy smokers is also of concern in Turkey. In earlier studies, 64% of the adult population ( 15 years in age and older) indicated that they smoke one pack or more cigarettes a day (PIAR 1988, in Asut, 1993). Unlike high-income countries, health care professionals smoke at high rates in Turkey. Individuals in Turkey who should best understand the dangers of smoking, i.e., doctors and medical students, appear uninfluenced by information. The smoking rates of male medical doctors (62.4%) matches the smoking rates of the general population of adult males (62.5%). A comparison of smoking patterns among medical school versus non-medical school students revealed that the prevalence of smoking among first-year medical students was 11.8% vs. 21.0% among non-medical students (Kocabas et al., 1994 ). There was a significant jump in the smoking rates in the last year of school for students in both groups, increasing to 30.2% among medical students and 32.4% among non-medical students. Gender and Tobacco Use. Gender differences are important in tobacco use in Turkey. In Turkey many of the professional women and women who live in urban areas are more likely to smoke than women in rural areas. Female smoking rates are reported to vary by occupation in several studies (Bilir and Onder, 2000; Firat, 1996). Saltik et al. 's ( 1988) 7

PAGE 23

study found a 11% smoking rate for female doctors in Elazig. In a more recent study, Bilir et al. (1999) found a 34.4% smoking rate for female physicians nationwide. Kocabas et al. (1994) found that female doctors and medical students were more likely to smoke than women in general: 44.1% of female doctors and 34.9% of female final year medical students were smokers, while only 10.9% of the females in the general population smoke. The high smoking rates among female professionals are illustrative of recent trends in women's smoking in Turkey. Traditionally in the Middle East and many Islamic countries, smoking has been seen as an inappropriate and immoral behavior for females (Dagli, 1999). The figures above, however, indicate that Turkish women's attitudes toward smoking are rapidly changing. This was also evident from a 1996 study of 538 randomly selected women in Istanbul, finding an overall smoking prevalence of 48% (Dagli, 1999). This pattern is replicated elsewhere in the Middle East. Maziak (2002) points out that smoking is associated with social liberalization among women in Syria. It is a well-known fact that transnational tobacco companies have identified women as an important target (WHO, 1996), using the Western symbol of the liberated woman to sell cigarettes to the Third World (Headden, 1998). Especially in developing countries, smoking has been promoted as a way of liberalization and Western modernism for women. Not surprisingly, a 1996 industry review noted that "the anticipated growth in cigarette sales in Turkey is based on the increase in smoking amongst women and the lack, at the present, of a strong anti-smoking lobby" (Hammond & Purcell, 1999:1 ). Gender plays a role in recreational activities in Turkey, with potential implications for tobacco use. Particularly in small towns, the sexes often 8

PAGE 24

separate for evening relaxation. The men frequent all-male coffee houses, where tobacco smoke is always densely present. The women gather in thei r homes for social teas. To be able to speak freely, they usually send their young adolescent males away. This usually results in these boys spending more time with peers in the evenings, sometimes even in coffee houses. Minors are not officially allowed into these places, where the association of smoking with the male identity is almost unavoidable Youth and Tobacco Use. The smoking rate among Turkish youth ages 15 to 18 was reported to be 20%, with 23 1% for males and 14.8% for females (WHO, 2003). Data from 1986 on youth smoking reported higher rates for males (31 %) and lower rates for fema l es (5%) between the ages of 15 and19 (Hammond & Purcell, 1999). The prevalence of smoking among youth has been provided by several other recent studies (Aiikas i foglu et al., 2002; Bilir et al., 1997, 2000; Cel i k et al., 2000; Derman et al., 1995; Erbaydar et al., 2002; Emri et al., 1998; Karababa et al., 1997 ; Ozcan& Ozcan, 2002). In Ankara, the capital of Turkey, 33% of high school students were classified as "ever smokers and 28.3% (32 % boys and 20% girls respectively) were current smokers, with a mean number of 7 cigarettes per day (Bilir et al., 1997). In Bilir et al.'s (2000) nationwide study among 1Oth graders, 32% of students were ever-smokers (64% for boys and 22% for girls), and 18% of the boys and 11% of the girls were current smokers. In the same study, there were high numbers of ex-smokers, 18% o f boys and 11% of girls. The most current nationwide data on 1Oth graders reported that 58% of the students had ever tried smoking, and 33% of boys and 18.3% of girls between ages 15-17 reported they were current smokers (Erbaydar et al. 2002). A marked increase in the rate of smoking by girls is clear: 1986 WHO 9

PAGE 25

data showed that 31% of boys and 5% of girls ages 15 to 19 were smokers in Turkey, an increase particularly for girls. Karababa et al. (1997) surveyed 990 Apprenticeship School students in lzmir. Fifty-four percent of boys and 53% of girls had tried smoking, but identified themselves as nonsmokers currently. Twenty seven percent of boys and 37.1% of girls were found to be current smokers, with an average age of 17 and reported age of uptake between 13-15; this was the only study that demonstrates higher smoking rates for girls than boys. A survey of 1066 high school students in Manisa, Turkey (Celik et at., 2000) found that smoking prevalence among male students was 17.2%, and 7.1% among female students. The mean age for beginning smoking was 13. Health Awareness and Knowledge of Health Consequences of Smoking According to Bilir and Onder's recent report (2000), 70-100 thousand Turkish people die due to smoking-related health problems each year, 10-15 times the number who die in motor vehicle accidents in Turkey. The most frequent cause of death in Turkey is cardiovascular disease, followed by cancer and then cerebrovascular diseases. In Turkey, cancer incidence has tripled between 1983 and 1990, with bronchoalveolar cancer identified as the primary cause of cancer related deaths (Bilir et al., 1997). Thus smoking plays an increasingly important role in the increased incidence of cancer. It is estimated that smoking may be responsible for 50% of cancers in Turkey. A 1996 study shows that of the major of causes of cancer deaths in Turkish men, carcinoma of the lung accounts for 30-40% of deaths and carcinoma of larynx accounts for 7.5% of cancer deaths. Moreover, lung cancer is the 10

PAGE 26

leading cause of cancer death in females, with a 16% relative mortality rate (Firat, 1996). The effects of smoking are not simply a matter of premature death. Smoking is also an added burden in terms of health costs and misery, predominantly in regions of the earth where misery is already higher than in most developed part of the world. A portion of these diseases will not be reflected in higher healthcare costs because, in many countries, these diseases will go untreated. By the same token, it is difficult to estimate direct and indirect healthcare cost of providing treatment to the Turkish smokers, and the loss in Turkish productivity. Tobacco-related disease costs US more than $123 billion annually with $50 billion in avoidable medical expenses and $73 billion in lost productivity. Over 25 years, this could amount to $3 trillion (Moskowitz, 1998). For developing countries struggling to provide basic services and a subsistence-level existence for their citizens, almost any loss in productivity is tragic. Over a period of a decade, 0.5% annual loss in productivity in a country translates to a 5% relative loss of wealth, potentially affecting the well being of thousands of people. In addition to productivity losses, cigarette purchases divert badly needed family income. As was noted above, an estimated 10 to 15% of Turkish household budgets are spent on tobacco It is far more efficient for a family to simply put their income to better use. In a survey of 1 066 high school students in Manisa, Turkey (Celik et al., 2000), 95% of the students stated that they were aware that cigarettes were harmful and 95 8% knew about the close relationship between smoking and lung cancer. Thus knowing the adverse health effects of smoking does not stop young people in Turkey from becoming smokers. Similarly, in another study in Istanbul, Turkey 100 medical students were asked about the adverse health effects of smoking. One hundred percent of 11

PAGE 27

students were aware of the adverse health effects of smoking but this knowledge was not enough to stop them from smoking (Derman et al., 1995). C. Actions of Transnational Tobacco Companies Turkey is a tobacco growing country; an estimated 500,000 farmers grow tobacco (Dagli, 1999). Their families account for 3 million people. Tekel, the state monopoly, is a huge state industry covering both tobacco and alcohol production and employing more than 36,000 people (Tobacco Reporter, 2001 ). The Turkish government has been following the recommendations of the International Monetary Fund (IMF) to privatize the state monopoly for many years, although this effort has been a very controversial and is moving slowly. Philip Morris is especially keen to see Tekel taken out of government hands. In 1997, Marco Terribilini, General Manager of Philip Morris in Turkey, said: Turkey is a significant market with major potential. ... Only 18% of the population is above 45 years of age .... The Turkish monopoly has to be privatized totally .... Philip Morris will play a role under any condition in this privatization (Dagli, 1999:117). When the privileges of the Turkish state monopoly were rescinded in 1990, the economic consequences for the Turkish population were significant. Despite the dramatic increase in Turkish tobacco consumption, use of domestically produced tobacco actually decreased. With the rise of tobacco imports, local farmers lost their market and their subsidies. In 1991, 329,000 tons of excess Turkish tobacco were burned. The percentage of American tobacco in domestically consumed cigarettes rose from 0% in 1987 to 66% in 2000. Three million people were affected by these changes. 12

PAGE 28

Many Turkish export companies went out of business. Aside from the obvious economic impact, the battle against Turkish youth's tobacco use became a battle against formidable American tobacco companies and their deep pockets. Currently Philip Morris's enjoys a quarter share of the Turkish cigarette market. History Until the early 1980s, Tekel held exclusive rights to manufacture and sell tobacco products. The monopoly's run-down factories produced inferior quality cigarettes that were harsh tasting (Country Case Study, 1999). In 1984 President Turgut Ozal, a former employee of the World Bank, led a drive to privatize the Turkish markets and prepare Turkey for membership in the European Union. As part of that drive, Tekellost many of its monopoly rights, opening up its market to foreign tobacco imports. However, Tekel maintained control over pricing and distribution. Consequences of this decision included: An increase in tobacco imports as tobacco advertising and promotion became more sophisticated and aggressive, targeting women and the young for the first time. Overall consumption per capita rose 23% from 1985 to 1991, in spite of the skewing effects of a high birth rate. For the population over 15 years of age, per capita consumption rose 41% from 1965 to 1992. As cigarette consumption increased, the amount of Turkish tobacco used in the cigarettes decreased. In 1991, Tekel had to burn 329,000 tons of excess tobacco. Many Turkish export companies went out business, dropping from 59 firms to 14. In 1987, Turkish cigarettes contained no American tobacco. In 2000, the amount was 66,000 tons. 13

PAGE 29

In 1990, Philip Morris signed an agreement with Sabanci Holdings, a powerful Turkish company exerting considerable influence on Turkish tobacco policy The parties agreed that the government regulations would change, effectively exempting Philip Morris from any regulation of its prices distribution and sales practices (Dagli 1997). Philip Morris paid Margaret Thatcher, former prime minister of the United Kingdom, $1 million US for her lobbying efforts in foreign countries (Cunningham, 1996). When she visited Turkey in 1992, the press reported that the purpose of her visit was to help Philip Morris gain greater access to Turkish markets. Neither Ms. Thatcher nor Philip Morris denied the allegation (Dagli, 1997) In 1993, R.J Reynolds created its Turkish cigarette factory with $100 million in US capital. Thomas Reibl, the CEO of R.J Reynolds, said at the time: We have put the locomotive on the right tracks. Now our aim is to speed it up. We will take advantage of this country Turkey is one of the world's biggest markets. We have good production, we are working with good people, have good strategies. Turkey is promising with its young population (Dagli, 1997:116). In 1998, a partnership agreement was signed between Tekel and British American Tobacco (BAT), the world's third largest tobacco company. These developments were followed by aggressive tobacco marketing by Transnational Tobacco Companies. As the Wall Street Journal notes: [T]he battle plan Philip Morris used to capture Turkey is the same one it has perfected setting up operations in nearly 30 countries worldwide. First, it lobbied heavily to eliminate the government's control of tobacco prices, enlisting the help of one of Turkey's most influential businessmen. Then it poured tens of millions of dollars into a state-of-the-art cigarette factory, where, competitors say, it carefully engineered its cigarettes to appeal to Turkish taste buds, but with a stronger kick than local smokes It dispatched salesmen to 130,000 stores across the country, successfully recruiting many of the 14

PAGE 30

mom-and-pop shops to support a vast in-store marketing plan. And it spent lavishly on an advertising blitz of cowboys and panoramic vistas, selling a cellophane-wrapped piece of America for $1.20 (Hwang, 1998; Hammond et al., 1999). Turkish smokers marvel at Marlboro's smooth taste and powerful buzz, as compared to harsher, local cigarettes made with Turkish oriental tobacco, which is naturally lower in nicotine content than the American leaf tobacco Philip Morris imports here. Turkey has seen financial and economic power used to employ influential political personalities, skirt laws, and dominate foreign commercial and governmental interests. Virtually unlimited funds fueled tobacco's legal machine, such that it never lost a lawsuit until relatively recently. As Cunningham states (1996:26): One representative of an advertising agency working for a tobacco company said that the client had 'more money than God.' Tobacco's enormous profit potential provides a tremendous incentive for keeping the industry's sales high. When it comes to efforts to combat tobacco control measures, the industry can write a blank check at any time. The entry of transnational tobacco companies in foreign markets is correlated with increased consumption (Kiebnikov, 1994; McKerrow, 1992; Poole, 1998; Tilly, 1991; Yach, 1986). Transnational tobacco companies promote increased consumption by out-marketing and out-spending the domestic producers. These companies are often far more sophisticated in their marketing strategies and promotional campaigns than the domestic producers. Tekel, for instance, never advertised. This is generally true of indigenous tobacco companies (Makary, 1998). Current research shows that the entrance of multinational tobacco companies into the Turkish market since 1984 has correlated with the increase in smoking prevalence, and increase in relative mortality from lung 15

PAGE 31

cancer for men from 25% to 40% and for women from 11% to 16% between 1985 and 1996 (Firat, 1996; Kocabas et al., 1994 ). Marketing Strategies of Transnational Tobacco Companies Turkey's population is becoming particularly young. Tobacco has strong inducements for targeting a growing youth market (Emri, 1997), although Cunningham's general assessment applies: "There is as much justification for promoting tobacco as there would have been for promoting rats during the Plague" (Cunningham, 1996:66). Transnational tobacco companies' marketing is backed by well-developed and time-tested strategies. Naturally, these strategies are particularly directed at young people (Headden, 1998; Nichter & Cartwright, 1991 ). The following is a sample of these strategies: Transnational tobacco companies distributed free cigarettes at disco parties, advertised heavily in youth magazines, and offered free tickets to a concert by a Taiwanese pop star in return for five empty cigarette packages (Tilly, 1990: 12). One Turkish high school senior reported that he was handed five free packs a day from sales representatives from Philip Morris and competitor R.J. Reynolds Tobacco Company (Hwang, 1998). Marlboro organized a 'music disco show' on 31 May 1991 World No-Smoking Dayat the Marmara University campus, Istanbul, Turkey (Dagli, 1997:116). During the passage of key anti-tobacco legislation in 1996 in Turkey:" ... tobacco control advocates started to hear that newspaper bosses who were going to lose advertising revenue, as well as advertisers, were upset about the bill. 16

PAGE 32

... eleven major newspapers ran the same headlines: 'smoke and be more intelligent', referring to an unpublished study presented at a meeting in the USA. Very kindly, Keith Ball who was phoned offered help. He woke Prof Stanton Glantz (it was 4 am in the USA) to ask about the study. Prof Glantz found the paper and faxed it in an hour. It had been misinterpreted by the Turkish press." (Dagli, 1997 : 116). That aggressive and successful marketing targets children is reflected in Emri and Bagci's (1998) study examining the recognition rates of cigarette brand names and logos by primary schoolchildren (age 8 or younger) in Turkey. They found that the Camel logo and the Samsun and Marlboro brand names were the most highly recognized of all product logos and brand names tested. Recognition of cigarette brands averaged 95.2%; by comparison, the recognition of Chee-tos (a children's snack food) was 58.1 %. It is ironic that worldwide mortality rates for children have dropped dramatically, only to have those children become tobacco customers at risk of premature death as adults (Nichter & Cartwright, 1991 ). D. Tobacco Control Efforts and Tobacco Companies' Response In 1988, the Minister of Health prepared Turkey's first anti-tobacco legislation, which was unanimously accepted at the parliament on January 9, 1991. President Ozal vetoed the bill, claiming that the restriction of advertising conflicted with freedom of trade. The Minister of Health resigned in protest. In 1996, Turkey enacted antismoking legislation that banned cigarette advertising from printed media, TV, and radio, and ends any tobacco company sponsorship of sporting events. Moreover, the sale of cigarettes to anyone under 18 was prohibited, and smoking was restricted in 17

PAGE 33

all public areas (Emri, 1997). Various other measures have been suggested, such as worldwide bans on tobacco promotion, banning sales to minors, improved health warnings, setting limits to harmful cigarette additives, and establishing smoke-free areas, and increasing taxes. The estimated price elasticity of demand for cigarettes implies that a 10% price increase will reduce consumption by 2% in the short run and 4% in the long run (Tansel, 1993). Turkish anti-tobacco legislation has been circumvented in various ways. Philip Morris and RJ Reynolds, for example, retailored their marketing strategies in which they print "price announcements" to newspapers which often "announce" that prices have not changed (Dagli, 1999; Simpson, 1997). In response to the ban on advertising, Philip Morris simply removed the name "Marlboro" from its advertisements, but left its distinctive chevron; the brand's red and white chevron is seen everywhere. Hundreds of small convenience stores are capped with lighted Marlboro signs. In Istanbul's trendy Ortakoy neighborhood, sidewalk cafes use red and white Marlboro umbrellas and trays, provided free by Philip Morris. A 19-year-old pointing to a red chevron in one store explains, "This doesn't say Marlboro, but we understand it. The law didn't change anything". Philip Morris also hosts parties at nightclubs and sponsors contests. The former head of Philip Morris's operations in Turkey said, "Of course, the first thing you do is paint the town red ... we painted every building" (Country Case Report, 1999). E. Significance of the Study Crosscurrents of cultures, demographics, and economic stresses make Turkey a compelling locale for studying tobacco use among youth 18

PAGE 34

I --The severity of Turkey's economic outlook has given today's Turkish adolescents the bleakest future in several generations. Simultaneously, Turkey continues to move westward in its cultural orientation, and the population is increasingly exposed to media images and marketing of tobacco products. Serious gaps in tobacco control literature exist between Turkey and the developed world, where most formal studies have focused on the incidence of consumption rather than the causes of consumption (Bilir et al., 2000). While the prevalence of smoking among Turkish youth has been examined, there have been very few studies addressing adolescents' reasons and motivations for smoking. Bilir et al.'s study (1997) of high school students in a single municipal area, Ankara, found that 44% of high school students stated that they smoke more with their friends. Thirty one percent of the students think that smoking helps to build a relationship, while 14% think that they seem more mature while smoking. Some adolescents smoke for relaxation, some for pleasure, and others to impress others Sixty nine percent of the high school students smoke when they are angry and 62 % smoke at rest. Twelve percent do not know what to do with their hands if they do not smoke. The reported reasons for not smoking included illness, parental influence, and cost. In addition, parental modeling may be involved in Turkish youth's development of smoking habits. In Karababa et a I.'s ( 1997) study, no relation was found between student smoking and parents' education or family structure, but parents' and colleagues smoking habits were positively correlated with student smoking. Erbaydar et al. (2002) and Ozcan & Ozcan (2002) identified risk factors for smoking among youth as follows: male gender, parent, relative and friend smoking, problems with family and school, dysfunctional family life, low grades, higher socioeconomic status, having more pocket money, spending 19

PAGE 35

more time with friends after school and in the evenings, feeling lonely, alcohol use, and neighborhood and school effects. In summary, what is known about the social influences on youth smoking in Turkey is the product of a handful of studies. More studies are needed, especially studies which employ a clear theoretical framework. Culture-specific information about the reasons adolescents give for engaging in tobacco use is required in order to design effective smoking prevention programs (Sarason et al., 1992). A better understanding of smoking behaviors will contribute to more effectively targeted tobacco control and health messages aimed at Turkish youth. This study uses survey to measure the risk associated with the social influences of gender, age, parents, siblings, relatives, friends, peers, teachers and schools. It also assesses the sociodemographic influences of gender, age, economic situation, academic achievement, and urban/rural location. Ethnographic interviews analyzed the first-hand impressions of students about their experience of smoking uptake and the influences that affected their decisions to smoke, including their recreational activities, ambitions, frustrations, and plans for the future. Moreover, they highlighted findings concerning smoking influences related, for example, to custom and gender. Students' impressions are cross-validated with the reports of adults such as teachers and school counselors. In this manner, issues of school policy and the adult perspective of the changing social environment can be taken into account. By combining the qualitative findings from interviews with the quantitative findings from the survey, a richer, more culturally specific description of the problem emerges, leading to a detailed model of the social influences on youth smoking. This provides conclusions that contribute to more effective tobacco control interventions, to benefit Dursunbey and other communities in Turkey. 20 ---J

PAGE 36

F. Overview of the Chapters This thesis is organized in eight chapters. Chapter One has provided an overview of the research aims, and background on tobacco use in Turkey. Chapter Two discusses theoretical approaches relevant to the study, with emphasis on Social Cognitive Theory, and presents the study hypotheses. Chapter Three describes the research design, preliminary studies, methods for the survey and its analysis. Chapter Four presents the setting, describing Dursunbey and its high schools, and characteristics of youth who participated in the survey are described. Chapter Five presents findings from hypothesis testing, the predictive model, and other survey results Chapter Six is a discussion of survey findings, with elucidation provided by findings from the study's qualitative data. Chapter Seven is a conclusion, including implications for future efforts to understand and control tobacco use in Turkey. 21

PAGE 37

CHAPTER2 THEORETICAL APPROACHES FOR UNDERSTANDING SOCIAL INFLUENCES ON SMOKING BY YOUTH A. Overview of Relevant Theories Although most of the literature on social influences of youth smoking reflects studies carried out in high-income countries, those studies provide a reasonable beginning point. A variety of theories, employing equivalent or overlapping constructs but differing in emphasis and directionality, have been used to explain factors relevant to smoking onset. Earlier theories are broadly categorized as intrapersonal models or interpersonal models of behavior. While the intrapersonal models used in tobacco research focus on the personal traits, states and skills of individuals as determinants of smoking behavior, the interpersonal models emphasize interpersonal interactions and the social environment that provide reinforcements, resources, and learning situations that influence smoking behavior (Lewis, 1997). Later models of health behavior combine these intrapersonal and interpersonal elements into broader models of influences where personal characteristics and the social environment interact. This chapter reviews literature on Social Cognitive Theory, social influences, and sociodemographic influences, covering the topical areas of role modeling, the influences of family, friends, peers, and teachers who are significant in the lives of adolescents. Socio-demographic factors are then treated, such as gender, age, academic achievement, urban/rural location, 22

PAGE 38

and disposable income. This is followed by an examination of social, cultural and psychological factors that are special to Turkey and other developing countries. The chapter concludes with a presentation of how Social Cognitive Theory is applied in this study in the development of the study hypotheses. In general, the behavioral theories in tobacco control have grown in sophistication and complexity as they attempt to adequately model the influences documented in the literature. The next few pages review a variety of commonly used theory in tobacco literature. Flay et al. (1999) emphasize the key role of these multifaceted theories in producing effective anti smoking interventions, stressing the need for strongly theory-directed studies that take into account the mutual influence of intrapersonal factors, culture, and the social environment. lntrapersonal models in an approximate chronological order of development, include the Health Belief Model (Hochbaum, 1958, Janz & Becker, 1984 ), the Theory of Planned Behavior and the Theory of Reasoned Action (Ajzen & Fishbein 1975), and the Stages of Change Model (Flay et al., 1983, Prochaska & DiClemente, 1982; Prochaska et al., 1987). The Health Belief Model describes health behaviors as the outcomes of personal cost-benefit decisions. It posits six determinants of behavior change: perceived susceptibility; severity of a disease; perceived benefits and barriers to action; internal or external cues to action; self-efficacy (one's confidence in one's ability to take action) (Bandura, 1997, Strecher & Rosenstock, 1997). Perceived barriers have been the most consistent predictor of preventive health behavior, and perceived severity of consequences of disease has been the weakest predictor (Zimmerman & Vern berg, 1994 ). Health Belief Model attempts to understand the determinants of health-related behaviors at the individual level in order to 23

PAGE 39

stimulate positive behavioral changes. It sees the individual as the only unit of intervention; groups, organizations, worksite and communities are almost absent from Health Belief Model. The Theory of Reasoned Action and the Theory of Planned Behavior (an extension of Theory of Reasoned Action) are the most important conceptualizations of the "attitude-behavior relationship" (Zimmerman & Vern berg, 1994 ). Generally, the strongest relationship has been found between attitudes and intentions, rather than between intentions and behavior (Zimmerman & Vernberg, 1994). Behavioral intentions to act and the role of normative beliefs in mediating between behavior and action are emphasized in both theories. The attitude toward the act is the product of outcome beliefs and outcome evaluations, and the subjective norm is the product of beliefs about how significant others feel about the behavior and motivations to comply with those other significant others (Zimmerman & Vernberg, 1994: 48). Thus, these two models include perceptions of others in their decision-making framework. Studies in the US show that nationwide use of specific substance among high school students usually declines with the increased perceived risk and social disapproval associated with the substance (Petraitis et al., 1995:69). In addition to attitudes and normative beliefs, the Theory of Planned Behavior posits that self-efficacy (i.e., perceived control over one's behavior) will affect the intentions and behaviors. Self-efficacy has two important forms: use self-efficacy and refusal self-efficacy. The first form represents adolescents' beliefs in their abilities to obtain and use substances. The second form represents adolescents' beliefs in their abilities to resist social pressure to begin substance use (Petraitis et al., 1995). This theory has been used to predict smoking behavior of youth by various researchers (Carvajal et al., 2000; Flay et al., 1998; Norman and Tedeschi, 1989). In several studies, intention 24

PAGE 40

to smoke predicts the smoking behavior several years later (Flay et al., 1998). In a limited sense, the Theory of Reasoned Action/Theory of Planned Behavior's inclusion of normative beliefs makes them attractive theories for studying social influences. They also include intentions and self-efficacy in their intrapersonal model. One problem, however, is that they do not explicitly include important forms of social learning such as observational learning. In spite of their use of norms, their greatest problem is that their treatment of the environment is limited to that aspect of the social environment, such that other features like social policy and physical surroundings are largely excluded. The Stages of Change model sees behavior as a process involving a several stages (Prohaska et al., 1997). The stage theory has important implications for understanding the different stages of youth smoking, which are usually identified as never users, triers, experimenters, and regular users (Flay et al., 1998). The focus of this model is plainly intrapersonal to the virtual exclusion of interpersonal interactions. The preferred theories used in this research are Social Learning Theory and Social Cognitive Theory. Social Learning Theory (Miller and Dollard, 1941; Akers et al., 1977) is an interpersonal theory that forms the basis for the more extensive Social Cognitive Theory (Bandura, 1986). The central principle of Social Learning Theory is observational learning, where individuals modify their behavior based on observing the behavior, rewards, and punishments of others. Social Learning Theory allows for role modeling, modeling of behavior, and social normative beliefs. (Flay et al., 1991). Like the Theory of Reasoned Action, the individual learns to adapt his/her behavior to the normative beliefs and expectations of others. Social Cognitive Theory expands on the principles of Social Learning Theory, 25

PAGE 41

incorporating Reciprocal Determinism as its central concept. In Social Cognitive Theory, interactions may occur between a person and his environment (either social or physical), a person and his behavior, and his behavior and his environment. Any of these three interactions may flow in either direction; for example, a person may influence his social environment, as well as be influenced by that environment. Social Cognitive Theory not only includes interpersonal constructs, but intrapersonal constructs as well: emotional coping responses, self-efficacy, values that a person places on outcomes (expectancies), and anticipated outcomes of behavior (expectations) (Bandura, 1997) Bandura ( 1997) extends the concept of reciprocal determinism to self-influence, to include self-control and self reward. With these constructs, Social Cognitive Theory is more than an interpersonal theory, offering a group of interactions among person, environment, and behavior. The framework of Social Cognitive Theory provides a good basis for describing social influences on smoking behavior, and was selected as the theoretical framework for my study. Social Cognitive Theory will be reviewed in greater detail later in this chapter. A model of social influences is incomplete without accounting for differences among individuals; a given social influence on tobacco use will not yield the same behavior for all individuals. This variation in response to exogenous influences is partially explained by mediating processes and moderating processes (Flay et al., 1997). In a mediation process, a variable of interest has influenced one or more intermediate variables, which in turn influences the behavior. In a moderating process, the influence of a variable is altered (strengthened, reduced, delayed) by another variable, typically a sociodemographic variable such as gender, SES, age, or ethnicity (Flay et al., 1997). To accurately gauge the effect of a social influence, it is therefore important to include sociodemographic measures as control variables. This 26

PAGE 42

-----------applies an adjustment for the mediating influence of potentially skewed sociodemographics in the study population, and also furnishes valuable data concerning subsets of the study population that may be at higher risk than the general population. Given these concerns, my study includes a number of sociodemographic variables in its measures. Comprehensive understanding of tobacco use in Turkey requires multiple diverse investigations into various dimensions of the problem. This section presents a broad view of the literature addressing tobacco onset but focuses on the interpersonal or social influences and sociodemographic influences. The topics on adolescent smoking introduced are smoking prevalence, smoking uptake, addiction, gender, age, SES, academic achievement, peer influence, family smoking, and media influences. Smoking rates vary by gender as indicated by a 47% smoking rate for males and a 12% smoking rate for females (the worldwide prevalence is 29%) (World Bank, 2000). These rates differ by stage of national development; high-income countries have a 39% male smoking rate, versus a 49% male smoking rate for low and middle-income countries. For women, the difference is reversed; 22% of women in high-income countries smoke, as opposed to 9% in low and middle-income countries. In the United States, rates of smoking uptake declined for male adolescents from 45% in 197 4 to 33% in 1985 (Fiore et al., 1989) During the same period, rates of uptake for female adolescents remained unchanged at 34%. Young people have reasons for initiation and maintenance of smoking. In the literature related to initiation and maintenance of smoking the following factors are widely listed for beginning curiosity, social norms, and social pressure (Sarason et al., 1992); peer pressure (Barber et al., 1999; Brynin, 1999; Smith et al., 1999; West et al., 1999); and signs of independence and Western modernism (Headden, 1998; Zhang et al., 27

PAGE 43

2000). Relaxation, pleasure, addiction, and an effort to impress others (Sarason et al., 1992) have been proffered as reasons for the maintenance of smoking. As Denscombe (2001: 161) states: [T]here is a need to investigate the meaning which smoking holds for young people. There is a need to understand how they view smoking and what significance it holds in terms of their perceptions of themselves. And there is a need to link the findings to the social context within which young people find themselves at the start of the new millennium. His method focuses on the hermeneutic understanding of human action in relation to some wider whole that gives it meaning. Several studies discuss smoking in terms of adolescent needs for identity. And desirability and in the context of tobacco consumption. Nichter and Cartwright (1991: 430) nicely state this as follows: [R]ather than being regarded simply as objects to consume, cigarettes become indices of social membership for adolescents who are searching for their identities or who want to escape the immediate reality into which they have been born. The impact of consumption on the construction of identities is emerging as a key concern of social theory (Miles et al., 1998). Since cigarettes are one of many consumer products, teenagers might perceive some meaning in them that goes beyond their face value. Simmel's (1971) discussion about the role of fashion as a means of constructing a relationship between individual and social concerns supports this idea. Given that most tobacco addiction begins at a young age, interventions must be targeted at the youth in the developing nations (Kessler, 1995; Smith et al., 1999; Tilly, 1991 ). Nicotine addiction starts with tobacco consumption. Addiction is sustained by its psychopharmacological 28

PAGE 44

relation with dopamine and the psychological state of satisfaction, and is characterized by compulsive drug-seeking and use--even in the face of negative health consequences-and also by neurochemical and molecular changes in the brain. Tobacco use certainly fits both of these criteria (NIDA, 1999). Most smokers accept that tobacco is harmful (perceived severity of a disease) and express a desire to reduce or stop using it, and nearly 35 million of them in the US make a serious attempt to quit each year. Unfortunately, less than seven percent of those who try to quit on their own achieve more than one year of abstinence; most relapse within a few days of attempting to quit (1990 National Household Survey, in Heishman et al., 1997). Genes involved in the brain's dopaminergic reward mechanisms are one possible source for genetic factors that influence smoking practices. To date, the strongest evidence involves polymorphism (variation) in the dopamine transporter (SLC6A3) gene. This gene codes for the dopamine transporter protein (OAT) which transports released dopamine, thereby limiting the level and duration of dopamine receptor activation. Recent studies suggest that variation in SLC6A3 is associated with the likelihood of being a smoker, age of smoking onset, and length of periods of smoking abstinence (Lerman, 2000). The prevalence statistics above show that males are at greater risk of smoking than females (Malcon et al., 2003; Flay et al., 1999). Of all the sociodemographic variables, gender is a frequent moderating variable in other influences on smoking. Some studies show that gender moderates social influences such as population migration (Brunswick and Messeri, 1984 ), level of parent education (Ary et al., 1988), number of friends who smoke (Brunswick and Messeri, 1984), mother's smoking status on male childrens' smoking status (Skinner et al., 1985), parents' smoking status on 29

PAGE 45

female childrens' smoking status (Charlton and Blair, 1989; Chassin et al., 1986), mother's smoking status on male childrens' uptake (Skinner et al., 1985), sibling smoking status on female childrens' uptake (Mittelmark et al., 1987), friends' smoking status on female adolescents (Charlton and Blair, 1989). Although gender is covered in a separate section in the literature review, some gender-specific findings are reported in other sections, for example that females are susceptible to peer pressure starting at ages 14 or 15 (Barber et al, 1999) and that teachers intervene more often for male students (de Moor et al., 1992). Smoking has been linked to greater self confidence and social skills in girls (Clayton, 1991; Michell and Amos, 1996), but less self-confidence in boys (Clayton, 1991 ). Other studies find that: girls are more susceptible to parent and teachers' smoking influences than boys at ages 12-13 (Lloyd-Richardson et al., 2002); concerns about body weight leads to uptake in females (Gritz and Crane, 1991; Camp et al., 1993); boys initiate smoking earlier than girls (Lucas and Lloyd, 1999); and smoking is linked with female eating disorders (Crisp et al., 1999) and issues of power (Fisher, 1976). Smoking has been linked to other sociodemographics factors: increasing age (Alexander et al., 1983), low SES (Maziak, 2002), and pocket money (Alexander et al., 1983). It has also been linked to low academic ambition (Bauman et al., 1984), low grade point averages (Newcomb et al., 1989), and low academic achievement (Bailey and Hubbard, 1990; Choi et al., 1997). Social influences have been documented in relation to smoking. Among these have been peer pressure (Barber et al., 1999; Brynin, 1999; Smith et al., 1999; West et al., 1999) and social norms supporting tobacco use (Sarason et al., 1992). Smoking uptake has been linked to parents' smoking (Chassin et al., 1984; Jackson and Henriksen, 1997), although some studies have found that only the father's smoking was significant 30

PAGE 46

(Ahlgren et al., 1982) or only the mother's smoking (Skinner et al., 1985). Although Alexander et al. (1983) found that sibling smoking was a significant influence, West et al. (1999) found this effect in younger adolescents ages 15 and 16. The incidence of ever-trying has been associated with exposure to positive portrayals of smoking in film (Sargent et al., 2001 ). Gray et al. ( 1996) found that 12-13 year olds have very concrete and literal interpretations of images of cigarette use in magazines that lead to negative impressions of smoking, but 15-16 year olds' interpretations were more abstract, leading to either negative or neutral impressions of smoking. B. Social Cognitive Theory Social Cognitive Theory emphasizes observational learning and, role modeling, as well as the mutual influences of individual, behavior, and society. It provides a theoretical structure for parental modeling, differential peer association, positive or negative attitudes toward smoking, perceived positive and negative consequences of smoking, the integration of smoking into lifestyle, and the interaction of the social environment with individual characteristics. In sum, Social Cognitive Theory furnishes a large variety of modeling tools for studies involving social influences on individual behaviors such as smoking, providing a solid theoretical foundation for studies such as this. Social Learning Theory (Akers et al., 1977) and Social Cognitive Theory (Bandura, 1977) are two theories categorized under interpersonal models. Social Learning Theory (SL T) presents a commonly used perspective on smoking. Social Learning Theory can be categorized under socialization theories in sociology. Socialization is the process by which 31

PAGE 47

Iindividuals learn to become members of society, both by internalizing the norms and values of society, and also by learning to perform their social roles (Marshall, 1998). Early Behaviorist Origins of Social Learning Theory Early learning theories explain behavior is mechanistically, in which human behavior explained based on positive and negative reinforcements that elicit the desired behavior (Baranowski et al., 1997:154). B.F. Skinner's behaviorist theory is essential to the early definition of social learning model. He describes the conditions under which behavior is modified, not explaining the mechanism by which this process occurs. In attempting to apply Skinner's methods to smoking cessation, the proper reinforcements for the desired behaviors would have to be identified. Here addiction provides an interesting challenge to the behaviorist model, since addiction provides powerful motivations for continuing smoking that are not visible to the conditioner. After all, for the addict, the effects of addiction carry with it some perceived benefits such as relaxation, stress relief, and stimulation. This demonstrates that although the behaviorist approach may be useful in a limited sense, it does not furnish all the tools necessary for modifying behaviors under all circumstances such as nicotine addiction. Vicarious Learning On the other hand, Bandura and Walters (1963) asserted that rewards do not have to be directly applied to the individual for learning to occur. Instead the individual can learn from observing the behavior and the consequent reward of others. In the case of the perceived rewards of 32

PAGE 48

smoking in youth, social acceptance and social facilitation is a common experience of adolescents. Applying the Social learning theory to smoking, it includes parental modeling, differential peer association, positive or negative attitudes toward smoking, perceived positive and negative consequences of smoking, and integration of smoking into lifestyle (Spear & Akers, 1988). People also learn norms, attitudes, and orientations by interacting with their social groups. Social groups control individuals' major sources of reinforcement and punishment and expose them to behavioral models and normative definitions (Akers et al., 1979:838) Under SL T/SCT, learning may even occur without direct observation of others (Petraitis et al., 1995). Spoken endorsements by role models may be sufficient for youth to adopt positive beliefs about substances like tobacco. Steps in the Acquisition of an Abusive Habit Social Learning Theory identifies a three-step sequential process in the abuse of substances by youth, starting with influences external to the individual and ending with altered internal characteristics (Petraitis et al., 1995). Key to this process is the adoption of cognitive dispositions, called definitions in Social Learning Theory, about the substance. The first step is imitation, in which youths model their own behaviors or attitudes on other's behaviors by simply observing and replicating the behaviors or, transposing them into forms more acceptable to the youth lifestyle. The second step is social reinforcement in which adolescents internalize definitions and exhibit behaviors and values approved by significant others. In the final step, the individual's own expectation of physical and social rewards perpetuates the use of the substance. 33

PAGE 49

Social Cognitive Theory Evolves from Social Learning Theory Social Cognitive Theory (SCT) is an extension of social learning theory, and has its roots in reinforcement theory (Baranowski et al., 1997). It begins with the assumption that individuals are more likely to engage in behavior they find enjoyable or makes them feel good about themselves, or that they know significant others find enjoyable. It integrates concepts of self-efficacy, emotional coping responses, and cognitive world-modeling into an evolved form of Social Learning Theory (Petraitis et al., 1995). Bandura may have renamed his enhanced SL T as the Social Cognitive Theory in order to clearly distinguish SCT from its early behaviorist roots (Stone, 1998). Reciprocal Determinism Social Cognitive Theory explains human behavior by the reciprocal determinism among behavior, environmental determinants, and personal factors (Bandura, 1977, 1997). This is the central organizing principle of SCT: In social cognitive theory, human agency operates within an interdependent causal structure involving triadic reciprocal causation (Bandura, 1986a). In this transactional view of self and society, internal personal factors in the form of cognitive, affective, and biological events; behavior; and environmental events all operate as interacting determinants that influence one another bidrectionally (Bandura, 1997:5). SCT identifies three types of interaction under reciprocal determinism (Bandura, 1977). Person-behavior interactions specify that one's behavior is influenced by one's thoughts and feelings, and that one's behavior will 34

PAGE 50

conversely influence one's mental life. While the environment may influence one's behavior, one's behavior may also change the physical and social environment. Finally, one's personal characteristics, either mental characteristics such as intentions, beliefs, perceptions, and expectations, or physical characteristics such as gender and behavioral pre-dispositions (Stone, 1998), may be influenced by the environment. Likewise, the environment may change personal characteristics. These interactions among may be two-way or three-way, simultaneous or delayed. The behavioral, environmental, and personal phenomena modeled under SCT may therefore be very complex. The Constructs of Social Cognitive Theory. Bandura posits ten constructs within the SCT model: Environment, Situation, Observational Learning, Behavioral Capability, Reinforcements, Outcome Expectations, Outcome Expectancies, Self-Efficacy, Self-Control of Performance, and Emotional Coping Responses (Baranowski et al., 1997). The Environment construct is meant to represent those factors that are external to the individual, consisting of the physical environment and the social environment (Baranowski et al., 1997). Just as the Environment construct describes factors external to the individual, the Situation construct represents the individual's perception of elements in the environment. This not only includes an awareness of the objects and actions in the environment, but also the roles of those objects and even cues about the social acceptability of social behaviors (Baranowski et al., 1997). Since this construct involves perceptions, those mental representations of the environment may be mistaken or distorted (Baranowski et al., 1997). Unlike factors in the Situation construct, factors in the Environment construct can 35

PAGE 51

influence behavior without any awareness of that factor (Baranowski et al., 1997). Humans often learn new behaviors in a social setting by observing the behaviors of others and discerning the desirable/undesirable outcomes for those others. The advantage of this vicarious learning is that outcomes are more predicable, often involving less time and risk than learning based on direct experience (Baranowski et al., 1997). This learning activity falls under the Observational Learning SCT construct. Individuals engaged in observational learning form a model of the environment and the functioning of the learned behavior in the environment (Bandura, 1986). This is the basis for family members having similar behavior patterns (Baranowski et al., 1997). Not all observationally learned behaviors are effectively put into practice. The Behavior Capability construct specifies that behavior has two prerequisites: knowledge of the behavior and knowledge of how to perform the behavior (Baranowski et al., 1997). This construct reflects the intellectual capacity, training, and skill needed to engage successfully in the behavior. Social Learning Theory and Social Cognitive Theory suggest that the most effective way to prevent tobacco uptake is to decrease the appeal of tobacco advocates and users as role models, while simultaneously increasing the appeal of role models who abstain from tobacco use (Petraitis et al., 1995). In sum, role models are important factors in the initiation and maintenance of tobacco use. Individuals with friends who smoke are more likely to feel peer pressure. The Reinforcements construct consists of three kinds of reinforcement (Baranowski et al., 1997). Direct reinforcement is operant conditioning, having positive and negative rewards. Vicarious reinforcement is synonymous with observational learning. Self-reinforcement refers to self control. Reinforcements may be either external or internal under Social 36

PAGE 52

Cognitive Theory, where external reinforcements have predictable value as seen by an independent observer of an event, whereas internal reinforcements are based on the individual's own perception of the reinforcement value of the event. In the Outcome Expectations construct (originally called antecedent determinants), the individual projects an outcome of a behavior, anticipating a result (Baranowski et al., 1997). Smoking may be seen as pleasurable, for example, because the individual has learned to associate that outcome with the behavior. This is not to be confused with the Outcome Expectancies (originally incentives) construct, in which the outcome of a behavior is valued either positively or negatively. According to Baranowski et al. (1997), the timing of the outcome falls under this construct, where the actualization of a value may be immediate or delayed. This is particularly significant for smoking behavior, where immediate gratification is often chosen over long term health benefits (McAlister, 1980). Self-Efficacy is a construct closely associated with self-confidence that one can perform an activity (Baranowski et al., 1997). Self-confidence is an important aspect of behavioral change, since it influences the level of investment an individual makes (Bandura, 1977). It also plays a role in the refusal of youths to take up smoking (DeVries et al., 1990). Bandura (1991) specifies several components in the Self-Control construct: setting goals, monitoring of one's behavior, assessing its effects against goals, and self reward and self-reprimand for one's behavior. In the Emotional Coping Responses construct, excessive emotions such as anxiety and hostility are recognized as an inhibitor of useful learning (Baranowski et al., 1997). Several techniques for overcoming these responses have been identified. These include stress management, effective problem solving, and problem restructuring. These techniques have particular relevance for smoking 37

PAGE 53

cessation interventions, and work well with the Social Cognitive Theory (Baranowski et al., 1997). Social Cognitive Theory is complex, having disparate constructs addressing a broad range of determinants of behavior, ranging from external environment factors to learning paradigms, to internal mental dispositions and beliefs, to disciplines of self-regulation. The number of choices of construct combinations is large, and the choice of interactions adds yet another layer of complexity. For this reason, studies based on Social Cognitive Theory usually limit themselves to utilizing one or two constructs (Stone, 1993). Having completed the review of Social Cognitive Theory, this chapter continues with a review of the literature concerning the influences that are relevant to my study. These include several aspects of the social environment: peer, friend, parent, sibling and teacher influences, as well as school policy. Sociodemographic and personal factors are also reviewed: age and gender, academic achievement and urban/rural location. A review of the literature specific to Turkey follows. C. Family. Peer. Friend and Teacher Influences Role models have a powerful influence on youth's use of tobacco and other experimental substances, supporting both the Social Learning and Social Cognitive theories (Petraitis et al., 1995). The influence of parents and friends are issue specific in which parental influences are stronger for future roles; peer influences are stronger for issues related to immediate adolescent life style. Tobacco use is one of many adolescent behaviors that result from interaction between individual characteristics and the competing 38

PAGE 54

influences of multiple social groups (Kandel, 1980: 256). Peer and Friend Influences Peer influence has two aspects: First, the idea that peers have a distinct culture in which they use sanctions to enforce group norms Second, the peer group is comprised of individuals who share certain key attributes, such as age, gender, and ethnic origin (Denscombe, 2001 ). Peer pressure and appearance-consciousness play a role as well. In Brynin's (1999) survey of adolescents aged 11-15 in 5000 households in the US smoking is believed to make young people look more fashionable, older, and sexier. Kandel et al. (1978) found positive correlations between marijuana use and friendships where her subjects were exposed to drug, alcohol and tobacco use, were offered those substances, or were exposed to positive attitudes towards those substances. Huba et al. (1980) and Bailey & Hubbard (1990) found similar evidence in the case of friends who used these substances. Huba et al. (1980) also found evidence that marijuana use is correlated with friendships where these substances were offered for use. Flay et al. (1983) and Fisher et al. (1988) have found evidence contrary to the idea that role modeling of friends causes experimental substance use. Friends were found to select each other from their common interests and activities involving experimental substances, rather than those friendships causing substance use. This "birds of a feather" syndrome is a point of significant controversy among researchers. Kandel (1985) finds that both friendship selection and friendship influence are factors in substance abuse. Nevertheless, an overwhelming amount of research shows the significance of peer and friend influence in the case of youth smoking 39

PAGE 55

(Poulsen et al., 2002; Alexander et al., 2001; Wang et al., 1999; Barber et al., 1999; Rose et al., 1999; Smith and Stutts, 1999; West et al., 1999; Flay et al., 1998; Michell, 1997; Flay et al., 1993, 1994; Norman and Tedeschi, 1989). kl alternative view is that smoking is a response to immediate need, often psychological. Peer pressure provides additional forces in favor of the adoption of tobacco, driven by fear of rejection. Adolescents who are at the extremes in the popularity spectrum are more likely to be regular smokers (Spear and Akers, 1988). Adolescent cigarette use can be predicted from the influence of best friends and friendship groups. These influences include peer smoking, peers' tolerance of smoking, and approval of smoking among best friends (Kandel, 1980; Rose et al., 1999). Alexander et al. (2001) found that current smoking status was significantly associated with peer networks in which at least half of the members smoked, also finding significantly increased risk where best friends smoked. Barber et al. (1999) describe peer pressure as an important predictor of smoking for males in all age groups and for females starting from age 14-15. Furthermore, West et al. (1999) shows that uptake of smoking between ages 15-23 was associated significantly with friend's smoking behavior, and Flay et al. (1998) found that the percentages of experimental and regular smokers rises precipitously with the increasing number of friends who smoke. Denscombe (2001 ), in his recent study, discussed peer pressure from a different perspective. His subjects did not acknowledge the impact of peer pressure on their smoking initiation because this conflicted with their sense of personal autonomy, independence and self-determination. They might be heavily influenced by their friends but they might perceive their smoking as a free choice. On the other hand, Turkish youth might not have strong desires to express their personal autonomy since they grow up in a more traditional 40

PAGE 56

society in which family and friends are very influential and used as reference points for defining self in other areas (Guneri et al., 1999). Thus, 1 prefer to use the term peer influence instead of peer pressure for my study. Family Influences Some studies have found family to be a significant influence on youth smoking (Distefan et al., 1998; Farkas et al., 1999). Jackson and Henriksen (1997) found a significant risk of uptake when one or both parents smoke. Children of current smokers were found to be at greater risk of intending to smoke and perceived easy access to cigarettes. Jackson and Henriksen also found that children of former smokers were still at risk of modeling smoking after their parents. Flay et al. (1998) found that parents' smoking and family conflict were primary predictors of youth transitioning from experimental smoking to regular smoking, a finding confirmed in other studies (Distefan et al., 1998; Biglan et al., 1995). Similarly, Kandel's (1985) study revealed that parents play a central role in the transition from marijuana use to other substances. In another closely related finding, Fearnow et al. (1998) determined that parental actions to prevent their children's smoking lessened under environmental stress, while parental permissiveness was associated with their smoking status. Adolescents are more likely to smoke when their parents' verbal anti-smoking messages are diluted by their own smoking status (Henriksen and Jakson, 1998; Baer and Katkin, 1981 ). Other studies, however, have found that parental and family influence has a minimal or limited role in adolescent smoking when compared to peer and friend influences (West et al., 1999; Rowe et al., 1994; McNeill et al., 1988; Skinner et al., 1985). West et al. (1999) determined that sibling smoking had a significant 41

PAGE 57

effect on smoking uptake for subjects only between ages 15 and 16. According to Rowe et al. (1994), siblings usually do not have friends in common, and therefore have different peer influences in regards to smoking; the differences in friend selection may be attributable to differences in genetic makeup and differences in the way siblings experience their shared family and social environments (Rowe and Plomin, 1981 ). Teacher Influences Teachers have also been found to be an influence on youth smoking (Zhang et al., 2000; McNeill et al., 1988). In a Korean study, 35% of students indicated that they were affected by their teachers' smoking (Lee, 1997). Poulsen et al. (2002) found that Danish students subjected to teachers' smoking outdoors on school premises had a significantly increased risk of being daily smokers or heavy smokers. Poulsen speculates that teachers' smoking outdoors is related to an increased frequency of exposure on playgrounds, since student exposure to teachers' smoking indoors was found to be mostly in staff rooms, where the duration of exposure was presumed to be briefer and less frequent. Poulsen notes that school policies aimed at limiting second-hand smoke may be partially misguided, and that total bans on school premises may be necessary to prevent students' role modeling of teachers' smoking. Griesbach et al. (2002), however, found that total bans on teachers' smoking actually increased their outdoor smoking on school premises. School Policies The topic of school policies was introduced in the above review of the literature on teacher influences. Furthering this exposition, Griesbach et al. 42

PAGE 58

(2002) found that lower levels of smoking perceived by students was associated with the consistent enforcement of school policies against student smoking. This study also found that teachers who smoke were less likely to intervene to curtail student smoking, a finding echoed by de Moor et al. (1992). In a classic demonstration of reciprocal determinism, de Moor also found that teacher intentions to intervene were strongly associated with the smoking prevalence of male students, but not female students. An Australian (Clarke et al., 1994) study disputes these results, concluding that neither smoking policies nor staff smoking had a significant impact on student smoking. A Welsh study (Moore et al., 2001 ), however, contradicts the Clarke et al. conclusion regarding student anti-smoking policy, finding a strong correlation between the degree of enforcement of student smoking restrictions and the incidence of student smoking; no correlation was found between student smoking and enforcement of teacher smoking restrictions. D. Sociodemographic Factors Age and Gender Gender differences in smoking behavior arise early in adolescence, although in the US many of these differences disappear by the later teenage years. Across the globe, boys' age of initiation into smoking is significantly earlier than girls (Lucas & Lloyd, 1999). Girls are particularly more susceptible to peer pressure at the 14-15 age level (Barber et al., 1999), and Lloyd-Richardson et al. (2002) found that girls are more susceptible to parental and teachers' smoking influences at ages 12-13 than boys. Lloyd Richardson, however, finds virtually no difference in risk between US 43

PAGE 59

adolescents of both sexes in all categories of smoking. This outcome is very different from that of Turkey and most of the world, where male smoking rates are much greater than female smoking rates (Firat, 1996). Some studies show that gender moderates social influences such as population migration (Brunswick and Messeri, 1984 ), level of parent education (Ary et al., 1988), number of friends who smoke (Brunswick and Messeri, 1984), mother's smoking status on male childrens' smoking status (Skinner et al., 1985), parents' smoking status on female childrens' smoking status (Charlton and Blair, 1989; Chassin et al., 1986), mother's smoking status on male childrens' uptake (Skinner et al., 1985), sibling smoking status on female childrens' uptake (Mittelmark et al., 1987), and friends' smoking status on female adolescents (Charlton and Blair, 1989). Smoking has been linked to greater self-confidence and social skills in girls (Clayton, 1991; Mitchell and Amos, 1996), but less self-confidence in boys (Clayton, 1991 ). Two explanations are usually given for female smoking, as opposed to male smoking. For girls, smoking is often associated with concerns about body-weight and shapeliness (Gritz and Crane, 1991; Camp et al., 1993), a factor that is particularly pronounced in females with eating disorders (Crisp et AI., 1999). For women, this pattern remains throughout life. Fisher (1976) found that women's smoking is linked to a preoccupation with issues of power, an association frequently exploited by tobacco companies worldwide (WHO, 1996; Headden, 1998). Academic Achievement Although academic achievement is not purely a demographic variable, I categorized it under this section for organizational purposes. Low academic performance has also been associated with the use of tobacco 44

PAGE 60

and other experimental substances (Choi et al., 1997; Bailey & Hubbard, 1990; Conrad et al., 1992; Jessor et al., 1991; Kaplan et al. 1984; Kandel et al., 1978). Bryn in ( 1999) claims that there is an association between adolescent smoking and difficulties at home and school, but also with greater self-confidence and presumably greater interest in appearance. Nutbeam et al. (1993) found a strong link between adolescent smoking and their perception of their school, their academic capabilities, and their future academic plans. Additional data support that the onset of tobacco use is related to poor academic achievement: 21% of below-average students were heavy smokers compared to 7% of above-average students (Elder, 1994; Surgeon General's Report, 1994). Flay et al. (1998) found an increased risk of experimental smoking and ever-smoking among students with poorer grades, but found an increased risk of regular smoking among the top grade level ("A" students), versus "B" students, an unusual finding not discussed in his study. Lloyd-Richardson et al. (2002), however, confirms the Flay results, showing that students with grades in the top quartile are more likely to become regular smokers than their counterparts in the second quartile. Urban/Rural Location and Economics Various studies report contradictory findings on urban-rural differences, and economic level. Children in rural areas have been found at greater risk of smoking uptake than their urban counterparts (Harrell et al., 1998; O'Connell et al., 1981), although this is disputed by Stanton et al. (1994). Some studies also have found an association between spending money and tobacco use in adolescents (Rissel et al., 2000; Alexander et al., 1983), although lower socioeconomic status has also been associated with 45

PAGE 61

youth smoking (Harrell et al. 1998; Stanton et al., 1994; Conrad et al., 1992). E. Social. Cultural and Psychological Factors in Turkish Tobacco Use Although Turkish youth obviously share numerous psychological characteristics with their Western counterparts, the mediation of those characteristics by society and culture can produce surprising differences and surprising similarities in smoking-related behavior. In his recent study Denscombe (2001) discussed the reluctance of his subjects to acknowledge the impact of peer pressure on their smoking initiation because this conflicted with their sense of personal autonomy, independence and self determination. Guneri et al. (1999) stressed that Turkish youth might not have strong desires to express their personal autonomy since they grow up in a more traditional society in which family and friends are very influential and used as reference points for defining self in other areas. Erbaydar et al. (2002) found that students who work are exposed to smoking workplaces during their working internships, and earn pocket money that gives them more opportunity than their non-vocational peers to buy cigarettes. Ozcan and Ozcan (2002) noted that most of the vocational school student population comes from lower-class families. Some of the Turkish patterns are replicated in other Middle Eastern countries. Maziak (2002) concluded that smoking is associated with poor academic performance and low SES in Syria, whereas Jarallah et al. (1999) found smoking to be associated with a lack of education in Saudi Arabia. Many of the factors influencing smoking in the developed world appear to be universal. Cunningham (1996) argues that the same tools 46

PAGE 62

which exploit adolescent needs in the US work effectively across the globe, operating on common principles of human nature everywhere. These principles are sometimes called the "5 S's": sophistication, slimness, social acceptability, sexual attractiveness, and status. Similarly, Headden (1998) states that feminine independence is a draw for women in developing countries. In Western cultures, adolescents frequently report curiosity, social norms, and social pressure as reasons for smoking uptake (Sarason et al., 1992). Maziak & Mzayek (2000b) found that peer influence is the greatest influence on smoking uptake among high school students in Syria, although parent and sibling influences are also significant factors. In Indonesia, the smoking status of best friends is the greatest determinant of smoking (Smet et al., 1999). In China, tobacco use is associated with peers, teachers, and mothers who smoke (Zhang et al., 2000). The data that do exist indicate that the development of the smoking habit is delayed, at least in Syria among the Middle Eastern countries, until adult supervision is relaxed in the post-secondary school years of young adults, corresponding to increases in smoking rates after age 18 (Maziak & Mzayek. 2000). Rissel et al. (2000) had similar findings for Vietnamese, Asian, and Arabic youth living in Sydney, Australia. Rissel cites closer family ties, a stronger commitment to education, and lack of pocket money as reasons for the delay in smoking onset. The brief review above leaves several factors unexplained in terms of why and how Turkish youth pick up and maintain the smoking habit. In Turkey, the adult smoking rates are 63% for men, and 25% for women yet adolescent smoking rates (ages between 15-19) have been found to be (only) 30% in Turkey. Taking the average of these adult rates, this indicates that the adult smoking rate is approximately 44%. This is highly suggestive that many Turkish youth are becoming regular smokers as young adults. 47

PAGE 63

Since individuals in the US rarely take up smoking after age nineteen (Tilly, 1991; Kessler, 1995; Corbett, 2001 ), clearly some factors are at work in Turkey that are not evident in the United States. F. Application of Social Cognitive Theory: Study Hypotheses This study's primary concern is with the social environment's influence on smoking status. Sociodemographic variables are also useful and necessary as control variables when modeling the influences of the social environment. Social Cognitive Theory emphasizes observational learning, role modeling, as well as the mutual influences of individual, her behavior, and society. The literature review provides a theoretical structure for parental modeling, differential peer association, positive or negative attitudes toward smoking, perceived positive and negative consequences of smoking, the integration of smoking into lifestyle, and the interaction of the social environment with individual characteristics. In sum, Social Cognitive Theory furnishes a large variety of modeling tools for studies involving social influences on individual behaviors such as smoking, providing a solid theoretical foundation for studies such as this. Three types of information were specified under Social Cognitive Theory: person-environment; person-behavior; behavior-environment. The first two interactions are the main focus of the study but the third interaction will be also included in the discussions where it is needed. This study incorporates constructs that are directly applicable to the social environment and socio-demographics, namely the Environment and Situation constructs. The observational learning construct is categorized under environment. Each hypothesis and measure in this study is 48

PAGE 64

identifiable with these three constructs (or with sociodemographics). As is typical for studies using Social Cognitive Theory, only one or two constructs are necessary to study the interactions of interest (Stone, 1993). These constructs and the potential influences examined are illustrated in Table 2.1. Table 2.1 Selected Factors Influencing Tobacco Use Previously Identified Risk Factors Social Cognitive Theory Constructs SOCIODEMOGRAPHIC FACTORS Environment and Person 0 School type 0 Male gender 0 Age 0 Origin (urban/rural) 0 Allowances (SES) 0 Economic situation (SES) 0 Academic Achievement SOCIAL ENVIRONMENT Environment FACTORS Interpersonal factors 0 Parent smoking 0 Sibling smoking Modeling 0 Relative smoking 0 Teacher Smoking 0 Friend smoking Perceived Environmental Factors Situation 0 Close friends' approval of smoking (social support) 0 Perceived prevalence Situation among peers (norms) 49

PAGE 65

------,,----------------------------------Hypotheses Various sources suggested potential hypotheses for the study. The existing literature indicates that family, friend, peer, teacher and school influences are globally at work on the youth of many cultures. Likewise, the influences of socio-demographic factors are documented across the globe: gender, age, place of origin, economic circumstances, and academic success. These, together with the preliminary focus groups, interviews and surveys, as well as the researcher's own experience as a citizen of Turkey, informed the creation and elimination of the potential hypotheses. Each hypothesis furnishes an opportunity to ask a series of questions in the survey that will reveal the mechanisms of smoking influences in greater detail. The hypotheses of the study are presented below, organized by major theoretical construct and topical area of influence. Sociodemographic Factors School Type. In the preliminary interviews, many perceived the industrial high school as "the smoking school." In Turkey, high school students are targeted by levels of academic achievement into different types of high schools: regular (academic), specialized vocational, and religious schools. Ozcan and Ozcan (2002) found a greater risk of smoking for students of some types of vocational schools in Ankara. In addition, testing a hypothesis related to type of school introduces a potentially significant control variable. This leads to the hypothesis: (H1 ): Adolescents smoking differs by school type. Gender. The literature strongly supports gender as a significant factor in smoking uptake and maintenance. The preliminary survey indicated that 50

PAGE 66

male adolescents in Dursunbey are at much greater risk of smoking than girls. In addition, gender is an important factor that will serve as a control when analyzing other factors related to smoking. A new hypothesis is added: (H2): Adolescent boys will be more likely smokers than girls. Age. The literature supports age as a significant factor in smoking status. The preliminary survey supported this association. Like gender, age is an important factor that will serve as a control when analyzing other factors related to smoking. The hypothesis is: (H3): The older the adolescent, the higher the probability of being a smoker. Place of Origin. Dursunbey is a mixed urban/rural community, making it an excellent area for studying potential differences between urban and rural smoking patterns. Indeed, this opportunity was one of the motivations in selecting of Dursunbey for this study. Like gender and age, place of origin also serves as an important control when analyzing other factors related to smoking. The hypothesis: (H4): Adolescent smoking differs by place of origin (urban/rural). Weekly Allowance. The literature reviewed indicates that there is a significant relationship between smoking status and spending money. The hypothesis is included: (HS): Adolescent weekly allowances will be correlated with smoking status. Economic Situation. Turkey has been undergoing severe economic 51

PAGE 67

stress over the last several years. With an unemployment rate of 10%, a 40% annual inflation rate, and currency devaluations, families in Turkey are under considerable economic stress. In Dursunbey itself, the slumping forest and coal industries are adding to the problems in the local economy. The preliminary survey indicated that neither family income nor weekly allowance was related to adolescent smoking. However, the general economic conditions in Dursunbey prompts one to check on youth's perception of their economic situation in relation to smoking. A hypothesis relating to adolescent's perception of their economic situation is added as insurance against missing a potentially important factor in youth smoking: (H6): Adolescents' perception of their economic status will be associated with their smoking status. Academic Success. There is a substantial body of evidence in the literature linking an increased risk of smoking uptake and experimental smoking to low academic achievement. To explore this possibility, a new hypothesis related to academic achievement is added: (H7): Adolescents perception of their academic achievement will be associated with their smoking status. Social Environment Factors Family Influence. Families are also an important part of one's social environment from which one learns through observation. The literature is mixed in its findings about the influence of family on one's smoking status, much of it finding that family is a lesser influence than friends and peers. Nevertheless, some studies suggest adolescents model the smoking behavior of their parents, and that children of smoking parents are more likely to shift from experimental smoking to regular smoking. This suggests 52

PAGE 68

that this topic is worth investigating in Dursunbey, where the relative strength of parental influence may differ somewhat from studies performed in the West. The literature on siblings is also mixed, indicating a vulnerability to the influence of siblings between ages 15 and 16. This is, however, within the age range in my study's target population, which is a prime age for smoking uptake. Interviews from the preliminary studies indicate that relatives exert some role modeling influence For these reasons three hypotheses related to family are tested in this study: (H8): Adolescents whose parents smoke are more likely to smoke themselves. (H9): Adolescents whose siblings smoke are more likely to smoke themselves (H10): Adolescents whose relatives smoke are more likely to smoke themselves. Teacher Influence. The reviewed literature from countries as dissimilar as Denmark and China show that teachers are a significant influence on adolescent smoking status. The role modeling influence of teachers was strongly suggested in the preliminary interviews and focus groups. Teachers have traditionally been revered in Turkey, although the interviews indicated that this special regard might be waning. Earlier Turkish studies have shown that 44% of teachers are smokers, potentially leading to the frequent exposure of their students to their smoking. It is hypothesized that: (H11 ): Adolescents who are exposed to teacher smoking are more likely to smoke. Friend Influence. The reviewed literature documents the influence of smoking friends on one's smoking status, not only in Western cultures, but 53

PAGE 69

also in countries as diverse as Syria, Indonesia, and China. Friends usually play an important part in social environment of individuals, and from the researcher's personal experience this may be particularly true of Turkey. It is therefore reasonable to suspect under SCT that observational learning and role modeling takes place among friends, and with this an association between friends' smoking status and one's own smoking status. The preliminary focus groups and interviews also indicate that smoking in Turkey often receives approval from adolescent friends This leads to two hypotheses: (H12): Adolescents whose close friends smoke are more likely to smoke themselves. (H13): Adolescents who receive positive reactions about their smoking from their close friends are more likely to smoke Perceived Smoking Prevalence Among Peers. The literature strongly supports the link between peer smoking and the risk of smoking. To the adolescent, the perception of smoking prevalence among peers involves a judgment on the sum of his experiences, which may involve both objective and subjective factors. Subjective perceptions of prevalence, however, could well color a youth's judgment about whether or not smoking is a normal and acceptable activity. A new hypothesis is added: (H14): Adolescents who perceive higher smoking prevalence among peers are more likely to smoke themselves. 54

PAGE 70

CHAPTER 3 METHODS A. Study Design A multi-method research design was employed for this study, although the primary method was a survey. Focus groups, interviews, and observations allowed insight into the context of tobacco use by youth in Turkey, and contributed to the development of the survey. These qualitative methods aided in forming the final set of hypotheses and creating understandable, culturally appropriate, and comprehensive items for use in the cross-sectional survey of all 1 01 h graders in the five high schools of Dursunbey. The initial work was useful in planning data collection procedures, eliminating potential problems before the actual survey instrument was administered. Data gathering took place over two seasons of fieldwork in Turkey (see Figure 3.1 ). Preliminary work in the field season one took place during June and July of 2001. After passive informed consent was obtained, a preliminary survey of 247 students in Dursunbey and lzmir aided the development of the field season two's final survey, supplemented by focus groups and semi-structured interviews of 32 students in lzmir. Lessons from those research exercises (see field season one study below) benefited the construction of the survey instrument used in the second field season. 55

PAGE 71

Figure 3.1. Timelines for the Data Collection and Write-Up June-July 2001 August 2001March, 2002 April 2-11 2002 April 8-12, 2002 April 15-26, 2002 April 12-May 25,2002 June 2002September 2003 Field season one study completed in Dursunbey and lzmir, Turkey (247 surveys, 32 in-depth semi-structured interviews and 3 focus group) Data entry and preliminary analysis of field season one study. Designed the final questionnaire, translated questionnaire into Turkish. Finalized the dissertation proposal. Obtained permission from Education Ministry; shared the questionnaire with school counselors, modified the questionnaire after counselors' feedback. Talked with relatives and friends about adolescents' and adult smoking. Pilot tested the survey; scheduled survey administration for each school; distributed passive parental consent forms. Interviewed school principals. Administered the survey to 1Oth grade students in 5 schools in Dursunbey (N=211 ). Recruited students for semi-structured interviews; began encoding the survey results into SPSS statistical software; reviewed survey data; refined semi structured interview guide. In-depth semi-structured interviews with students and interviews with school counselors, teachers, and principals (N=70). In addition, three focus groups with students. Transcribed interviews; analyzed interviews using Microsoft Word; translated the supporting quotes. Completed inputting survey results; analyzed the survey data; wrote up survey results. The second field season's data collection in April and May 2002. The purpose of the final survey was to collect baseline data regarding the 1oth graders' smoking behavior. Like the first season, the final survey was accompanied by focus groups and semi-structured interviews. This investigation allowed me to identify social factors and smoking patterns 56

PAGE 72

among youth. The 57-item questionnaire was designed to obtain data on demographic variables, exposure to smokers in their family environment, in peer and friend groups, and in school environment. Demographic variables of school type, age, gender, family income, perceived economic situation, weekly allowances, place of origin, family, and friend influences constitute the independent variables. The dependent variable is the individual's smoking status categorized according to respondents' self-identification using WHO's question of smoking in the past thirty days. All interviews, focus groups and surveys were conducted and analyzed in Turkish. Some survey questions that did not translate well were discussed with Turkish speaking people from the local community and school counsellors to develop appropriate vocabulary. Passive parental consent was again obtained, followed by the administration of the questionnaire in all five of Dursunbey's high schools, consisting of industrial vocational, girl's vocational, religious, and regular schools. This cross-sectional survey included all 1oth graders in the town of Dursunbey (n=211 ). The study inclusion criteria were that students must be enrolled in a high school, be enrolled in 1Oth grade, attend school on the day(s) of survey administration, and consented to the study. Of the 57 survey questions, sixteen questions were developed based on input from previous focus groups. The focus group study also aided in the selection of forty-six questions drawn from various previously validated surveys. I conducted the qualitative portion of study using semi-structured in depth interviews with a grant received from Research for International Tobacco Control, International Development Research Center in Ottawa, Canada. A convenience sample of the survey population (n=42) was selected, together with students from other grades (n=1 0). These interviews helped me to interpret the survey results and are incorporated into the 57

PAGE 73

Dursunbey the following year. The qualitative part of the study was used to fill the gaps identified in the quantitative part. The initial survey was designed to be exploratory rather than hypothesis-testing. Its 62 questions asked about individual attributes such as age, gender, family income, monetary allowances, knowledge, attitudes, behavior, stress, decision-making and family, friend influences, social networks and norms, as independent variables. An individual's smoking status was categorized based on the respondent's self-identification, according to the World Health Organization's classifications of daily smoker, occasional smoker, ex-smokers and nonsmokers (WHO, 1997). The qualitative components of the Field Season One allowed the researcher to explore meanings, definitions, and other cultural aspects of tobacco use by Turkish youth in diverse settings. The semi-structured interviews allowed the teenage subjects to describe their smoking behavior in their own words, giving the researcher another view of social contexts in which smoking behaviors occurs. The combination of survey and semi structured interviews provided two perspectives for understanding youth knowledge, attitudes, and behaviors about smoking uptake and maintenance of smoking. Construction of the Survey for the Field Season One The survey questionnaire consisted of multiple choice and open ended questions. The survey questionnaire (See Appendix A) included fifteen open-ended questions to allow teenagers to express their ideas in their own categorizations, rather than the researcher's categorization. Forty six questions were drawn from the WHO Global Youth Tobacco Survey, the COMMIT Youth Survey (1993), the AMC Arresting Smoking Uptake Using 59

PAGE 74

Interactive Multimedia (2000) survey, and the Marmara University (MARUN) survey (2001 ). Table 3.1 shows the structure of the questionnaire that was used during the field season one. Table 3.1. Structure of the Questionnaire Social and symbolic influences Locus of control Self -efficacy Motivations/reasons for smoking or not smoking Intentions to smoke Access Cessation Perceived health dangers of smoking -Perceived list of diseases linked to smoking -Beliefs and expectations about future prospects and health -Pro-tobacco messages -Anti-tobacco messa es -Brand preferences -Parents and teachers who smoke -Perceived prevalence of smoking -Perceived parental approval of smoking -How many of their friends smoke -Offers of ci arettes from friends -Relationships with family -Do you think you can give up smoking? -Do you think not smoking is a way to express your independence? -Self-reported refusals of offers of cigarettes -Performance in school -"Who encouraged you to try your first cigarette?" -'Who were you with when you tried your first cigarette and where were you?" -Wh do ou smoke now? -Do you think you will smoke a cigarette anytime during the next ear? -How do they obtain their cigarettes -Have they ever urchased single ci arettes -Do you want to completely stop smoking? -How man times have ou tired to uit smoking? In addition, during the survey questionnaire development, students were given small pieces of paper and asked to list the reasons why they started to smoke or why their peers started to smoke in four high schools in 60

PAGE 75

Dursunbey from grade 9 to 11. Focus Groups Focus groups were conducted to orient the researcher to current Turkish adolescent attitudes and beliefs about smoking, as well as their smoking/nonsmoking behaviors. During the summer of 2001, a focus group was carried out in Dursunbey, consisting of 7 girls and 2 boys from low to middle income levels, with ages between 15 and 17 (10th and 11th graders) One individual was a nonsmoker and the rest were smokers. In addition, two focus groups were conducted in the urban city of lzmir. These focus groups enabled the researcher to explore smoking issues from the perspective of the teenager's own meaning and experiences. Among the questions asked were: When and how did they start smoking? Why do they smoke now? In which environment do they smoke the most? What are their reasons for smoking and what role do their peers play in the decision to smoke? What are their reasons not to smoke? The focus group study also aided in the selection of 46 questions drawn from the WHO Global Youth Tobacco Survey. Semi-Structured Interviews During the Field Season One 32 students were interviewed in lzmir, Turkey. Students who were interviewed were asked to tell their stories: e.g., how did they start smoking, and when and where did they try their first cigarette? Why do they smoke now? Or for nonsmokers, why do they think young people smoke? What does smoking mean to them in their social environment? In addition, questions were asked to identify the peer pressure and family influences in smoking (see Appendix A) 61

PAGE 76

Field Season One Data Collection and Sampling Convenient sampling procedures were used in Dursunbey and lzmir. After an initial focus group in Dursunbey provided insights into adolescent attitudes about smoking, a questionnaire was given to 247 high school students of mixed grade levels in lzmir and Dursunbey. This was followed by 32 semi-structured interviews given to subsample of the survey subjects in lzmir. The study was completed with two more focus groups in lzmir, the third largest city in Turkey. A total of 17 4 students from all five schools were surveyed in Dursunbey: the religious high school (n=37), the health vocational school (n=21 ), the regular high school (n=43), and the industrial vocational high school (n=70). In four high schools, all the 1Oth graders who were available during the administration of the survey were surveyed. In addition, one class of glh graders from the industrial vocational high school class, one class of glh graders from the regular high school, and one class of glh and 11 lh graders from Religious high school were surveyed. For the survey administered in lzmir, students (n=43) from two 1Oth grade classes of the lzmir Karsiyaka Regular High School, and 32 students from various EVKA-4 high schools were surveyed. The administration procedures were identical to those followed in Dursunbey. All the semi-structured interviews were conducted in lzmir. The set of subjects for the interviews was a subset of the subjects who completed the Field Season One survey, where the parental consent forms covered the contingencies for both methods of data collection. Students participating in the survey were chosen by convenience sampling for participation in the interviews. Thirty-two teenagers from various grades (9th, 1 olh, 11th) were interviewed, with the majority of them 1Oth graders. These teenagers were 62

PAGE 77

from the EVKA-4 Regular High School (n=16), the Industrial Vocational High School (n=6), the Karsiyaka Regular High School (n=2), and high school dropouts (n=B). Smokers were represented by higher numbers t han nonsmokers in the semi-structured interviews. Female and male segments of the sample were at rough parity ( 14 girls and 18 boys). Analysis The open-ended survey questions were coded first, then the questionnaire results were entered into a SPSS database. Chi-square tests were used to look for associations between smoking status and predictor variables. Semi-structured interviews were analyzed by using content analysis. Descriptive data regarding key factors and conditions influencing smoking attitudes and behaviors of Turkish youth were developed through in-depth, systematic examination and interpretation of the interview data and identification of key patterns Field Season One Study Results Study results showed high smoking prevalence in the study population. Seventy-four percent (N=246) of the subjects ever tr i ed smoking, with 60% of girls and 82% of boys Twenty-one percent of the students identified themselves as occasional smokers, and 24% as regular smokers. Boys were 3.1 times (95% Cl, .18, .58; p< .000) more likely than girls to have ever tried smoking, and boys were 3 14 (95% Cl, .19, .54; p< .000) times more likely to be a current smoker than girls. Boys were also 2.1 times (95% Cl, 1.05, 4.1) more likely than girls to be regular smokers and 1.2 times more likely to be occasional smokers (95% Cl, .44, 3.4; p<.OOO). 63

PAGE 78

Family influences were mentioned during the Field Season One interviews, but in the survey, father and sibling smoking were not significant, although the mother's smoking was. The number of close friends who smoke was significantly associated with adolescent smoking status, both in the field season one survey results and interviews. Grades were significantly correlated with smoking behavior. Household income was not significantly correlated with smoking behavior, and neither was weekly allowance. From the interviews, it was clear that the subjects thought that smoking was relatively benign, and less serious than drinking, gambling, or drugs. Most interviewees knew about the harmful effects of smoking, but were undeterred by this knowledge. Smoking was a circumstantial and unplanned activity. They observed that others find that "smoking is enjoyable," "it helps to forget worries," or is even an activity associated with celebration. Often the decision to try smoking is spontaneous, the reaction to an upsetting event where friends offer them a cigarette, saying, 'Take one. It will help you relax." Often in focus groups and interviews, students described practicing smoking with their friends. They practiced inhaling in groups, some believing that "if you don't inhale, you are a fake and are not a real smoker." Some believe that the smoke must be inhaled to be harmful; others believe that it is harmful if not inhaled, since this might result in mouth cancer. These are some of the myths about tobacco use found in the qualitative portion of my field season one study (Appendix B, includes more detailed information on preliminary results from quantitative and qualitative fieldwork, Summer 2001 ). PJthough the field season one study offered a good background for understanding youth smoking behavior in Turkey, It did not include a 64

PAGE 79

random sample of the students within these schools and I wanted to get more detailed information about the interpersonal influences on youth smoking. As a result I conducted another survey during the second field season ("the survey that is the subject of this dissertation"). C. Field Season Two in Dursunbey. 2002 Field season two has two components survey and semi-structured interviews. Both the survey (N=211) and semi-structured interviews (N=70) were conducted in Dursunbey, Turkey. The Survey in Dursunbey, 2002 Study Sample. This is a cross-sectional survey of all 1Oth graders in five high school of Dursunbey, Turkey (N=211 ). Criteria for inclusion in the study included: being enrolled in the 1Oth grade in a Dursunbey high school, school attendance on the day(s) of survey administration, and informed consent. The researcher returned to two high schools where 4 students were absent on the initial day of administration. The mean age for the 1Oth grade sample population was 16 years, and is significant because prior studies indicate that 39% of Turkish smokers start uptake in an age range of 15 to 17 years, and 20% of them between 11 to 14 years (Asut, 1993). The detailed information about the study area and population can be found in Chapter three. Data Collection Procedures. IRB approval from the University of Colorado at Denver was obtained in June 2001. The same approval was extended for another year in 2002. In addition, permission to conduct the 65

PAGE 80

survey was granted by the local bureau representative of the Ministry of Education in Dursunbey. The principal of each school was then consulted about the schedule for administering the survey. In April 2002, permission for a high school tobacco study was obtained from the Dursunbey office of the Ministry of Education. Informed consent forms were distributed to the classes in the weeks prior to the administration of the survey instrument in each school. Parent(s) were to sign if they refused their children's participation. Since Turkish parents place a high degree of trust in the authority and integrity of school administrations, parental permission was obtained without exception. After passive parental consent was obtained, the questionnaire was given in all five of Dursunbey's high schools, consisting of industrial vocational, girl's vocational, health vocational, religious, and regular schools, in April and May of 2002. The first two weeks of the data collection involved meeting school counselors and principals, refining the survey questionnaire. One of the school counselors was consulted to finalize the questionnaire. The survey items were discussed with one of the Dursunbey school counselors in terms of clarity and relevance of the questions, in order to create a student friendly questionnaire. This resulted in several recommended changes. After the counselor's suggestions were incorporated into the questionnaire, a pilot test of the questionnaire was conducted, first administered in a face-to-face interview to two students (one girl and one boy). The researcher solicited feedback from the students about their understanding of the survey questions and their interpretation of the concepts and phrasing used, as well as the criteria the students used in selecting their responses. This test uncovered a number of difficulties in the questionnaire. The students had problems, for example, when some of the skip patterns were not clear to them, or when the questions were formatted using the Likert-scale in table 66

PAGE 81

form. Using this feedback, the researcher reduced the size of the tables, dropped some of the question tables from the questionnaire, and eliminated some of the survey questions. As a next step developing the Field Season Two survey, the refined survey was administered to one class of twenty-five gth graders. During this phase, it became clear that the students were not carefully reading the instructions, which were now seen as extensive in length. This meshed well with an observation by one of the school counselors, that many of these students are not accustomed to taking surveys. Several refinements in the instructions and administration procedures were made; the researcher shortened instructions and prepared a fifteen minute detailed verbal explanation to preface administration of the survey. The researcher planned to be present while students were taking the survey, so that the students could seek clarification about the survey questions. This exercise also helped to estimate the actual administration time, which in practice consumed the entirety of a 45-minute class period. The questionnaire was administered by the researcher during class hours designated by the teacher and school administration. Parental Informed Consent and Student Assent Form. For the survey, students in selected schools were given a consent form for the parent(s) to sign if they chose to refuse their students' participation. In the student assent form, students were told that the participation was strictly voluntary and they could stop any time they wanted. Students were told that the information they were to give would be used to develop better education programs for young people like themselves. Students were assured that the questions that asked about their background would only be used to describe the types of students completing this survey. The information would not be used to find out their name, and no names were ever to be reported. 67

PAGE 82

Field Season Two Study Instrument. The questionnaire was designed and modified with the knowledge gained from field season one study and drawing from various validated questionnaires from prior studies. The final survey questionnaire had 57 -items (See Appendix A), including 54 multiple choice questions and 3 open-ended questions. Eight questions on the questionnaire were created with knowledge gained from the prior focus group study. Knowledge gained in the field season one was invaluable in improving the survey instrument and strengthening the study as a whole. The number of answers to some multiple-choice questions was reduced to cover potential low-frequency responses. Most importantly, it became clear that the Field Season One survey was too long, ranging in topics from demographics, to locus of control, social influences, health awareness, and perceptions of tobacco companies. Field Season One study helped to cut down the length of the survey to focus on socio-demographics and social influences, so that virtually all students could complete it in one class period. During the focus group, the researcher realized that teenagers' definitions of "a smoker" differed from the definition in tobacco control literature. The latter defines a smoker as someone who smokes one or more times within thirty days prior to the survey. Teenagers in the focus group did not consider themselves smokers if they only smoke occasionally, and if they do not carry a pack or inhale. Inhaling cigarette smoke was an important issue for teenagers because most of them indicated that they initially practiced inhaling with the assistance of their smoker friends. In addition, the researcher realized that brand issues surfaced frequently during the field season one. This led to the creation of eight survey questions about the reasons for brand preferences. At the end of the 68

PAGE 83

preliminary study, three of those questions were dropped to streamline the survey. Various other questions were deleted from the survey Questions about locus of control were removed as the study focused on social influences. Similarly, questions were dropped concerning intra-personal variables such as beliefs, rebelliousness and risk-taking behavior. It became clear in interviews that virtually no negative publicity about tobacco companies was reaching these youth; I removed the question, "do you think cigarette companies target teens to replace smokers who die?" Some questions were added to the survey In the field season one survey, many students were unsure about their family income, but in the interviews they certainly knew their weekly allowance. After interviewing a school counselor, the income levels in the family income question were adjusted and a question about weekly allowance was kept. The interviews revealed that adolescents were buying single cigarettes, rather than just full packs. As a result, a question was added about single-cigarette sales. During the data analysis, I will mainly focus on the variables related to social environment, interpersonal influences, and sociodemographic influences. Translation of the Questionnaire. The construction of the final survey began in the fall of 2001. The questionnaire blended 49 of questions adopted from prior studies conducted in Colorado, Turkey, and worldwide with 8 new questions from the focus group and interview results. The researcher then translated the questionnaire into Turkish; the translation was reviewed for accuracy and clarity by Candan Duran-Aydintug, Ph.D., a University of Colorado at Denver sociologist of Turkish national origin. 69

PAGE 84

Measures In this section study measures dependent and independent variables will be described in details. Dependent Variable. The World Health Organization (WHO, 1998:76) provides precise definitions for smoking status, categorized as daily smoker, occasional smoker, and non-smoker. Each category is discussed below. A Smoker is someone who, at the time of the survey, smokes any tobacco products either daily or occasionally, as defined under those two composing categories In this study there was a total of 88 students (42% out of 211) who met this definition. Smokers may be further divided into two categories of daily smoker and occasional smoker. A daily smoker is someone who smokes any tobacco product at least once a day. In this study 54 {61 %) students out of 88 were daily smokers. An occasional smoker is someone who smokes, but not everyday. Thirtyfour (39%) people out of 88 smokers were occasional smokers. Occasional smokers can be further divided into two groups: continuing occasional and experimenter. Continuing occasional smokers are people who have never smoked daily, but who have smoked 100 or more cigarettes {or the equivalent amount of tobacco) and now smoke occasionally. In this study a total of 13 students met this definition. Experimenters are people who have smoked less than 100 cigarettes (or the equivalent amount of tobacco) and now smokes occasionally. In this study 21 students met this definition (Table 3.2 presents study results based on these definitions). 70

PAGE 85

Table 3.2. Definitions of Adolescent Smoking (N=211) Smoking Status ever has has smoked o n smoked, smoked smoked 20 or more even a smoked not a 1-99 has smoked days in puff? in last 30 whole 100 or more last 30 days never Smoked lifetime use when current maybe maybe (/) puffs Qj yes no no 0 level experiments yes maybe maybe of use established yes maybe no frequent yes yes no n o A nonsmoker is someone who, at time of the survey, does not smoke at all. In this study123 (58.3%) students indicated that they d i d not smoke in the past 30 days. Non-smokers may be further divided into two categories: Never-smokers and ex-occasional smokers. Never smokers are those who either have never smoked at all (n=35) or have never been daily smokers and smoked less than 100 cigarettes (or the equivalent amount of tobacco) in their lifetime (n=78). Ex-occasional smokers are those who were formerly occasional, but never daily, smokers and who smoked 100 or more cigarettes (or the equivalent amount) in their lifetime (n=1 0). Smoking status as an outcome variable was measured by the question (017): "During the past 30 days, on how many days did you smoke cigarettes?" This question was answered by seven categories: (1 =0 days; 2=1 to 2 days; 3=3 to 5 days; 4=6 to 9 days; 5= 1 0 to 19 days; 6=20 to 29 days; ?=All 30 days). To form a dichotomous dependent variable for logistic 71

PAGE 86

regression analysis, subjects answering anything but "0 days" were classified as smokers. Independent Variables Table 3.3 and Table 3.4 present the independent (explanatory) variables and their descriptions. Table 3.3. Explanatory (Independent) Sociodemographic Variables :'>: ,, .;, ,; r Q!:Js(;;tipti(:m .. .';,::, : aa e < . . mg .DS coo A Demographic Factors ....... . School type The type of school 1 =religious high school students enrolled 2=health vocational school 3=regular high school 4=girls vocational school 5=industrial vocational school Gender What is your gender? 1=male O=female Age What is your age? 15 to19 Family origin Where did you originally O=urban come from? 1=rural Perceived economic How do you perceive 1=very poor situation your economic status? 2=poor 2=average 3=good 4=very good Weekly Allowances (in How much money do 1=None USD) you receive from your 2=< 0.38 parents each week that 3=0.38-0.77 you can spend anyway 4=0.78-2.30 you want? 5=2.31-3.83 6=>3.83 Perceived academic How do you perceive 1=very good achievement your academic 2=average success? 3=poor The review of literature helped to determine the choice of independent variables Bandura's Social Cognitive Theory provided the 72

PAGE 87

theoretical foundation. As a result the independent variables were categorized using this framework. Under the demographic factors domain the following questions were asked: school type, gender, age, family origin (urban/rural), levels of weekly allowance, perceived economic situation, and academic performance. Table 3.4 presents the social environmental factors and perception of the social environment. Social environmental variables measured whether the adolescent's parent, siblings, and other household members smoke (O=no, 1 =yes). Teacher influence was measured exposure to teachers' smoking on school premises (O=rarely/never, 1 =everyday). Friend influences were measured by two questions: close friend's smoking status (O=none, 1 =1 friend, 2=2 friends, 3=3 friends, 4=4 friends, 5=not sure), and close friends' approval of smoking (1 =approve, 2=not sure, 3= don't approve). Perceived smoking prevalence among peers is an ordinal variable but for the purpose of the analysis is treated as continuous (1 =less than 10%, 2=10-25%, 26-50%, 51-75%, 76-90%, 6=more than 90%). 73

PAGE 88

Table 3.4. Explanatory (Independent) Social Environmental factors Variables Does your mother or O=neither smokes father smoke? 1 =one smokes 2=both smoke Sibling smoking Does any of your siblings 0= no smoke? 1= Relative smoking Does any one else in 0= no household smoke? 1= Teacher smoking How often do you see O=rarely/never teachers s ? Friend smoking Out close friends O=none how many of them are 1=1 friend smokers? 2=2 friends 3=3 friends 4=4 friends 5=not sure Close friends' approval smoking Perceived prevalence of How many percentage of 1=1ess than 10% smoking among peers your peers do you think 2=10 to 25% smoke at least once a 3=26-50% week? 4=51-75% 5=76-90% 6=more than 90% Data Entry After the data collection, 211 questionnaires were entered into the SPSS statistical software by the researcher. The data entry sheet was created with all 57 questions and their sub-items. Open-ended questions were categorized where it was needed. The data entry and cleaning process took two months. 74

PAGE 89

Analysis Analytic techniques included descriptive statistics and logistic regression analysis using SAS 8.02 and SPSS 11.0, under guidelines described in Tabachnick & Fidell (2001 ). Each hypothesis was tested using logistic regression, followed by separate tests for socio-demographic and social environment factors. Logistic regression was used again to produce a final model based on these constructs. Models were built and tested using classification of cases, coefficient of determination via the Nagelkerke adjusted R2 statistic, and chi-square for goodness-of-fit tests. Models using constructs were tested for goodness-of-fit using the Hosmer & Lemeshow statistic. Semi-Structured Interviews in Dursunbey, 2002 As originally conceived, this study combined a cross-sectional survey with semi-structured interviews, with the interview subjects being a convenience sample of the survey subjects. Sharing the purpose of the field season one interviews, the qualitative portion of the study was designed to further explore the cultural influences on youth smoking. Semi-structured interviews in the field season two asked many of the same questions as the field season one study interviews. For smokers, it asked how the subjects started smoking, when and where they tried their first cigarette, what influenced them to start smoking, and why do they smoke now? A typical subject's day related to smoking? What do they see as benefits and disadvantages of smoking? What people around them smoke? How does that influence them? For nonsmokers, why do they think their peers smoke? 75

PAGE 90

------------For both smokers and nonsmokers, what does smoking mean in their social environment (See Appendix A). Data Collection and Sampling. During the conduct of the survey, a subset of the survey subjects i n each school was interviewed The 42 student interviews averaged 45 to 60 minutes in length. Some of the students were interviewed more than once, since they were able to articulate their smoking very well. Some of the principals, school counsellors and teachers were interviewed more than once. There were total of 15 secondary interviews. Secondary interviews allowed researcher to discuss the emerging themes in greater detail. Students were assured that everything they said was confidential and would not be revealed to school administrators and parents. While the researcher followed an interview script, she also pursued relevant topics raised by the interviewees. This opened the interview process to explore new cultural and social factors from the perspective of the informants. Additional interviews were conducted with 3 teachers and 2 school counsellors, 5 principals and 4 vice-principals Due to the emerging themes, one middle school principal and vice principal were informally interviewed, along with 4 adults and 10 students from other grades. Thirty interviews were taped-recorded. Analysis. As the transcription and analysis of the qualitative data proceeds, it is clear that its value lies in its description of the links between adolescent smoking and cultural/institutional factors such as Turkish education policy, long-standing tradition, and economic hardship. All records of the interviews have been examined. The repetition of themes occurring in the interviews suggests that they can be confidently used to interpret the 76

PAGE 91

results in the quantitative portion of the study. Further analysis for sub codes will take place in the future. All the notes from these interviews were read for the emerging themes. All of the taped interviews were listened and transcribed. Seventy interviews have been formally transcribed and coded under the Open Coding method. The Social Cognitive Theory informed the coding process and helped in organizing the themes and categories. Codes were constructed based on a preliminary scanning of the text as recommended by Crabtree & Miller (1999), the relevant text in the transcripts were highlighted. Emerging themes and categories under these themes were created. The coding consists of tables associating direct quotes with themes identified by the participants, such as "motivation to lose weight", "hospitality ritual", "mixed messages from teachers and family members", "cigarette brands as status symbol," "relaxation," "boredom," or "passing the time." The codes and emerging themes were discussed with Candan Duran-Aydintug, Ph.D., a Turkish sociology professor at the University of Colorado. In the process of analyzing the transcribed data, the main themes, categories and subcategories under these themes have been identified. The adult interviews particularly shed light on the role of tobacco in the lives of the youth. These adult interviews raised concerns about the structural level of societal change and social issues on tobacco use. All of these data was triangulated to have the full understanding of adolescent smoking behaviour from different point of views. The quotes related to major themes were translated into English. 77

PAGE 92

Methodological Limitations For the most part, the findings of this study can be generalized to the population of Dursunbey. The study gives useful insights in understanding rural towns with conservative populations. It should be noted that smoking among females is much more common in larger cities. The social norms and high acceptability of smoking in Dursunbey is common in Turkey. Although the smoker is defined as smoking even one cigarette within thirty days of the survey, teenagers often did not see themselves as smokers. Various reasons were given: occasionally smoking did not constitute smoking in their view, or if they did no buy the packs themselves. They also do not see themselves as smokers if they do not inhale. Many say that the practice of inhaling qualifies as real smoker, "if you don't inhale you are faking it." Some of the teenagers said that "if you don't inhale you will get throat cancer." The researcher relied on the youth's self-report in determining smoking status. It is possible that some of the students did not report their smoking or some of them underreported the frequency of smoking. Also many other measures relied on adolescents' self-report rather than direct measure. 78

PAGE 93

CHAPTER4 DURSUNBEY, ITS HIGH SCHOOLS, AND CHARACTERISTICS OF THE STUDENTS A. The Town of Dursunbey This chapter describes the setting of the study at two levels: the small city of Dursunbey, and the five high schools in the city. In addition, the life style and daily activities of youth in Dursunbey will be described. Ethnographic observations will be utilized in describing the daily lives of these students. The study area for this research, Dursunbey, Turkey, is a small city or district town in the Aegean region that is one of the most populated regions of Turkey. The town of Dursunbey is on the border of the Aegean and Marmara regions of Turkey and is situated within the province of Balikesir. Located about 200 kilometers south of Istanbul, it is one of nineteen towns of its size or larger in Balikesir. Figures 4. 1 show the maps of Turkey and the province of Balikesir with the town of Dursunbey respectively. In this chapter, I first introduce the study area, Dursunbey, Turkey. Then, I provide an explanation of the data population. Other health scientists have investigated tobacco use by youth in cities and larger towns; no other researcher is focusing on the communities of smaller size that represent important segments of the Turkish population. 79

PAGE 94

The majority of Dursunbey's land is covered with forest and the town itself has an altitude of 639 meters 1 above sea level. Its climate is one of extremes, with cold winters and hot summers. Dursunbey covers an area of 1,952 square kilometers. It has a total of 643 kilometers of roads; twenty two percent is asphalt and seventy-eight percent is gravel (Kaymakam Office, 2002). While seventy nine percent of the villages has sufficient drinking water, twenty one percent of them lacks sufficient drinking water supplies. An adequate sewage system exists in twenty villages however, there are inadequate sewage systems in six villages. As of 2002, the construction of new sewage systems has been in progress for fourteen villages The town of Dursunbey is the center of health care and education for its 103 villages There is one hospital with 75 beds, plus one health center in the town and four in the villages. There is one district library with 15,347 books. Many village students come to Dursunbey for their education. In my survey sample (N=211), 50. 2% (105) of the tenth grade students come from the area's villages. In 1994, a branch of Balikesir University's Vocational College opened in Dursunbey. This school offers education in forest industry, construction, and fashion design. Currently, it hosts about 200 students. 1 Equal to 2 096 feet. 80

PAGE 95

Figure 4. 1. Map of Turkey istanbul ( \ . 4KOO$Cii Ont#tl . ... . : Gemtik .. Bursa < . : :-' .. ':''-:: .:. Mal'lisa . AJ'I8t9fia izmir .rr. 15 U i D m Samsun Trabzort Town ofDursunbey in Balikesir Province, Turkey Dursunbey has both urban and rural characteristics. Based on 2000 census data, the urban population of 14,774 was mainly concentrated in Dursunbey; the rural population was 32,928, scattered among one hundred three villages Twenty-eight percent of these villages have a population size of below two hundred people. There are sixteen villages with a population of five hundred or more people, and only three villages have populations of above one thousand people. Dursunbey has a very young population, reflecting Turkey's general age demographic According to local Government health organizations, 23. 1% of the population is under 15 years of age; 38. 6% is between the ages 81

PAGE 96

of 15 to 40. The greatest proportion of the town's population is young (approximately 61. 7%). Due to the lack of employment opportunities, many young people leave Dursunbey for the bigger cities, the latest exodus occurring after the economic crisis of February 2001. Two of the municipality's government banks closed down in 2002. The largest remaining employer of local residents is the struggling logging industry, which provides jobs for around 1000 people. Dursunbey hosts the second largest forest in Turkey. Sixty percent of its land is covered with forest, however, over-cutting has led local businessmen to start bringing timber from Russia. Another Dursunbey industry, coal mining, is on the wane in Odakoy, Cakirca and Hamzacik, three of the district's villages. The agriculture in the region consists mostly of subsistence farming with a minor cash crop component. Dursunbey has a religiously conservative population with a low level of education. The literacy rate is 97% by 1997 figures (Dursunbey Governor's office, 2002). The town of Dursunbey has six primary schools2 and the remaining 45 schools are in the 103 villages. The total number of pupils in primary schools is 6,320 in the greater Dursunbey area. Girls and boys constitute 3,094 and 3,226 of the students respectively. In addition, there are five high schools with 885 students (278 girls). The total number of students in all educational institutions is 7,205 (i.e. primary and high schools combined). The five high schools have a total of 103 teachers; 226 teachers are working in Dursunbey's middle schools. There is one teacher for every 25 students. During April and May 2002, I conducted a survey with the census of all of Dursunbey's 211 tenth graders, ages 15 to 17, using a 57 -item, self2 Primary school in Turkey consists of 8 years compulsory education that includes 5 years of elementary school and 3 years of middle school. 82

PAGE 97

administered questionnaire. Nine variables in the survey described the study population: school type, gender, age, mother's and father's educational attainment, household income, perceived economic status, student's weekly allowances, place of origin, and place of residence. Each variable is discussed separately below. B. High Schools in Dursunbey There are five high schools in Dursunbey. The high schools represent a wide range of school types that are common in Turkey. They are regular high school with multiple programs, religious high school, health vocational high school, female vocational high school, and industrial vocational high school. There are a total of 885 students in these five high schools, in grades 9 through 12. Table 4. 1 represents the population distribution among five high schools. Table 4.1. School Type School Girls Boys Total Type Number Percent Number Percent Number Percent Religious 14 100.0 14 6.6 Health 16 72. 7 6 27.3 22 10.4 Regular 31 42.0 43 58. 1 74 35. 1 Girls 7 100.0 7 3.3 Industrial 9 9.6 85 90.4 94 44.5 Total 63 29.9 148 70. 1 211 100.0 Although all these schools lack facilities for physical education, organized soccer teams are very popular among the male students. For a 83

PAGE 98

I limited number of students, these schools offer intensive classes for college preparation. Dursunbey Regular High School Dursunbey High School is the largest regular high school with 304 students (190 boys and 114 girls). Usually the students of this school have high academic aptitude and plan to get into college after graduation. However, the rate for college acceptance is as low as 5 or 6 students per year due to an intense competition for college entrance. Annually, 1.5 million students participate in the nation-wide placement examination for college; of these, approximately 100,000 students get accepted to enter. The curriculum of this school is geared toward the college preparation exam. In addition to hard science classes, psychology, sociology, logic, philosophy, and literature classes are taught. This school also has a training division for accounting. The students who enter into this program take a nationwide placement exam. These students work outside the school as interns for two semesters. Some of these students stay at the dormitories. This high school draws nearly 45% to 50% of its population from surrounding villages. Due to the distance from the school, some of these students live in a private dormitory. There are 7 4 1oth graders in this school in four classes. Forty-two percent of these students were females. Religious High School There are 102 students (70 boys and 32 girls) in this school. More than 90% of its population comes from villages. Usually these students are from low-income farm families and live in a school dormitory. These students have a very slim chance of getting into college, due to lack of 84

PAGE 99

college capacity in their chosen specialty. The curriculum of this school is similar to regular high schools in terms of science classes and literature. In addition to these classes, they study the Koran and prepare for the entry level religious cleric positions that would available to them after graduation, especially in the villages. There was only one tenth-grade class in religious high school, an all-male class of 14 students. Health Vocational High School There are a total of 123 (61 boy and 62 girl) students in this school unlike the other schools, this school is managed under the Ministry of Health. It is one of 382 of its type of school in Turkey. The students of this school are placed based on a nationwide exam. As a result, this school enrolls students from various other cities and towns in the region. Fifteen percent of them come from Dursunbey, 20 percent from cities and towns other than Dursunbey, and 65% of them come from villages of Dursunbey and villages of adjacent towns (e. g. the towns of Sindirgi, Bigadic, Kepsut). Since this school does not provide a dormitory for its out-of-town students, most out-of-town students rent apartments or stay with relatives. There was only one tenth-grade class in this school with 22 students; 73 percent of them were females. Until 1999, the students of this school received additional consideration for college acceptance. Graduates could expect employment both in local hospitals and in other cities. However, the prospects for employment for these graduates collapsed with the Turkish economy and shrinking government expenditures. The last alumnus to find a job graduated from this high school in 1994. 85

PAGE 100

Girls Vocational High School This school has 57 students, with nearly all of them coming from Dursunbey. Girls who go to this school learn home development, childcare, and other household-related subjects. In addition to this regular curriculum, classes are taught in subjects such as literature, geography, and mathematics. Most of these students do not plan to attend college, and would have virtually no chance if they did. There was one tenth-grade class in this school, having only 7 students. Industrial Vocational High School This school hosts 299 students (286 boys and 13 girls). Fifty to sixty percent of its students come from villages and stay in the school dormitory. There were seven 1Oth grade classes in this school with 94 students. The school curriculum prepares students for work in the forest industry, the metal industry, and in the mechanics of various engines. In addition, regular science, health, literature, and foreign language courses are taught. The students usually take internships in local industries during the school year and summer break. They are usually not very academically motivated and do not plan to pursue a college degree, something which is beyond the financial means of many of these students' families. This school also houses the Anatolian technical high school, which accepts students through a nationwide exam. Unlike their counterparts in the rest of the school, students in the Anatolian technical program have four years of coursework, which includes English and computer instruction. As a result, these students have an academically different profile than the rest of the high school. 86

PAGE 101

C. Characteristics of the Students Who Were Surveyed This description of the study participants is based on the knowledge I gained during the in-dept interviews and my personal experiences. As in many small towns in Turkey, segregation of the sexes for recreational activities is common in Dursunbey. Men spend much of their free time in the smoke-filled coffee houses. Many young boys spend their time after school in the internet cafes, which are smoke-filled like the coffee houses. Boys socialize in internet chat rooms, in many cases frequenting internet cafes to smoke. Despite the fact that minors are legally not allowed to coffee houses, they still visit coffee houses, since rules restricting access to minors are not strictly enforced. Other common activities for boys during the summers are fishing, swimming, and picnics. Boys cited activities such as watching action movies and soccer games, playing cards, billiards and other games. A picnic area at the outskirts of the town called "sucikti" is a common hang out nowadays for boys and girls. Boys also help their family businesses, working at coffee houses, markets, and other small shops. Girls are much more restricted than boys in their social activities. They usually spend time after school at home with their mothers, and are sometimes allowed to gather at girlfriends' homes. Their mothers usually socialize at the social teas in the afternoons, with girls usually accompanying their mothers. Picnics are also among the common activities for females. Another activity for girls is watching TV, commonly viewing soap operas and paparazzi programs. Girls knit and prepare their dowry for their future home. This is a common activity even for girls who continue to their high school education, since their chances of going to college are very slim. They know that their only option is to marry and have a family. The role of 'homemaker" is still socially normative for women. 87

PAGE 102

As was mentioned before, a two-year college opened in Dursunbey in the 1990 's. College students from larger cities have altered Dursunbey's social environment somewhat. In recent years, three pastry and coffee houses opened. However, many students complain about the lack of recreational opportunities in town, a problem so acute that even high school principals and teachers echoed this point during interviews. High schools lack government funding for sport equipment and facilities. The only active leisure activity clubs are the choir in Dursunbey regular high school and the soccer team, and a folk dancing group that performs in national celebrations like independence day. The major concern of these young people is the lack of job opportunities and the high level of unemployment in town. Most would not be able to start a family without their family's financial support. They have little hope for the future, since having a high school education does not guarantee jobs or entrance to college; many students have very little motivation to study. In fact, the major complaint of teachers and principals was that many students were not academically motivated. From my personal observations, it appeared that many students, especially boys, did not spend much time doing homework, or they do it at the minimal level. Many of the adult educational professionals expressed a concern for Turkish cultural "degeneration" (the word is the same in Turkish and English) and its potential relation to increased tobacco use. They cite the dramatic drop in per capita income, a sense of hopelessness and a lack of goals on the part of the young as contributing factors. This lack of hope for the future was frequently communicated during the student interviews. The students of Dursunbey know that almost none of them will get into college (only 7-8% of students reach college nationwide and many of these come from special "Anatolian" high schools for high achievers and private high schools. ) 88

PAGE 103

These students know that even with a college diploma, the prospects for getting a job in the depressed Turkish economy are grim. According to the principal, the last student from the health vocational high school to receive a job graduated in 1994. This lack of possibilities for their students is demoralizing for the teachers themselves. This was a common sentiment expressed among the interviewed educational professionals The connection of this sense of hopelessness with tobacco was only indirectly referenced in the student interviews, where smoking was frequently connected with idleness, "having nothing to do," "passing the time," and "relieving stress." This may be connected with a lack of motivation to study (given the lack of rewards), but it may also be a reflection of Turkish education policy for regular and vocational high schools, which de-emphasizes homework in favour of preparation for exams. In the following section, the sociodemographic characteristics of the study participants will be discussed. Age and Sex Distribution The usual age for 1Oth graders i s 16. The study population ranged from 15 to 19 years of age; 93 were 16 years old. The mean age was 16. 61 (SD=O. 85) As seen in Table 4. 2, there are more male students in the age group between ages 17 and 19. The older students tend to be male, since girls reaching high school are usually very academically motivated and graduate on time, whereas some male students tend to repeat a grade. In my study sample there were 29 boys between ages 18-19, while there was only one girl in the same age group. Seventy percent of the study population was males compare to only 30%. Girls are more likely to drop out of school system after the compulsory 89

PAGE 104

education of 8th grade. Table 4. 3 shows the percentages of place of origin for study participants. Since these towns and cities are similar in size and characteristics, all are combined under urban category. Fifty percent of the students in the study population are from urban areas. Table 4. 2. Age Categories Age Girls Boys Total Categories Number Percent Number Percent Number Percent 15 years old 10 83.3 2 16. 7 12 5 7 16 years old 39 41.9 54 58. 1 93 44.0 17 years old 13 17. 1 63 82.9 76 36.0 18 years old 1 3.8 25 96.2 26 12.3 19 years old 4 100.0 4 1. 9 Total 63 29.9 148 70. 1 211 100.0 Family Origin Urban Origin. Students fitting this description come from the town of Dursunbey, and other towns and cities, some of which are outside the region. The extra-regional students attend the health vocational school or the Anatolian technical school, the latter being attached to the industrial vocational school. These extra-regional students were accepted to their schools through nationwide exams. 90

PAGE 105

Table 4. 3. Family Origin Family Origin Category Number Percent Dursunbey Urban 80 37.9 Another Town Urban 8 3.8 Another City Urban 18 8.5 Village Rural 105 49.8 Total 211 100.0 Rural Origin. These students come from the villages of Dursunbey and the villages of other towns. These students usually return to their villages during the summer to help their farm families. Their family incomes range from low-income to middle-income. Fifty percent of the students are rural in the study sample. Place of Residence Seventy percent of the students in the sample indicated that they live with their family and 4% indicated that they live with relatives. The rest either live with friends in rented apartments (7%) or dormitories (19. 4%). Table 4. 4. Place of Residence Place of Residence Number Home with family Home with relatives Out of home with friends/alone Dormitory Total 148 8 14 41 211 91 Percent 70. 1 3.8 6.6 19.4 100.0

PAGE 106

Educational Attainment of the Parents Table 4. 5 shows the parents' educational attainment. In the study sample, 20. 4% of the mothers had no schooling and 63% were elementary school graduates. Five percent of the fathers had no schooling and 59. 3% were elementary school graduates. Twenty five percent of fathers were high school graduates and 7. 6% were college graduates Table 4. 5. Parents' Education Education Mother Father Number Percent Number Percent Secondary 15 7. 1 68 32.2 School/College Elementary School 133 63.0 125 59.3 Some Elementary 20 9.5 8 3.8 School No School 43 20.4 10 4. 7 Total 211 100.0 211 100.0 Economic Status Table 4. 6 displays the student family income level in US dollars and Turkish liras, reported as monthly salary. Some Turkish economic statistics help to interpret the table. Monthly food expense for a family of four is roughly four hundred dollars. Bread costs 15 cents per loaf and a family of four would consume three to four loafs a day. The minimum wage is two hundred dollars a month, and the government poverty level stands at five hundred dollars. Given that the lack of Dursunbey's job opportunities and high unemployment rates, the residents have a low to middle income level. 92

PAGE 107

Monthly Household Income. Many of these families grow their own food and raise their own stock animals for family use, as well as for sale. Twenty-seven percent of families earn less than minimum wage, and only 10% of families earn more than three hundred dollars monthly. Forty-six students indicated that they don't know their families' income. Table 4. 6. Monthly Household Income Monthly Income (in Turkish In US Number Percent Liras) Dollars 50 million 38 23 10.9 51-150 million 37-115 33 15.6 151-300 million 116-230 56 26.5 301-450 million 231-345 32 15.2 451 million+ 345+ 21 10.0 Don't know 46 21.8 Total 211 100.0 Students' Perception of Economic Status. Nearly 60 percent of the students reported that their family's economic situation is average. Two percent of the students perceived that their economic situation was very poor while same percentage of students perceived their economic situation as very good. Six percent of the students perceived their economic situation as poor and 32% perceived their economic situation as good. Table 4. 7 shows the adolescents' perception of their economic situation. 93

PAGE 108

Table 4. 7. Perceived Economic Situation Perception of economic situation Very poor Poor Average Good Very good Total Number 5 13 121 67 5 211 Percent 2.4 6.2 57.3 31.8 2.4 100.0 Weekly Allowance for Discretionary Spending. Table 4. 8 shows the adolescents' weekly allowances. Allowances ranged from 8% reporting that they receive no allowances to 8% reporting that they receive in excess of $3. 50 US per week. In terms of ability to buy cigarettes, the upper tier of allowances can buy two or more packs per week. Thirtyfive cents will purchase an inexpensive pack of Turkish cigarettes, whereas a premium foreign cigarette like Marlboro will cost as much as $1. 75. Table 4. 8. Weekly Student Allowance Weekly Allowance (in In US Turkish Liras) dollars None Less than 500,000 500 I 000-999,000 1,000. 000-2,999. 999 3,000. 000-4,999. 999 More than 5 million TL. Total < 0. 38 0. 38-0. 77 0. 78-2.30 2.31-3.83 >3.83 94 Number 17 47 31 64 35 17 211 Percent 8. 1 22.3 14. 7 30.3 16.6 8. 1 100.0

PAGE 109

Academic Success Table 4. 9. Academic Achievement Academic Girls Boys Total Success Number Percent Number Percent Number Percent Very good 2 3.2 5 3.4 7 3.3 Good 32 50.8 63 42.6 95 45.0 Average 27 42.9 76 51.4 103 48.8 Poor 2 3.2 4 2. 7 6 2.8 Total 63 100.0 148 100.0 211 100 Table 4. 9 shows students' perception of their academic achievement. Three percent of sampled students perceived their academic success as very good while 45% perceived their academic success as good. Forty nine percent of the students perceived their academic success as average. Finally, approximately 3% of students perceived their academic success as poor. 95

PAGE 110

CHAPTER 5 SURVEY FINDINGS A. Introduction The results are organized into 4 sections: section one addresses the integrity of the data; section two presents descriptive statistics for all measured variables; section three examines the study's major hypotheses; and section four presents results from models testing the relative contribution of demographic and social environmental variables to the prediction of smoking behavior. B. Section One: Data Integrity The Dursunbey data set was screened for missing data, outliers, and normality. Cross-tabs and scatter plots were created to examine the pattern of subject responses to the survey and to detect outliers. From these investigations, a discrepancy was detected in the responses to the two outcome variables that measured smoking status. Specifically, students who were occasional smokers had a tendency to minimize their smoking. In response to question twelve, three students out of 211 answered the question: "Do you smoke now?" as "No" or "Rarely less than once a month;" but on question seventeen, "During the past 30 days, on how many days did you smoke cigarettes?" they answered "one to two days or more". To correct this discrepancy, question seventeen, a World Health Organization validated question, was retained. For these specific study subjects, the rest of their 96

PAGE 111

answers were closely reviewed to see if there were any more discrepancies, but no additional problems were detected. A large number of cases with missing data can create bias and problems with the generalizability of results (Tabashnick & Fidel!, 1996). This problem was ameliorated in advance by data collection procedures designed to minimize incomplete answers on the survey. In each classroom the researcher took 15 minutes to explain the questionnaire and the procedures to the subjects. The researcher was also present during the 3035 minutes administration in each classroom and allowed students to ask questions if they were not clear about questions on the survey. In addition, the researcher reviewed each survey for possible mistakes (e.g. overlooking some questions, or not understanding skipping patterns) at the time students turned in the survey. If problems were detected the students were asked to complete the missing or faulty questions. Table 5.1. Variables with Missing Values Variable Name Family Income Place of Origin Friend's Approval of Smoking Missing% 22% 0.9% 0.5% Table 5 1 presents the variables with missing values. The family income variable had 22% missing values Since the missing values were larger than ten percent, this variable was removed from the analysis (Aday, 1996). The question addressing the best friends' approval of smoking had only one missing value and the missing value was replaced with the mean. The analysis was run with and without the missing value and the results were similar. This completed the data corrections of individual variables. The 97

PAGE 112

data analysis for the hypotheses and model testing were run based on the full sample of 211 students. The occurrence of zeros in contingency tables (IV x DV) requires collapsing categories, or if the variable is ordinal modeling it as if it were continuous. The perceived economic situation and academic achievement categories were collapsed to avoid the empty cells. In addition, the ordinal variable "exposure to teacher smoking" and nominal variable "place of origin" were dichotomized to ease the interpretation in the logistic regression analysis. Reducing the number of variables leads to parsimonious, numerically stable, and more easily generalized regression models In addition, too large a number of independent variables for the sample size increases the estimated standard errors (Hosmer and Lemeshow, 1989) and increases the risk of making a Type I error. To minimize these risks some of the variables were combined to form a simple measure. For instance, mother and father smoking were initially two separate variables; since, however, there was no difference in smoking outcomes based on mother and father smoking those two variables were combined to form parental smoking. C. Section Two: Descriptive Statistics Description of Sample Table 5.2 presents the demographics of the sample. All 1 01h graders (ages 15-19) were surveyed in all five high schools in Dursunbey (N=211 ). The health vocational school, religious school, and girls vocational high school had small sample sizes relative to the Regular and Industrial high schools. Seventy percent of the sample were male, with median age of 17 98

PAGE 113

years (range 15-19 years). Forty-nine percent of the sample was from urban areas. Table 5.2. Description of the Study Population Variable Schools Religious School Health Vocational School Regular High School Industrial Vocational School Girls Vocational School Gender (male) Median Age (15 -19 years-of-age) Mean Age Place of Origin Dursunbey (Urban) Another town (Urban) Village (Rural) Description of Variables Results (N=211) 6.6% 10.4% 35.1% 44 5% 3.3% 70 0% 17 years 16.6 (SD=0.85) 38 0% 11.0% 51. 0% Table 5 3 shows the results of univariate analyses for each variable. The measurement of the respective variables determined the particular types of statistics reported. None of the continuous variables were normally distributed (Shapiro-Wilk test of a normal distribution) The desirable outcome for Shapiro-Wilk test is to fail to reject the null hypothesis. The probability value for age, economic situation, academic achievement, and perceived prevalence were less than 0.05, we conclude that these variables are not normally distributed. However, Q-Q normality plots for age and perceived prevalence variables suggested that the distributions were pretty close to be normal. Logistic regression analysis does not assume the 99

PAGE 114

multivariate normality. Thus I will use the variables without any transformation. Table 5.3. Descriptive Statistics of Variables Variable N Type Mean so MIN MAX Name Dependent: 211 Binary .42 0 1 Current smoking Independent: SociodemograQhics School type 211 Binary 3.7 1 5 Age 211 Con tin. 16.6 .85 15 19 Gender 211 Binary .70 0 1 Economic situation 211 Con tin 2.3 .60 1 3 Allowances 211 Categ 3.5 1.4 1 6 Origin 211 Binary .50 0 3 Academic 211 Con tin. 2.5 .55 1 1 achievement Social Environment 211 Binary 74 0 1 Parent smoking 211 Binary .29 0 1 Sibling smoking 211 Binary .15 0 1 Relative smoking 211 Binary .48 0 1 Teacher smoking 211 Categ. 2.7 .90 1 4 Friend smoking 211 Categ. 2.2 .59 1 3 Friend approval 211 Cantin 3 5 1.5 1 6 Perceived Qrevalence SD= Standard Deviation MIN= Minimum Value MAX= Maximum value Bivariate Analysis To Describe RelationshiQs between Variables Normality (ShapiroWilks Test) .86 <.05 .76 <.05 .71 <.05 .92 <.05 To assess the potential problems of collinearity in the final models, existence and strength of the relationship between variables, then between independent variables and the dependent variable, were tested. Table 5.4 presents the level of significance of Pearson Correlation test for each 100

PAGE 115

variable. Age and gender related significantly with other Independent variables. Since literature suggests that age and gender influence youth smoking behavior, I adjusted for these confounders in the regressions. Table 5.4. Correlations Between Variables Variable Age Gender School Econ Allowance Success Origin IV Age .. Gender .408 School .103 .271** Econ -.129 -.256** -.212** Allowance .210** .198** 047 .298** Success -.108 162* 225** .181** 019 Origin .048 .256** 095 -.235** -.092 -.142* Parent .146* -.063 .179** .111 -.011 080 -.095 smoking Sibling .103 -.034 -.063 -.032 .003 002 -.039 smoking Relative .213** .131 -.141* -.004 .133 -.034 187** smoking Teacher 177* .231** .193** -.045 .122 077 -.100 smoking Friend -.141* -.143* 025 .102 105 .136* -.175* smoking Approval 173* -.311** -.086 .103 -.072 137* 046 Perceived .063 .219** .272** -.038 .055 -.016 -.058 prevalence DV Smoking .383** .426** .173* -.088 .134 -.307** -.015 status u Correlation is significant at the 0.01 level Correlation is significant at the 0.051evel Factors significantly correlated with smoking status at alpha level of 0.01 included teacher smoking, friends' approval of smoking, age, gender and academic success. Factors significantly correlated with smoking status at alpha level of 0.05 included high school type, relative smoking and perceived smoking prevalence among peers. 101

PAGE 116

Table 5.4. Correlations Between Variables (Continued) Variable Parent Sibling Relative Teacher Friend Friend Perceiv smokng smokng smokng smokng smokng approv Prevalnce IV Age Gender School Econ Allowance Success Origin Parent smoking Sibling -.034 smoking Relative -.015 104 smoking Teacher 006 090 .124 smoking Friend 019 034 002 -.020 smoking Friend .122 -.143* 029 .064 .159* approval Perceived .170* -.061 -.012 .270** .032 088 prevalence DV Smoking 067 .087 152* 305** 106 327** 153* status ** Correlation is significant at the 0.01 level Correlat i on is significant at the 0.05 level The Study Population and Smoking Behavior Table 5.5 presents the smoking status of the high school students in Dursunbey, Turkey. Eight-three percent of the students had tried smoking at least once, with a mean age of initiation of 11.8 years. Fifty-six percent of the students had smoked 1 0 cigarettes or fewer over their lifetime, and 36% had smoked more than 100, meeting the World Health Organization's definition of being an occasional or daily smoker, or an ex-occasional or ex daily smoker who has quit. Forty-two percent of the students reported that 102

PAGE 117

they had smoked in the past 30 days (n=88). Twenty-six percent of the 211 students were daily smokers (n=54 ), and sixteen percent (n=34) were occasional smokers. For data analysis purposes, occasional and daily smokers were combined as smokers. Thirty-two percent of the smokers smoked between 2 to 20 cigarettes per day. Smoking behavior was not described based on gender due to the small sample size for girls. Although, gender differences are a clear and compelling difference, the extent of the gender differences cannot be determined in this study but should be included as an important component of future analysis of smoking behavior among Turkish youth. However, the smoking behavior of the study population will be described based on gender to give a sense about the study population in detail. 103

PAGE 118

Table 5.5. Smoking Behavior in the Study Population Variable Female Male Total (N=63) (N=148) (N=211) (%)a (n) {%t {n) (%t Ever-tried smoking No 24 38 1 11 7.4 35 16.6 Yes 39 62.0 137 93.0 176 83.4 Mean age of initiation (6 11.8 11. 8 11.8 17 years-of-age) 2.7 Number of cigarettes smoked life time None 24 38.1 11 7.4 35 16.6 1-2 puffs 15 23.8 22 14.9 37 17.5 1-1 0 cigarettes 16 25.4 28 18.9 44 20.9 11-99 cigarettes 3 4 8 17 11. 5 20 9 5 More than 1 00 5 7.9 70 47.3 75 35.5 How many days have you smoked in the past 30 days? None 57 90.5 66 44.6 123 58.3 1-2 days 2 3.2 8 5.4 10 4.7 3-5 days 1 1.6 12 8.1 13 6.2 6-9 days 3 2.0 3 1.4 10-19 days 1 1 6 7 4.7 8 3 8 20-29 days 8 5.4 8 3 8 30 days 2 3.2 44 29 7 46 21.8 Number of cigarettes smoked per day? None 57 90.5 66 44.6 123 58.3 Less than 1 1 1.6 12 8.1 13 6.2 1 2 3.2 10 6.8 12 5.7 2-5 1 1.6 22 14.9 23 10 9 6-10 2 3 2 17 11.5 19 11. 5 11-20 14 9 5 14 6.6 More than 20 7 4.7 7 3.3 a Row percentages, Column percentages 104

PAGE 119

Demographic Factors and Smoking Behavior. As defined in the Methods chapter, a current smoker is as someone who has smoked in the past 30 days at least one or more times. A non-smoker is defined as someone who did not smoke in the past 30 days, including students who never tried smoking, ex-smokers, and triers. Table 5.6 indicates that smoking rates among school types varies widely, from 14% for the girls' vocational school to 64% for the religious school. It must be noted that gender is confounded with these two school types, in that the religious school had all male students in 1oth grade, and the girls school did not have any boys. Since boys are more likely to smoke, smoking rates were higher for the religious high school. Table 5.6 shows that 10% of girls (n=6) and 55% of boys (n=82) reported smoking in the last 30 days, indicating the possibility of a strong gender effect in the risk of smoking. The gender difference appears compelling, but the sample size for girls is not large enough to allow for meaningful conclusions about the extent of differences between female smokers and nonsmokers. Table 5.6 also shows that smoking rates are virtually the same for urban and rural 1Oth graders, at 43% and 41 %, respectively, and that smoking rates rise steadily with age. While none of the 15 year olds smoked, all of the 19 year olds were smokers. It should also be noted that only four students were 19. Thirty-four percent of the students reported their economic situation as good, 57% reported their economic situation as average, and 9% reported their economic situation as poor. Representation in these categories was proportionate between smokers and non-smokers. 105

PAGE 120

Table 5.6. Sociodemographic Factors and Smoking Behavior Variable Non-Smoker Smoker Total (N=123) (N=88) (N=211) {n} {%t {n} {%t {n} {%t School Type Religious 5 36.0 9 64 3 14 6.6 Health Vocational 17 77.0 5 23 0 22 10.4 Regular High 54 73.0 20 27.0 74 35 1 Girls Vocational 6 86.0 1 14 3 7 3.3 Industrial Vocational 41 44.0 53 56.4 94 45.0 Gender Male 66 45.0 82 55 4 148 70 0 Female 57 90.5 6 9.5 68 30.0 Age 15 12 100 0 12 6.0 16 66 71.0 27 29.0 93 44 1 17 37 49.0 39 51.3 76 36.0 18 8 31.0 18 69.2 26 12.3 19 4 100.0 4 1.9 Family origin Urban 61 58.0 45 43 0 106 50.0 Rural 62 59.0 43 41. 0 105 50.0 Weekly allowance (USD) None 6 35 3 11 65 0 17 8 1 < 0.38 36 77.0 11 23 4 47 22.3 0.38-0.77 22 71.0 9 29 0 31 14.7 0.78-2.30 34 53 1 30 47.0 64 30. 3 2.31-3.83 20 57.1 15 43 0 35 16. 6 >3.83 5 29.4 12 71.0 17 8 1 Economic situation Poor 12 67 0 6 33 3 18 8.5 Average 62 51.2 59 49.0 121 57. 3 Good 49 68.1 23 32 0 72 34.1 Academic achievement Poor 1 17 0 5 83 0 6 3.0 Average 48 46.0 56 54 0 104 49 0 Good 74 73.0 27 27.0 101 48 0 a Row percentages, Column percentages 106

PAGE 121

The weekly allowance students received varied from none to >3.83 (USD). Eight percent of the students indicated they do not receive a weekly allowance. Similarly, 8% of the students received more than 3.83 (USD) allowance. Most of the students allowances clustered between .38 to 2.30 (USD). Of the adolescents who received more than 3.83 (USD) weekly, 71% were smokers and 29% were nonsmokers. Table 5.6 shows that 3% of the subjects reported their academic performance as poor, 49% reported their performance as average, and 48% reported performance as good. Six percent (5 of 88) of smokers reported low performance versus less than 1% (only 1 of 123) of non-smokers. Thirty one percent of smokers reported their perceived academic performance was good, compared with 60% of the non-smokers. Social Environment and Smoking Behavior. The study included six variables that examined social environment factors. Table 5.7 presents the descriptive statistics for each variable. Six percent of the students reported both their parents smoked, 62% reported that one parent smoked, and 32% of reported neither parent smoked. Of the students who smoked, 72% (64 of 88) reported one or both of their parents smoked, whereas, 64% (79 of 123) of the non-smoking students reported one or both parents smoked. Sixty seven percent (59 of 88) of smoking students indicated that a sibling smoked, whereas 57% (71 of 123) of non-smoking students indicated that a sibling smoked. Fifteen percent of the students reported a relative (non nuclear family member) living in their household smoked, with 22% of smokers and 11% of non-smokers reporting a smoking relative in the household, respectively. 107

PAGE 122

Table 5.7. Social Environment Factors and Smoking Behavior Variable Non-Smoker Smoker Total (N=123) (N=88) (N=211) {n) {%t {n) {%t (n) {%t Parent smoking Neither 44 64 7 24 35.3 68 32 2 One parent 71 54 6 59 45.4 130 61. 6 Both parents 8 61. 5 5 38.5 13 6.2 Sibling smoking No 91 61. 1 58 38.9 149 70.6 Yes 32 51.6 30 48.4 62 29.4 Relative smoking No 110 61.5 69 38.5 179 84 8 Yes 13 40 6 19 59.4 32 15. 2 Teacher smoking No 80 72.7 30 27.3 110 52.1 Yes 43 42.6 58 57.4 101 47.9 Friend smoking Not sure 9 47 4 10 52.6 19 9 0 1-2 friend 52 73.2 19 26 8 71 33 6 3-4 friend 20 26 0 57 74.0 77 36 5 None 42 95 5 2 4 5 44 20.9 Friend approval of smoking Approve 6 33 3 12 66.7 18 8 5 Not sure 63 49.6 64 50.4 127 60 2 Don't approve 54 81.8 12 18.2 66 31. 3 Perceived prevalence Less than 10% 18 58 1 13 41.9 31 14 7 10-25% 20 66.7 10 33.3 30 14.2 26-50% 30 68.2 14 31.8 44 20 9 51-75% 28 68 3 13 31.7 41 19 4 76-90% 22 44.9 27 55.1 49 23.2 More than 90% 5 31.3 11 68.8 16 7 6 Row percentages, Column percentages 108

PAGE 123

Forty-eight percent of students reported seeing teachers smoking at school on a daily basis. The smoking students and non-smoking students reported this measure at 70% and 35%, respectively Thirty-four percent of students reported having 1 or 2 friends who smoked, 37% reported having 3 or 4 friends who smoked, and 21% reported having no friends who smoked. Of the smoking students, only 2% (2 of 88) reported no smoking friend, whereas 34% (42 of 123) of non-smoking students reported no smoker friends. Twenty-one percent of the smoking students reported having 1 or 2 smoking friends and 65% reported having 3 or 4 smoking friends. For the non-smoking students, 42% reported having 1 or 2 smoking friends and 16% reported having 3 or 4 smoking friends. Nine percent of students reported their friends approved of smoking, and 31% reported their friends do not approve of smoking. Of the smokers, 14% reported their friends approved, versus 5% of the non-smokers. A large percentage of both smoking and non-smoking students were uncertain about their friends' approval of smoking. The perceived prevalence of smoking among peers varied widely, with the perceived prevalence approximately evenly distributed among the percentile ranges, from less than 10% to the 75-90% range. Relatively few students perceived smoking prevalence in excess of 90%. D. Section Three: Test of Hypotheses The hypotheses were tested using logistic regression. Logistic regression is used when the outcome variable is discrete (Y=O if the outcome did not occur and Y= 1 if it did occur). Logistic regression quantifies the association between a risk factor (independent variable) and a 109

PAGE 124

disease (or any dichotomous event), after adjusting for other variables (Motulsky, 1995:269). In the following section, the logistic regress ion results presented include prediction accuracy, coefficient of determination, and odds ratios In addition, the Hosmer and Lemeshow Goodness-of-Fit Test (using the SAS 8.02 and SPSS 11.0 software) was used to determine the best fit of the models to the data. Classification Accuracy. This result assesses the success of a model at classifying cases for which the outcome is known (Tabachnick and Fidell, 1996) In the present study the primary focus was on examining the percent of smokers and non-smokers classified correctly. Coefficient of Determination. The Nagelkerke (1991) statistic gives an adjusted coefficient of determination (adjusted R2 ) for logistic regression. This coefficient explains the proportion of variance in the outcome variable explained by the independent variables in the logistic regression model. Odds Ratios. An odds ratio is the change in the odds of an outcome (dependent variable) occurring when the value of one or more of the predictors (independent variables) increases by one unit. For example, a one unit increase (year) in age corresponds to a three-fold increase in the likelihood of being a current smoker. Goodness of-Fit. This is formally evaluated by using the Hosmer and Lemeshow statistic where a good model produces a non-significant chi square (Tabachnick and Fidel!, 1996:598). In other words, when the test statistic is non-significant, the model is a good fit. Test of Sociodemographic Factors In this section Seven hypotheses were tested that related to the sociodemographic factors in relation to adolescent current smoking 110

PAGE 125

behavior. Each variable was analyzed separately but results are presented in one table (see Table 5.8). (H1) predicted a difference in smoking rates based on school type. This hypothesis was confirmed. The adolescents who attended the health vocational school were about 77% (OR= 0.23) less likely to be smokers than students in the industrial schools, p = 0.007 (95% Cl: 0.08, 0.67). In addition, students who attended the regular high school were about 71% (OR=0.29) less likely to be smokers than the students in the industrial school, p < 0.000 (95% Cl: 0.15, 0.55). Finally, students who attended the girls' vocational school were about 87% (OR=0.13) less likely to smoke than students in the industrial school. This result was not statistically significant due to the small sample (and resulting loss of statistical power) in the girls high school (95% Cl: 0.02, 1.11; p-value: 0.63). The overall model was significant (x2= 23.9, df = 4, p<.001); 15% of the variance in smoking behavior was accounted for by school type. The prediction accuracy for the baseline was 58.3% (i.e. the model containing only a constant correctly predicts 58.3% of the smokers without school type included), and the model with school type classified 66% of the cases correctly. The baseline model uses the most frequent occurrence of the outcome variable as the basis for predicting the outcome. In this study, the most frequent occurrence between smokers and non-smokers was the non smoker category (n=123, 58.3%). For this and all subsequent logistic analyses the baseline accuracy was 58.3%. Table 5.8 includes statistical test for this analysis and the logistic regression results from the six other sociodemographic variable hypotheses tests. 111

PAGE 126

Table 5.8. Hypothesis Testing for Sociodemographic Factors Variable x2 Of pN-fil OR 95% C/ value H1: School 23 9 4 0.00 0 .15 Religious 1.39 0.43, 4.47 Health 0 23** 0.08, 0 .67 Regular 0 29*** 0 15, 0 55 Girl's 0.13*** 0 02,1 .11 Industria Ia H2: Male 43.6 1 0 00 0.25 11.8*** 4.80, 29.0 H3: Age 32.8 1 0 .00 0 19 2.9*** 1.92, 4 24 H4: Rural (vs 0.49 1 0.83 0.00 0.94 0.54, 1 63 urban) H5: Allowances 19.2 5 0.00 11. 7 None 0.76 0.18, 3 23 < 0.38 0.13** 0.37, 0.44 0 38-0.77 0 17** 0.05, 0.63 0 78-2.30 0 .37 0.12, 1 16 2.31-3.83 0.31 0.90 1.07 >3 .83a H6: Economic 1.6 1 0.19 0 .01 0.74 0.47, 1.17 situation H7: Academic 20 3 1 0 00 0.12 0.30*** 0.18 0 .52 achievement Reference point N R2 =Nagelkerke R Square OR= Odds Ratio 95% Cl= 95% Confidence lnteNal ***P <0 001, ** P < 0 01, P< 0 05 (H2) predicted adolescent boys will be more likely to smoke than adolescent girls. This hypothesis was also confirmed. There is a clear relationship between gender and current smoking behavior, as predicted in the literature in the US and elsewhere. As seen in Table 5.8, 25% of the variability in 112

PAGE 127

smoking behavior was accounted for by gender, with male students being 11.8 (95% Cl: 4.8, 29.1; p< 0.001) times more likely to be current smokers than females. The overall model was significant at the 0.01 level (x2 = 43.6, df = 1, p< .001 ), and this model predicted 66% of the smoker and nonsmokers correctly. (H3) predicted older students will be more likely to smoke than younger students. This hypothesis was also confirmed. The odds ratio shown in Table 5.8 indicates that a one year increase in age increased the likelihood of being a smoker by about three times (OR=2 9), p<.OOO (95% Cl: 1.9, 4.2). The overall model was significant p<.001 (x2 = 32.8, df = 1) with 66% of the responses correctly classified. The adjusted R2 indicated that 19% of the variance was explained. (H4) predicted a difference in smoking rates as a function of place of origin. This hypothesis was not supported. (H5) predicted the greater the allowance, the greater chance of being a smoker. This hypothesis was confirmed. As with all the logistic regression tests in this study, the model with the constant only classified 58.3% of the cases correctly; but after weekly allowance was added the classification accuracy increased to 64%, p < 0.01 (x2 = 19.2, df = 5), with 12% of the variability in smoking behavior accounted for by allowance. Weekly allowance of predicted adolescent smoking behavior for students who received less than $.38 (OR=0.13) and for the adolescents who received 113

PAGE 128

$.38 to $.77 (OR=0.17) weekly allowance. Adolescents in both groups were less likely to smoke compared to adolescents who received more than $3.83 weekly. These results were significant p-value: 0.001 and 0.008 (95% Cl: 0.04, 0.44; CI:0.4, 0.44, respectively). Although not statistically significant, there was a trend in weekly allowance with students who received $0.78 to $2.30 and $2.31 to $3.83 weekly were less likely to be current smokers than students who received $3.83 weekly. (H6) predicted adolescent smoking will be different based on perceived economic situation This hypothesis was not supported. (H7) predicted adolescent perception of academic achievement will be associated with their smoking status. This hypothesis was supported. After academic achievement was added to the model, the prediction accuracy increased from 58.3% to 64%, p< 0.001 (x2 = 20.4, df = 1) with 12% of the variance explained by academic achievement. Adolescents with good academic performance were 30% less likely to be smokers than the students with average performance. In turn, the students with average performance were 30% less likely to be smokers than the students with poor academic performance, p< 0 001. (95% CI:0.18, 0.52). Test of Social Environment Factors Seven hypotheses were tested relating to the influence of family, friends and teachers on the smoking behavior of the adolescents. Each 114

PAGE 129

--------------------------------------------------------------variable was analyzed separately. Table 5.9 presents the results for each of the seven hypothesis tests. Table 5.9. Hypothesis Testing for Social Environment Factors Variable i Of p-N-Ff Odds 95% Cl6 value Ratios H8: Parent smoking 0.96 1 0.33 0.00 1 3 0 78, 2.08 H9: Sibling smoking 1.60 1 0.20 0.01 1 5 0.81' 2.67 H10: Relative smoking 4.77 1 0.03 0.03 2 3* 1.08, 5.01 H11: Exposure to 19.9 1 0 00 12.2 3.6*** 2.02, 6.40 teacher smoking H12: Friend smoking 73.4 3 0 00 40.0 23.3*** 4.34,125 Not sure 7.7** 1.69, 34.7 1-2 friend 59.6*** 13.2, 269 3-4 friend None11 H13 : Friend aQQroval of 25 1 2 0.00 15 1 smoking Approve 8.9*** 2.81, 28 8 Not sure 4.5*** 2.24, 9.35 Don't approve a H14: Perceived 4.26 1 0.04 0.27 1 2* 1 02, 1.48 a Reference point bAta =0.05 95% Cl= 95% Confidence Interval ** P <0 001, ** P < 0 .01, P< 0 05 (HB) predicted adolescents whose parent smokes are more likely to smoke themselves. This hypothesis was not supported. (H9) predicted adolescents whose siblings smoke are more likely to 115

PAGE 130

be a smoker than adolescents whose siblings do not smoke. This relationship was not confirmed. (H10) stated that adolescents who have other smoking family members in the household are more likely to be a smoker than adolescents who do not have other smoking family members. The results showed relatives' smoking added to the accuracy of predicting smoking status from 58.3% to 61%, p< 0.05 (x2 = 4.8, df = 1 ,), with 3% of the variability in smoking behavior accounted for by relative smoking. Adolescents who have relatives who smoke Jiving in their household are 2.3 times more likely to be a smoker than adolescents who do not have relative who smokes in their household, p-value: .031, this is statistically significant at the alpha level 0.05 (95% Cl: 1.1, 5.1 ;). (H11) predicted adolescents who observe teachers smoking on a daily basis are more likely to be a smoker than adolescents who do not see teachers smoking on a daily basis. This hypothesis was confirmed. Adolescents who witnessed teachers smoking were 3.6 times more likely to be smokers than those who had not witnessed their teacher smoking (OR=.36, Cl: 2.02, 6.40), p< 0.001 (x2 = 19.9, df = 1). The model accounted for 12% of the variance in smoking status, with 65% of the cases correctly classified. (H12) predicted adolescents who have smoker friends are more likely to likely to be a smoker than adolescents who do not have smoker friends 116

PAGE 131

This hypothesis was confirmed. After friends' smoking status was added to the model, the accuracy of predicting smoking status increased to 76.3% (l = 73.4, df = 2, p< 0.001 ), with 40% of the variance in smoking behavior explained by friends smoking. Adolescents who have 1 or 2 friends who smoke are eight times more likely to smoke than adolescents who have no smoker friends, p= 0.008 (95% Cl: 1.7, 34.7). Furthermore, adolescents who have 3 or 4 friends who smoke are 60 times more likely to smoke than adolescents who have no smoker friends, p= 0.000 (95% Cl: 13.2, 269). The confidence intervals for the friends' smoking variables are wide. This shows that there is a lot of variability in my data and I have too small a sample size. As the sample size increases, the width of the confidence intervals generally decreases. The confidence interval can more precisely estimate the value of the population parameter when N is large. (H13) predicted adolescents who receive positive reactions about their smoking from their close friends are more likely to smoke than adolescents who do not receive positive reactions. This hypothesis was supported. After friends' approval of smoking behavior was added to the model, the accuracy of predicting smoking status increased from 58.3% to 62%, p< 0.001 (x2 = 25.1, df = 2), with 15% of the variance in smoking status explained by the friends approval of smoking. Adolescents whose close friends approve of their smoking behavior were 9 times more likely to smoke than adolescents whose friends do not approve of their smoking, p= 0.000 (95% Cl:2.8, 28.8). Adolescents who are not sure whether their close friends would approve of their smoking (or not) are 4.6 times more likely to smoke than those adolescents whose friends do 117

PAGE 132

not approve of their smoking. This result was statistically significant at p< 0.001 (95% Cl: 2.2, 9.4). (H14) Adolescents who perceive higher smoking prevalence among peers are more likely to smoke than adolescents who perceive lower smoking prevalence among peers. This hypothesis was supported but the relationship was weak. After the perceived prevalence among peers was added to the model, the accuracy of predicting smoking status increased from 58.3% to 61 %, p< 0.05 x2 = 5.02, df = 1) with 3% of the variance explained by the perceived prevalence of smoking among peers. As seen in Table 5.9, a one unit increase in perceived prevalence increased the likelihood of being a smoker by 1.2 times. In other words, adolescents who perceive greater smoking prevalence among their peers are 1.2 times more likely to be a smoker than their counterparts, p = 0.027 (95% Cl:1.02, 1.48). E. Section Four: Model Testing In the previous sections the contribution of individual variables to smoking status was examined. In the present section models combined the independent variables to determine their contribution to the prediction of smoking status. Sociodemographic Factors and Smoking Behavior For the first step in building and testing a model of smoking behavior, a logistic regression was run using all of the sociodemographic variables: 118

PAGE 133

school type, age, gender, place of origin, allowances, economic situation and academic achievement. Table 5.10 presents the results for this analysis. The overall model was significant p< 0.001 (x2 = 96.5, df = 14), with 77% of the responses correctly classified. Forty-nine percent of the variation in smoking status accounted for by sociodemographic factors. When the model included all the sociodemographic variables, school type was not significant. On the other hand, place of origin initially was not a significant predictor of adolescent current smoking, but was significant in the presence of other variables. Students who were from the villages were 39% less likely to smoke than students who were from Dursunbey and other cities. The inclusion of the other independent variables in model reduced the error which probably produced this significant result. Place of origin was mildly correlated with gender, economic situation, and academic achievement. To explore the contribution of place of origin, a series of logistic regressions adding the sociodemographic variables into the model one at a time were conducted. When academic success was added into the model, place of origin become significant. The qualitative interviews provide a plausible explanation for this. During the interviews, many adolescents who were from Dursunbey or other cities referred to students from villages as serious students, and they were well aware that their families were using their limited resources to educate them. School counselors also suggested that many of he students who come from villages were poor, so that they would have no money to spend on cigarettes. Indeed, some teachers give money to these students to help them out. In addition, these students are usually shy and their social skills are not developed, even naive. On the other hand, students from urban towns and cities are described being more outgoing. Social skills were previously reported in the tobacco literature as a 119 J

PAGE 134

predictor for smoking, especially for girls (Clayton, 1991; Mitchell and Amos, 1996). Table 5.1 0. Full Model for the Sociodemographic Factors and Current Smoking Behavior Variable p SE Odds 95%CI P-Value6 Ratio School type Religious 0.31 0 .75 1.36 0.32, 5 85 0.68 Health 0 .17 0 84 1.18 0.23, 6.11 0.84 Regular -0.39 0 44 0 .68 0.29 1.62 0.38 Girl's -0.75 1 .35 0.47 0.03, 6.70 0.58 Industrial a Male gender 2.59 0.66 13 3 3.66, 48.6 0.00 Age 0.86 0.27 2.4 1.39, 3.99 0.00 Rural (vs urban) -0.94 0.41 0.39 0.17, 0.87 0.02 Allowances None 1.86 1 .08 6.40 0.77, 53.2 0.09 < 0.38 -1.13 0 .82 0 .32 0.07, 1 60 0.16 0.38-0 77 -0.59 0.83 0.56 0 .11,2. 85 0.48 0.78-2 30 -0.05 0.76 0.95 0.21, 4.25 0.95 2.31-3 83 -0 85 0.82 0.43 0.09, 2.13 0.30 >3. 83 a Economic situation 0.28 0.38 1 .32 0.63, 2.77 0.46 Academic -1.29 0.37 0.27 0.13, 0.57 0.00 achievement Reference point p= Slope coefficient (change in log odds for increase of 1 unit in x SE=Standard Error of f3 95% Cl= 95% Confidence Interval Allowance proved to significantly predict smoking status. Students having a low-end weekly allowance of less than $0.38 were less likely to smoke, compared to students having high-end allowances. Unlike the 120

PAGE 135

medium-level allowances, the smoking rate reduction for low-end allowances was statistically significant. Surprisingly, adolescents who have no weekly allowances are 6.4 more likely to be a current smoker. This contradictory finding needs to be investigated further. Although this finding was not statistically significant students with a weekly allowance ranging from $0.38 to $3.83 were less likely to smoke than students with a weekly allowance of more than $3.83. Finally, gender, age and academic success were strong correlates of adolescent current smoking. To achieve parsimony with the sociodemographic factors model and continue with model building, a backward elimination logistic regression analysis was conducted. The backward elimination resulted in removal of school type and economic situation without reduction in the 76.3% predictive accuracy (x2 = 94.4, df=9, p< .001 ), with 49% of the variation in smoking status accounted for by the remaining demographic variables: age, gender, place of origin, allowances, and academic success. Table 5.11 shows the odds ratios and confidence intervals for the remaining variables in the model. Hosmer and Lemeshow's Goodness-of-Fit test was used to assess whether or not the model was a good fit to the data. The Hosmer and Lemeshow test proved non-significant (p = .53), indicating that the model is a good fit. 121

PAGE 136

Table 5.11. Final Model for Sociodemographic Factors and Current Smoking Behavior Variable p SE Odds 95%CI P-Value6 Ratio Male gender 2.67 0.59 14.4 4.56, 45.7 0.00 Age 0.91 0.26 2.5 1.50, 4.10 0.00 Rural (vs urban -0.92 0 .39 0.40 0.19, 0.85 0.02 Allowances None 1.57 1.04 4.82 0.63, 37.0 0.13 < 0.38 -1.30 0 .80 0.27 0 06, 1.30 0.10 0.38-0.77 -0.68 0.83 0.50 0.10, 2.53 0.41 0.78-2.30 -0 .21 0.75 0.81 0.19, 3.50 0.77 2.31-3 83 -0.97 0.80 0.38 0 08, 1.82 0.23 >3. 8311 Academic -1.26 0.35 0.28 0.14, 0.56 0.00 achievement Reference point bAta. =0.05 p= Slope coefficient (change in log odds for increase of 1 unit in m) SE=Standard Error of p 95% Cl= 95% Confidence Interval Social Environment Factors and Smoking Behavior For the second step in building and testing the model of smoking behavior, a logistic regression was run using all of the social environment variables: parent, sibling, relative, teachers, friends' smoking status, friends' approval of smoking, and prevalence estimates. Table 5.12 shows the full model for the social environment factors and smoking behavior. The prediction accuracy for this model increased from 58.3% to 78% (x2 = 91.0, df = 9, p< 0.001 ), with 47% of the variation in smoking status accounted for by the social environment factors. Exposure to teacher smoking and number of close friends who smoke were the best 122

PAGE 137

predictors of adolescent current smoking behavior. With the introduction of these two influences, the effects of three variables lost their significance: having a relative who smokes, friends' approval of smoking, and perceived prevalence among peers. Table 5.12. Full Model for the Social Environment Factors and Current Smoking Behavior Variable fJ SE Odds 95% Cl PRatios valueb Parent smoking 0 .23 0 34 1 26 0 65, 2.44 0 .49 Sibling smoking 0.01 0.40 1 .01 0.47, 2 20 0 .98 Relative smoking 0.49 0 49 1 64 0 63, 4 29 0.32 Exposure to teacher smoking 1.26 0.37 3.54 1 70, 7.34 0.00 Friend smoking Not sure 2.91 0 .91 18.3 3.04, 109 7 0.00 1-2 friend 1.85 0.80 6.3 1 34, 30.4 0 .02 3-4 friend 3.6 0.81 36.6 7.53 177 7 0 .00 Nonea Friend aggroval of smoking Approve 1.27 0 74 3 .58 0 .84, 15.3 0.09 Not sure 0.76 0.47 2.14 0 86, 5.33 0.10 Don't approve a Prevalence estimate scale 0.04 0.12 1.04 0.82, 1.32 0.75 a Reference point Reference point; 6 At a. = 0.05 ; 95% Cl= 95% Confidence Interval To achieve parsimony for the model of social environment factors, a backward elimination logistic regression analysis was conducted. The backward elimination resulted in removal of four variables: parent, sibling, relative smoking friend's approval of smoking, and prevalence estimate from the model without any reduction in the prediction accuracy 78% (x2 = 86.2, df=4, p< .001 ), with 45% of the variation in smoking status accounted 123

PAGE 138

for by the remaining social environment variables: friend smoking and teacher smoking. Table 5.13. Final Model for the Social Environment Factors and Current Smoking Behavior Variable f3 SE Odds 95% Cl P-va/ue6 Ratios Exposure to teacher 1.23 0 35 3.4 1.72, 6.85 0.00 smoking Friend Smoking Not sure 3.23 0.88 25.2 4.53, 140.6 0.00 1-2 friend 1.98 0 78 7.22 1.57, 33.3 0.01 3-4 friend 4.02 0.78 55.9 12.1' 257 2 0.00 None a a reference point; 5At a. = 0.05; 95% Cl= 95% Confidence Interval The Hosmer and Lemeshow test was used to assess the model's goodness-of-fit to the data. The test proved non-significant (p =.99), indicating the model was a good fit to the data. Table 5 13 shows the final model for the significant social environment variables and smoking behavior. Sociodemographic and Social Environment Factors After exploring the effect of sociodemographic and social environment variables separately, I wanted to test the effect of all the significant variables from the previous models. To accomplish this, the demographic and social environment variables were entered into logistic regression simultaneously. The overall model was significant at p<.001 (-/ = 124.5, df = 13). The model predicts 81% of the responses correctly. The adjusted R2 indicated that 60% of the variance was explained by the final 124

PAGE 139

model. Table 5.14 presents the final model for sociodemographic and social environment variables predicting current smoking behavior. Table 5.14. Final Model for Predicting Adolescent Current Smoking Behavior Variable fJ SE Odds 95% Cl P-Ratios va/ueb Male gender 2.22 0.659 9.18 2.52, 33.4 0.001 Age 0.64 0.279 1.90 1 09, 3.28 0.022 Rural (vs urban) -0.79 0.437 0.45 0.19,1.07 0.071 Allowances None 2.52 1.157 12.4 1.28, 119.4 0.030 < 0.38 -.406 0.834 0.66 0 13,3.42 0.626 0.38-0 .77 140 0 893 1 15 0.20,6.63 0.875 0.78-2.30 .590 0 787 1.80 0.37, 8.43 0.453 2.31-3.83 -.261 0.825 0.77 0.15,3.88 0.752 >3.838 Academic achievement -.882 0.386 0.41 0 19, 0.88 0.022 Exposure to teacher smoking 1.06 0.404 2.90 1.32, 6.41 0.008 Friend smoking Not sure 2.62 1.02 13 7 1.87, 101. 2 0.010 1 2 friend 1.81 0.849 6.09 1.15, 32.2 0.033 3-4 friend 3.24 0 868 25.5 4 65, 139.8 0.000 None a a reference point bAt a. = 0.05 95% Cl= 95% Confidence Interval Gender, friend, and teacher smoking, age and academic success contributed significantly to the model. Allowance and place of origin had a mild relationship with adolescent smoking, but were not significant at an alpha level of 0.05. 125

PAGE 140

The Hosmer and Lemeshow test was run to test the significance of the model. The test was not significant (p=.96), indicating the model fit the data. Based on the "compiled" analysis results, it can be concluded that the factors which most accurately predict adolescent current smoking are gender, age, teachers' smoking, friends' smoking, and academic performance. Other Survey Findings In this section descriptive information about the adolescents' smoking behavior will be presented. Table 5.15 presents the companions accompanying adolescents when they first tried smoking. This often occurs in a social or environmental situation, such as at a picnic or friend group gathering where friends offer them cigarettes, or in a stressful situation where smoking is attempted as a quick and easy coping strategy for stress reduction. Table 5.15. Companions When Adolescents Try Smoking Companions Frequency Friends from school 65 Other friends 58 Alone 31 Relatives 11 Mother I father 7 Others 2 Total8 174 a 35 students never tried smoking and 2 missing values. Percent% 30 8 27.4 14.7 5 2 3 3 0.9 82.4 Adolescents were also asked in the survey "Why do you think people of your age start to smoke?" and 'Why did you start smoking?" Although these questions were not part of the hypothesis testing, it is still helpful to 126

PAGE 141

review their descriptive statistics. Table 5.16. Why do you think people of your age start to smoke? Reason FrequencY' Percent % Their friends smoke 154 73.0 Curiosity 146 69.0 Look cool 110 52.0 Look grown up 104 49.2 Relaxation 81 38.3 Look important 73 34.5 Friend's pressure 63 30. 0 Their parents smoke 52 24.6 Something to do 46 21.8 Rebel against teacher and parents 18 85.3 Their siblings smoke 11 5.2 a Since students could mark more than one answer column total does not add to 211. Table 5.16 and 5.17 presents the answers to these questions in the descending order of frequency cited. The most commonly cited reason (73%) by the adolescents for their peer's smoking was "They have friends who smoke." Curiosity (69%), looking cool (52%) and looking grown up (49%) were also commonly cited reasons for starting to smoke. As was reported in earlier studies, smokers in the current study often seemed to view their own behavior and that of others differently (AIIbutt et al., 1995). When students were asked about their own smoking, curiosity was the most commonly (56%) mentioned reason for smoking initiation Friend (38%) and family (28.4%) influences were also mentioned. Modeling after friends (16%) and adults (10.4%) was also part of the reason for initiation. 127

PAGE 142

Table 5.17. Why did you start to smoke? Reason Percent % Curiosity 118 56.0 Encouragement from friends 80 38.0 Family members smoked 60 28.4 Look impressive 41 19.4 Imitate friends behavior 33 15.6 Imitate adults behavior 22 10.4 Rebel against parents and teachers 20 9.5 Gain acceptance by friends 18 8.5 a Since students could mark more than one answer column total does not add to 211. These reasons reported by the adolescents support my hypotheses on interpersonal influences and modeling behavior and give us clues about the future interventions for youth smoking prevention. 128

PAGE 143

CHAPTER6 DISCUSSION A. Introduction This dissertation has examined the social environmental and sociodemographic influences on youth smoking in Dursunbey, Turkey In doing so, it tested fourteen hypotheses using logistic regression analysis of cross-sectional survey data. The smoking status of adolescents was measured by the past 30 days of smoking prevalence and was referred to as "current smoking". This measure of smoking status was studied as a function of the influences of parents, siblings, relatives, close friends, teacher smoking, and adolescents' age, gender, place of origin (urban/rural), type of school, perception of their academic achievement, economic situation and perceived smoking prevalence among peers. Ten hypotheses were initially confirmed and four were not; of the confirmed ten, three appeared to be confirmed when considered in isolation, but were later rejected after control variables were included in the regression. In this section, I further illuminate the details of these influences, and will explain the discrepancies between some of my findings and previous studies. Throughout, I discuss the implications of these findings for Turkey, and then theoretical implications of the findings. Finally, I raise sampling and other methodological issues at the appropriate points in the discussion, and identify the limitations inherent in this study. 129

PAGE 144

The qualitative study conducted in concert with the survey provides a useful source of data for explanation of some of the survey results. The qualitative study used semi-structured in-depth interviews of a subset of the survey population (n=42) and students from other grades (n=10). Its purpose was to understand smoking behavior from adolescents' own perspectives and words. I also interviewed principals, teachers, school counselors and adults from the town (n=18). The results of this qualitative study are also reported elsewhere, in an interim report submitted to its funding institution (Yuksel, 2003). I use these qualitative findings to interpret my survey results and gain a deeper understanding of these influences. B. A Review of Findings Incidence of Ever-Smoking The ever-smoking rates are high among 1Oth graders (83% ), with 93% of boys and 63% of girls reporting that they had already tried smoking. An initial comparison of these rates with Erbaydar et al. (2002) shows that Dursunbey's ever-smoking incidence is a good deal higher than the general ever-smoking incidence for youth in Turkey, reported at 58%3 with 66% and 53% for boys and girls, respectively. The Erbaydar study represents the most recent nationwide data on youth smoking currently available (15 provinces were surveyed representing a wide array of geographic locations). It was conducted virtually simultaneously with my study, with its closest study site being Ayvalik, another district in Balikesir province in which 3 Calculated by combining Erbaydar et al.' s 15 year old and 16-17 year old ever-smoking rate statistics. 130

PAGE 145

Dursunbey is located. In an earlier study, Bilir et al.'s (2000) nationwide study among 10th graders showed ever-smoking rates of 32% overall, with 64% for boys and 22% for girls4 Incidence of Current Smoking In my study, I classified subjects as "current smokers" if they smoked a cigarette within the past 30 days. Forty-two percent of the adolescents were current smokers, 55% of boys and 9.5% of girls. The daily smoking rates in this study population was 26% (35% for boys and 3.2% for girls). The daily smoking rate for boys in my study is comparable to the 33% daily smoking rate in Erbaydar et al.'s study (2002) and the (32%) daily smoking rate in Bilir et al.'s study (1997). However, daily smoking rates for girls in Dursunbey are much lower (3.2%) than larger size cities in Erbaydar et al. (2002) study and Bilir et al. (1997) study (18.3% and 20% respectively) in Turkey. This measure might again be explained by Dursunbey's smaller size, where informal and traditional relationships are still in effect compared to larger cities in the Erbaydar et al. and the Bilir et al. studies. Traditionally, Turkish society disapproves of women smoking unless they are professionals In such a small-town culture, the traditional pressures against non-professional women smoking are more easily exerted. Influence of Sociodemographic Factors Gender. A comparison of Bilir et al. (2000), Erbaydar et al. (2002) and my study demonstrates a large increase in the ever-smoking rate for 4 Calculated by subtracting Bilir et al.'s non-smoking rates from 100%. Even i f Bilir et al.' s non-smokers include some ever-smokers, the ever-smoking percentages calculated here represent the maximum possible ever-smoking rate derivable from his statistics. 131

PAGE 146

girls over a period of two or three years. Given that the tobacco advertisement ban took place in Turkey in 1996, one might assume that smoking rates would decrease between 2000 and 2002. Nevertheless, it seems likely that girls are becoming more curious about smoking and are experimenting at high rates. This might increase the current smoking rates for girls in the near future. This conclusion might be disputed because these study populations differ by age, geographic location, and development level. As shown in the results of this dissertation, many girls leave school after compulsory education of 8 years, skewing results for girls between studies based on grade level and those based on age. The geographical contention is supported by both the Erbaydar and Bilir study results, which show large variations in smoking rates among differing Turkish communities. Both the Erbaydar and Bilir studies are nationwide, however, making it less likely that such a large shift in girls' ever-smoking rates (from 22% to 53% between the two studies) is due to geographic sampling differences. However, both Erbaydar et al. and my studies (2002) show a much higher rate of ever smoking among girls at the 1Oth grade level than in Bilir et a I.'s 2000 study. Taken together, the Dursunbey, Erbaydar et al. and Bilir et at. studies make the increase in girls' ever-smoking rates a credible possibility that deserves further investigation. It is difficult to explain the remaining 8% to 10% Dursunbey rate differential for boys by comparing the Dursunbey survey statistics to national statistics, given the known geographic variability of smoking rates within Turkey. Here the qualitative data provide additional information. Interview subjects spoke frequently of having little to do with themselves except smoke. This may suggest that the small size of the community and lack of recreational alternatives might lead to experimentation with cigarettes. 132

PAGE 147

Gender explained 25% of the variability in smoking behavior. Other factors discussed in this section frequently involve gender, and are best reviewed in the discussions of those factors. Age. The positive relationship between age and the current smoking of adolescents was confirmed. The likelihood of being a smoker was increased (OR=2.9) by in increase in the age. Although the regular age group for 1Oth graders is 16, my study included students who are 18 and 19 (N=30), mostly due to students who repeat the same grade I have already noted the positive skewing effect of these older students. This finding is also consistent with the previous literature in Turkey and other countries. School Type. Schools are an important context in which to investigate adolescent cigarette smoking, since they constitute a large part of students' social environment. During interviews, the industrial vocational high school was identified as the smoking school by the principals, counselors, teachers and students of other schools. This made it a good reference point for logistic regression analysis of school type, since the odds ratios of all the other schools would probably be less than 1.0, making the interpretation of the results easier. Thus, the industrial high school was used as the basis for comparison of all the other high schools in Dursunbey. Ozcan & Ozcan (2002) found a significant effect of school type among high school students in Ankara, Turkey. According to their study results, students from industrial girls' schools and commercial high schools were more likely to smoke (OR=1 .6, 2.4, and 2.7, respectively) compared to private high schools. When considered in isolation from other independent variables, school type was found to be a significant factor in student smoking. Adjusted for gender, however, this factor was found to be non-significant. Smoking 133

PAGE 148

rates varied among the schools. The girls' vocational school had the lowest smoking rate (% 14, n=1 ), but this difference was not statistically significant due to in part the small sample size (n=7). The health vocational school had the second lowest smoking rate among the five high schools (23%, n=5). This result was statistically significant before controlling for gender (p < .05), but disappeared after gender was added to the regression. The religious school had the highest smoking rate, but once again its significance disappeared due to its small sample size (64%, n=9). The great majority of the religious school's students were males. Having mostly male students was reported as a problem by the school counselor, since these boys lack the opportunities to associate with the opposite sex and develop confidence instead they try to prove themselves by smoking. The health vocational school counselor and principal mentioned the policy change that took place in 1999 that reduced the importance of vocational schools. This resulted in a reduced chance of getting into college for vocational school students nationwide; "right now students of vocational schools have no hope for their future." According to the principal, the last student from the health vocational high school to receive a job graduated in 1994. This lack of possibilities for their students is demoralizing for the teachers themselves. This was a common sentiment expressed among the interviewed educational professionals. In Dursunbey, all of the high schools had non-smoking policies for students, and some had a designated smoking room for the staff. However, it was not uncommon, even during the survey administration, to see teachers in the school halls and stairs smoking during their breaks while they were watching the students. Many teachers and principals raised this issue, stating that they cannot effectively deliver anti-smoking messages because they themselves smoke. This makes whatever they say less 134

PAGE 149

believable since students usually challenge their smoking by saying: "If it is this bad, why do you smoke yourself?" Furthermore, many students reported that they smoke on school property in the restrooms or corners where school staff cannot see them. Although students are not supposed to leave the school property, some of them leave the school campus during the 10 minute break to smoke outside. New legislation on school non-smoking policies further limits school authority to school campuses. Formerly, schools were allowed to enforce non-smoking policies off-campus, even to the point of making night visits to coffee houses and other public places. During my interviews with the principals in some of the schools, I observed school administrators lack of commitment in dealing with smoking problem since they do not want to confront students, and do not want to alienate students even more. Additionally, one of the principals reported that some parents made complaints that they send their children to school for education, but instead teachers focus on their smoking. Dealing with student smoking issues was clearly a sensitive topic from the school administration's point of view. In general, principals and teachers felt unarmed in dealing with the issue. One of the principals pointed to the new government policy that decreases the punishment of smoking on school grounds, "If our government allows students to smoke, what can we do about it?" On the other hand, I witnessed the health vocational high school principal's strong anti-smoking rules on school property. This principal stated that he was motivated to strictly enforce anti-smoking policies because health was the subject area of his school, and his students should set an example to other high schools. Many students in this school indicated that they feared the principal's strong rules and his commitment to enforce them. As a result, they do not bring their cigarette packs to school. Some of the 135

PAGE 150

students mentioned leaving their cigarette packs with market owners right outside the school or hiding their cigarettes, since they are regularly searched for cigarette packs in school. I suspect that due to this fear, some of the health high school students underreported their smoking. In the health vocational school, we have an instance of strict on campus policy enforcement combined with no significant difference in smoking rates. The industrial high school was also diligent in enforcing student non-smoking policies, but also had no significant gender-adjusted difference in smoking rates, measured against other schools. However, not having a statistical difference does not mean that these policies are ineffective in reducing or preventing smoking among adolescents. These policies are still helpful to create anti-smoking norms. The quantitative and qualitative data indicate that two actions might reduce student smoking rates. First, since other studies indicate that total bans on teacher smoking actually increase student exposure to smoking in school yards, efforts should focus on teacher cessation programs, as well as tobacco education programs to inform teachers about their high degree of influence due to role modeling and sending mixed messages. Second, schools have been disarmed in combating tobacco off-campus. This lack of authority can be compensated to some degree by strict enforcement of current tobacco legislation. Market owners, for example, should not be selling to minors, much less keep cigarettes for them so that they can evade school searches. Future interventions should encourage parents and school administrators to demand the enforcement of current laws. Urban/Rural Place of Origin. Turkey is in transition from a rural to an urban lifestyle. As a result, rural-urban differences and modernization have been the central topic of sociological research in Turkey. Therefore, I 136

PAGE 151

wanted to see if there was any difference based on type of community in adolescent smoking behavior. The urban and rural origins were equally represented in the study population. This allowed me to compare the adolescents' smoking behavior based on their origin. There was no association between place of origin and adolescent smoking when individually tested. However, in the presence of other variables in the model place of origin approached significance (p= 0.071 ). On a larger scale, ever-smoking rates in Dursunbey were found to be among the highest in Turkey, compared to the rates found in Erbaydar et al.'s (2002) study, reflecting a pattern of higher rural ever-smoking rates found in the developed world by Harrell et al. (1998), Stanton et al. (1994), and Conrad et al. (1992). This was true for both boys and girls. In other parts of this discussion, I have linked these higher rates to a lack of recreational alternatives in a small-town environment. Current smoking rates for boys in Dursunbey were comparable to the Turkish national average for that age group, however the rate of current smoking was much lower for girls. Earlier in this discussion, I linked this phenomenon for girls to the small-town environment where traditional values against girls' smoking are more effectively discouraged compared to larger cities. Economic Situation and Weekly Allowance. The relationship between adolescents' perceptions of their economic situation and smoking behavior was worth investigating. In many countries cigarette use is reported to be more common among the low-income level than high-income level. However, in Bilir et al.'s (1997) study in Ankara, when the income level increased, the smoking rates increased also. In my study however, 33% of students in a poor economic situation were smokers (Table 5.6), a rate similar to that for students whose economic situation was good. Adolescent 137

PAGE 152

perception of their economic situation failed to predict adolescent smoking. Two factors suggest a possible explanation for this finding. The first is that awareness of smoking issues in Turkey is not differentiated by social class or professional employment, as indicated by Bilir et al. (1997) finding of high smoking rates among doctors, teachers, politicians and journalists. The second factor is documented earlier in the preliminary study, that cigarette brands are actually used as status symbols, an influence that apparently affects the low and high-income youth of Dursunbey equally. These factors indicate that awareness of the importance smoking issues is more or less equally absent among all levels of SES in Dursunbey, leaving affordability as a possible economic differentiator of smoking behavior among the student population. On the other hand, the weekly allowance of adolescents predicts the adolescent's current smoking behavior only for students who receive less than $0.38 weekly and for the adolescents who receive $0.38 to $0.77. Adolescents in both groups were less likely to smoke compare to the adolescents who receive more than $3.83 weekly. Although it was not significant at alpha .05, there was a trend that students who received $0.78 to $2.30 and $2.31 to $3.83 weekly were less likely to be current smokers than students who received $3.83 weekly. This finding was consistent with the previous smoking literature. The reviewed literature also showed an increased risk with increased allowance (Erbaydar et al., 2002; Risse! et al., 2000). Academic Achievement. Survey findings showed that academic achievement was a predictor of adolescent smoking, with students having lower achievement being at greater risk of smoking. Adolescents who perceived their academic performance as poor were 3 times more likely to 138

PAGE 153

be a smoker than the adolescents who perceived their academic performance as average, and students who perceived their academic performance as average were 3 times more likely to be a smoker than students who perceived their academic performance as good. Erbaydar et al. {2002) found a similar relationship for the boys; those with low and moderate academic achievement were more likely to be smokers than students who reported very good school success. On the contrary, Erbaydar et al. {2002) also found that the very good students were at greater risk of smoking than students who were merely good. Previous research also supports my findings. Newcomb et al. (1989) found that academic orientation is a central organizing principle influencing adolescent smoking. Nutbeam et al. (1993) listed three components of adolescent "alienation" and feelings of powerlessness leading to smoking: low academic achievement, positive/negative perception of school, and future plans for education. This was echoed in my qualitative results, where principals and teachers often stated that the low achievers were most likely to be smokers. I had measured two of Nutbeam et al.'s three components of alienation, having asked in my survey about students' future plans for taking the Turkish national college entrance exam. Even though this variable was not incorporated in my original hypotheses, I ran a chi-square test for this variable, followed by a logistic regression analysis. The risk of current smoking was negatively associated with plans to take the college entrance exam (OR= 2.1, Cl = 1.2 to 3.7, p-value= .011 ). Rural and small-town students face formidable barriers to attending college. Nationwide only 8% of the students taking the exam are accepted into college. To have any chance of passing the college entrance exam, a student must take special (and expensive) preparatory courses that many families cannot afford. Small school districts are generally not funded or 139

PAGE 154

I supported at the level of schools in larger metropolitan areas, such that most Dursunbey families who can afford it send their children to private or Anatolian high schools in Balikesir or lzmir. Thus, the attendee of a Dursunbey high school starts his or her high school career knowing that the odds of going to college are small. In my graduating year of high school, for example, an usually high number of 5 Dursunbey students were accepted to college. In a typical year, 1 is accepted. Given these facts, it is appropriate to question the role of Turkish educational policy and education resources in creating a sense of powerlessness in its students. This question is supported by my qualitative study: one of the most common statements heard in my interviews was that "I don't have a future." Based on the literature cited in this section, one can conclude that Turkish education policy itself might be a contributing factor in Dursunbey's high adolescent smoking rate of 42%. Influence of Social Environmental Factors This study provides information about how various elements of smoking exposure through interpersonal influences (i.e. peers, close friends, school, parents) are related to the risk of current cigarette smoking among adolescents. Adolescents are embedded in a set of social networks that include relationships with family, close friends and peer groups. Furthermore, how adolescents perceive their social environment might be as important as the actual environmental factors. As a result, adolescents' perception of norms about smoking and social support for their smoking will also be interpreted. 140

PAGE 155

Family Influences. The study results show that adolescents are exposed to high amounts of smoking in their social environment. Sixty-six percent of the students reported that their fathers smoke, a rate greater than the male adult smoking rates in the general population in Turkey (59%). On the other hand, mothers' smoking was reported as 8.1 %, which is lower than nationwide average for females of 26.4% In smaller cities like Dursunbey, and in rural areas, women usually stay at home and assume the traditional roles of mother and homemaker, and would face social disapproval if they smoked. Even though this study did not measure the adult female population for professional employment, first-hand knowledge allows me to state confidently that this is the case. Housewives also face the typical impediment of not earning money outside the home, so that many have little money pocket money to spend on cigarettes. These facts of life in Dursunbey largely explain the low smoking rates among the mothers reported in the study. My study showed virtually no difference in adolescents' smoking status between households where neither parent smokes and households where one or both parents smoke. In contrast, Erbaydar et al.'s study (2002) found a relationship between parent smoking and risk of being a current smoker, and girls were more influenced than boys by their mother's smoking behavior. Similarly, boys were influenced by their father's smoking behavior more than girls. However, children's role modeling of parents' smoking behavior may be particularly weak in districts like Dursunbey. In my qualitative study, a number of students mentioned the lack of education of their parents and an educational gap they perceive between their parents and themselves. The current generation of adolescents is growing up with a rapid adoption of Western values and media exposure to Western culture. This may exacerbate their perception of the gulf between their parents and 141

PAGE 156

themselves, making the parents poor role models for developing adolescents as they claim their own independent identity. There is more to family influences, however, than parental smoking statuses and role modeling. In my qualitative study, a number of interviewees (particularly the girls) stated that they were under stress due to pressure from their families. Several of these students linked this stress to communication problems with their uneducated or illiterate parents. Pandina & Schuele (1983) found that overbearing parental restraints on adolescent behavior lead to smoking. Based on this qualitative lead, I performed a chi square test on my survey database, finding a significant correlation (p < 0.001) between stress due to family pressure and smoking status. I also found a significant (p <0.001) correlation between stress due to "not being allowed to have friends of the opposite sex" and smoking status. These facts indicate that generational education gaps may be creating a "culture shock" to the families of rural Turkey, producing greater risks of smoking for adolescent girls. It must be noted, however, that I did not find a significant statistical relationship between parents' education level and smoking I also did not find a link between parents' education level and stress due to family pressure. Nevertheless, if ever-smoking rates are increasing for girls, a fuller investigation of these potential effects is warranted. Social pressure on girls is a commonly cited reason for girls to smoke as resistance to social pressure from families. Some of my interview subjects stated that girls who smoke in public would be considered loose and of low moral character. This finding is identical to that found by Nichter et al. (2002) when studying U S. Filipino and Mexican American subjects: "Smoking would bring shame on their parents as others would think that they had not been raised properly." On the other hand, families are somewhat resigned to their boys eventually smoking. For males, it is an 142

PAGE 157

expected right of passage. If men do not smoke and drink, they are usually described as effeminate by society-at-large: "He is like a girl. He does not smoke and drink." There are some cultural differences between Turkey and Western cultures in the perception of smoking. In Turkey, it is culturally considered to be inappropriate to smoke in front of older family members. This is the same for boys and girls. This norm often masks health as a reason not to smoke; you should instead avoid smoking out of respect for your elders. There are even some "modern" families who allow their children to smoke in front of them. Turkish smoking prevention programs should address this issue and emphasize the harmful effects of smoking over simply being disrespectful to family elders. One of my informants told me a story about how her brother started smoking. After their parents discovered his smoking and terminated his allowance, the brother turned to his grandmother, who did not take his budding habit seriously. The grandmother funded his habit and allowed him to smoke surreptitiously in her home. This story reflects how cigarettes are often perceived as benign toys for boys that are frequently accommodated. Families have other influences on their children, as stated by the U S Surgeon General (1994). They can be a positive influence, discouraging smoking, discussing health issues, and encouraging involvement in school. In my study, 44% of adolescents reported that their parents know that they smoke. On the other hand, 70% of adolescents reported that both of their parents disapprove of their smoking, and only 8% of the students indicated that both of their parents approve of their smoking. There was no significant association between parental approval of smoking and adolescent smoking What makes these figures interesting is the qualitative data. During my interviews, principals and other adults reported that, yes, smoking is not good for adolescents but "at least they are not into drugs or drinking like in 143

PAGE 158

the big cities." This made me realize that parents' disapproval of their children's smoking is often superficial. Knowing that they cannot follow their children everywhere, they usually pretend not to know that their children smoke. Smoking is still perceived as somewhat benign compared to mind altering drugs like alcohol. After all, everybody does it. It is clear that parents lack the skills required to deal with sensitive issues. This demonstrates a need for interventions offering parenting classes that teach parents the tools for dealing with smoking and problem behavior in their children. There was also no association between adolescents' current smoking and the smoking status of siblings. Twenty-nine percent of the students indicated that they have siblings who smoke (55% of the nonsmoking students and 45% of the smoking students). One possible explanation for this lack of influence might be the traditional emphasis in Turkish families that older siblings should be good role models for their younger siblings. Older brothers and sisters actively try to prevent their younger siblings from getting into trouble, even to the extent, as was commonly mentioned in interviews, that "their elder brothers or sisters smoke, but still they hide their smoking from them." For this reason, adolescents might not associate siblings with smoking. There was, however, one significant family influence on smoking. Having a smoker in the household other than a nuclear family member was significantly correlated with current smoking of the adolescents. For the ease of interpretation, after this point I will refer to these other people in the household as relatives. Other household members may include grandparents, uncles, cousins, and nephews who stay with the family for their education or work. It is common that young relatives who live in the villages come to Dursunbey to work or to go to school and stay with their relatives in the town. It is highly possible that these relatives are similar in 144

PAGE 159

age to the adolescents in the household and that they might associate together in social activities. The dynamic of their relationship would be different from the sibling relationship. Although there is less consistent support for the influence of parent smoking and approval (Conrad et al., 1992; Chassin et al., 1986) than the influence of peer smoking in the tobacco literature, future investigations should specifically explore the question of why family influences do not predict adolescent smoking outcomes in this study population. A study dedicated to exploring family influences in greater detail might find other factors that my survey study did not reveal. Research employing other measures of family influences such as family bonding (Hirschi, 1969; Elliot et al., 1983), family functionality (Diana, 2000), and family communication patterns (Ritchie, 1991 and Tims & Masland, 1985) might be useful in future studies. Influence of Friends. The effect of having close friends who smoke was the most significant predictor of adolescents' current smoking. The study results indicated a strong positive relation between the number of their close friends' smoking and adolescent's current smoking. This relationship was significant even when controlled for gender and age. This result is consistent with the existing literature documenting the strength of the association between adolescents' current smoking and friend smoking (Erbaydar et al., 2002; Poulsen et al., 2002; Alexander et al., 2001; Wang et al., 1999) Having at least 1 to 2 smoking friends increased the likelihood of smoking by eight times compared to adolescents who have no smoker friends. Having 3 to 4 smoker friends increased the likelihood of smoking by sixty times. This result was potentially confounded by effects of friendship selection. From the cross-sectional survey data, I was not able to determine 145

PAGE 160

if this finding was a result of friend influence or rather that smokers selected each other as friends. In future studies close friends effects should be longitudinally assessed. According to the study results, adolescents who reported that they are not sure how many of their close friends smoke (n=19) were 23 times more likely to be smoker It was difficult to interpret this finding since number of smokers and nonsmokers were equally divided in this group. During the analysis I realized that this problem was caused by the design of the question, since I allowed the students to answer not sure. I assume that some of the students did not want to take the time to answer this question carefully, or found it challenging (e.g., who qualifies as "a friend"). Although survey design practice requires providing neutral options, I would not allow the not sure option in the future studies where the likelihood of not knowing is small. Friend influences were the most pointedly mentioned during the in depth semi-structured interviews in the context of emotional coping mechanisms Some boys mentioned that they were taunted by friends into smoking with the phrase, "sOt kuzusu" literally milk lamb, or "so young that you're still drinking mother's milk." Some boys stated that smoking is equivalent to being a grown man among young boys. Furthermore, the thrill of hiding smoking activities was said to be a peer group experience. Another reason cited for smoking in groups of friends was passing the time; adolescents stated that they felt like they were doing something while they smoke. Teacher Influences. An important finding to emerge from this study is the positive association between teachers smoking in the school setting and adolescents' smoking behavior, with an odds ratio of 3.6 (CI = 2.2 to 6.4, p < 146

PAGE 161

.001 ). Forty-seven percent of the students reported observing their teachers smoking everyday. Teacher influences have never been identified as a predictor of student smoking in previous studies in Turkey. From earlier studies, however, we know that smoking rates have been measured for male teachers, from 47% to 62%, and for female teachers, from 34% to 54% (Bilir et al., 2000; Bililr et al., 1997;Saltik et al., 1989). A striking finding from the perspective of tobacco control researchers in the US is that 68% of smoking teachers indicated that they smoke in front of the students, 37% while working and 81.3% during their breaks in the work place (Bilir et al., 1997). Having a teacher who is a smoker might make adolescents more likely to perceive smoking as something positive and acceptable, developing favorable personal beliefs. In addition, most of the families in this study population have an elementary school education or less. As a result, adolescents might feel closer to their teachers, rather than parents, making them more powerful role models. Alternatively, students who smoke might be more aware of teachers who smoke, which might account for some of the association between adolescent smoking and exposure to teachers smoking. This alternative would be a more benign explanation of the survey data, in the sense that teachers' smoking would not cause adolescent smoking. The qualitative interviews offered more information on teacher smoking. Many students mentioned the mixed messages they receive from their teachers. On one hand, teachers tell them not to smoke, but on the other hand, the teachers themselves smoke. A health vocational school counselor and some of the teachers from regular high school mentioned that some teachers send their students to the market to buy cigarettes for them. The school counselor quoted the students' view on this: "On the one hand, 147

PAGE 162

you send us to buy cigarettes for you, on the other hand you tell us not to smoke." It is obvious that any kind of anti-smoking message teachers deliver would be undermined by their own smoking behavior. Some students even mentioned bumming cigarettes from a particular teacher. Students perceived this particular teacher as very understanding of their smoking behavior. As mentioned under school influences above, the most effective means of minimizing this harmful influence is cessation programs for teachers. Perceived Environmental Factors. Two variables were tested under perceived environmental influences, close friends' approval of smoking and perceived smoking prevalence among peers. Close friends' disapproval of smoking was associated with lower rates of current smoking. Adolescents whose close friends approve of their smoking behavior were 9 times more likely to smoke than adolescents whose friends do not approve of their smoking (p < .05). The adolescents who were not sure whether their close friends would approve of their smoking were 4.6 times more likely to smoke than those adolescents whose friends did not approve of their smoking (p < .001 ). This result is consistent with the existing literature (Kandel, 1980; Rose et al., 1999). Like the frequency of friends' smoking, it can be argued that this measure of friends' approval of smoking was confounded by effects of friendship selection. This effect disappeared in the presence of other variables, i.e., gender and the number of close friends who smoke. The perceived prevalence of smoking was not a strong predictor of adolescents' current smoking, having only explained 3% of the variance in the smoking behavior. Adolescents who perceive greater smoking prevalence among their peers are 1.2 times more likely to be smokers than their counterparts (p < .05). When the model was adjusted for friend and 148

PAGE 163

gender influences, this variable did not contribute at all. One explanation for the weak association between this variable and smoking behavior is that both smokers and nonsmokers perceived high smoking prevalence among their peers Many of the adolescents (50%) perceived unrealistically high smoking rates among their peers. It was a common sentiment during the in depth interviews that "everybody smokes." This might lead adolescents to think that smoking is a normal and acceptable behavior in society. Although current smoking rates are high in my study population, it would be still helpful to explain to adolescents that not everybody smokes and that 58% of the study population is nonsmoking. C. Theoretical Implications of the Findings Bandura's Social Cognitive Theory provides a useful framework for understanding smoking behavior. My goal was not to test Bandura's theory, but to use it as guide to organize my data and make sense of it. The variables I have tested can be categorized under three of his constructs (environment, situation, observational learning). I also should note that his constructs overlap and the divisions between them are sometimes seem arbitrary. He described the environment as anything that is external to the individual. Since it would be very difficult to measure everything external to the individual, I measured the interpersonal factors under environment. Although Bandura does not use the term interpersonal factors, it is commonly used in tobacco literature and it fits under his environment construct. I operationalized the environment construct by using the interpersonal factors of parent, sibling, relative, teacher, and friend smoking The mechanisms of role modeling and social learning that underlie 149

PAGE 164

the processes of interpersonal influences can be categorized under the observationalleaming construct of Social Cognitive Theory. Social Cognitive Theory views individuals in terms of the reciprocal effect of the person behavior, environment-person, behavior-environment interactions. Bandura sees individuals as both product and producer of their environment. This is Bandura's most significant contribution, since he is trying to describe the interaction of personal characteristics with social influences and physical structures within the environment in a dynamic process. Since Bandura is a psychologist, however, he does not discuss structural and cultural issues such as norms. Bandura does not talk about the socialization process, where these norms become embedded into the individuals' lifestyle. Bandura talks about a person's perception of his or her environment (i.e., situation construct) as an important factor. Although he does not discuss the mechanisms that lie behind the process of how people establish these perceptions, his situation construct could accommodate norms as perception of the social values. Therefore, perceived smoking prevalence among peers and perceived approval of smoking behavior by close friends can be placed under his situation construct. In the previous literature there is an association between adolescent smoking and the perception of high smoking rates among their peers. This association is also documented for perceived support for smoking from close friends {Kandel, 1980; Rose et al., 1999). As part of an intervention project, Bandura recommends correcting these misperceptions of the environment to promote healthful norms. This recommendation has been widely used in youth smoking prevention programs as the social norms approach, in which young people are told about the actual smoking rates in their community and among their peers. 150

PAGE 165

In the present study, the social cognitive model of social influences and sociodemographics explained 60% of the variance in smoking behavior Figure 6.1 displays the final model from the logistic regression analysis. Figure 6.1. Predictors of Adolescent Current Smoking Gender Female Male Person-Behavior Male OR= 9.2 Academic Achievement Increased risk by poor academic achievement OR = 2.5* *Inverse of the 0.4 Teacher Smoking See teacher smoking everyday Environment Behavior OR= 2 9 Smoking Status Smoked within last 30 days Age Increased risk per year Person Behavior OR= 1.9 per year Number of Friends Smoking None 1-2 friends 3-4 friends Environment Behavior Previous studies have demonstrated that role modeling (parent, sibling, friend, and teacher smoking behavior) is associated with adolescent smoking (Alexander et al., 2001; Flay et al., 1998; McNeill et al., 1988; Poulsen et al., 2002; West et al., 1999; Zhang et al., 2000). My study supports this relationship to smoking only with smoker friends and exposure to teacher smoking (environment-person interactions where the environment 151

PAGE 166

influences the person through role modeling, followed by person-behavior interactions where the person expresses smoking/non-smoking behavior) Support was also found for a relationship between smoking behavior and academic achievement (a person-behavior interaction). Finally, person behavior interactions were found between the biological personal factors of gender/age and smoking behavior. Figure 6.2 illustrates the reciprocal interaction between the significant study variables. Figure 6.2. Reciprocal Interaction of the Predictors of Adolescent Current Smoking (teacher and friend Person (gender, age, academic achievement) Behavior (current smoking) There was no support for role modeling of parents, or siblings influencing adolescent smoking behavior. Furthermore there was no support for the theoretical relationship between adolescents perception of their environment (environmenUsituation) and adolescents' behavior (current smoking). This finding was inconsistent with other's findings of a positive correlation between smoking and adolescents' perception of smoking rates among peers and between smoking and the perceived support for smoking (Conrad et al., 1992; Flay & Phil 1985). The relationship between school type/place of origin (environment) and current smoking (behavior) was not supported. The perceived economic situation and weekly allowance were served as proxy variables to test the socioeconomic status (SES) In the 152

PAGE 167

final model there was no relationship between these variables and adolescent current smoking behavior. Even though Bandura does not discuss SES as part of his model, I argue that SES can be successfully included and examined as part of environment-person and person-behavior interactions. Because SES variables portray a good example of his triadic, dynamic and reciprocal interaction of the environment, personal factors, and behavior. In Figure 6.3, I enhance the model represented in Figure 6.2 by incorporating theoretical constructs and a variety of variables from the qualitative results. To develop the model in Figure 6.3, ethnographic information obtained in the qualitative study is included with variables gathered in the survey, as well as "facts on the ground" such as knowledge of legislation, traditional values, and resources available to the community. I discuss the relations among these environmental and sociodemographic variables, as well as the formal relations and constructs given by the Social Cognitive Theory. Social Cognitive Theory and its antecedent Social Learning Theory accommodate relations with social norms (Zimmerman and Vernberg 1994; Flay and Petraitis, 1991; Akers et al., 1979) and with intrapersonal variables such as beliefs and stress (Baranowshi et al., 1997). I include these in the model as well. Figure 6.3 and Figure 6.4 display the specific factors that influences youth tobacco use based on mixed method findings. 153

PAGE 168

-------Figure 6.3. Partial Model for Rural Environment Using Qualitative and Quantitative Findings PB3 Rural Location Traditional Values Female Gender EPS Obs & Direct Learning PB12 Smoking Status Age R2 Lack of Recreational Outlets l EPB lntrapersonal: Boredom Figure 6.3 represents a display of the relationships in the data about rural environment. The relationships are represented by lines; each line has a label composed of two components: a character code representing the type and direction of the relationship, and a unique numeric identifier that links the relation to a description Table 6.1 describes the influences and relations in Figure 6 3 If traditional Turkish values are believed to influence the observational learning of adolescent males, for example, that influence is labeled EP6, an environment-person relation under Social Cognitive 154

PAGE 169

Theory. Several types of relations are used. The foremost of these are Bandura's three types of influences: person-behavior (PB), environment person (EP), and behavior-environment (BE). lntrapersonal cognitive processes are given a single letter "C". The existence of these processes is postulated under Social Cognitive Theory as emotional coping responses and self-control. Table 6.1. Influences in the Partial Model for Rural Environment R1 R2 PB7 EP8 C9 PB10 Location Traditional Recreation Boredom Values Location minimizes recreational outlets alues encourage male smoking Smoking Status Survey Measure: Age increases risk of smoki Boredom Lack of recreational outlets increases boredom Some relations are beyond the scope of Social Cognitive Theory, specifically among factors in the social environment. One example of this is the rural location of Dursunbey being associated with traditional Turkish values; another example is legislation limiting the authority of school administrators' off-campus. In the terminology of Social Cognitive Theory, 155

PAGE 170

these relations might have been loosely termed "environment-environment" influences, but I avoid calling these "influences" to prevent confusion with the three influences described by Social Cognitive Theory: they are instead labeled with a single letter "r". It must be emphasized that these relations are not claimed to be causal, but merely associational. Influences that were directly measured by the survey are given heavier lines. Gender and age were directly measured in the survey and found to be significant risk factors for smoking. These demographic factors are shown as direct influences that may be explained by other influences shown in parallel. Qualitative data and cultural background information suggest that traditional Turkish cultural values, part of the social environment, encourages male youth smoking while discouraging female youth smoking. This is an environment-person influence, leading to both observational and direct learning of gender-specific social norms and the person-behavior influence of that learning on smoking (and non-smoking) behavior. Direct observation and experience indicates that Dursunbey's rural location is associated with its lack of recreational outlets for youth. The interview data strongly suggest that a portion of Dursunbey's youth feels that there is little to do with free time, which may in part be due to a lack of recreational alternatives. If so, then the environment influences the person's state of boredom, leading to an emotional coping response. The person-behavior influence of that coping response leads to smoking behavior. Figure 6.4 represents the influences in the school environment using qualitative and quantitative findings. 156

PAGE 171

Figure 6.4. Partial Model for School Environment Using Qualitative and Quantitative Findings Education Policy : Mandatory 8 yrs R13 t Younger Ages Mixed Age Groups at School /Older Ages R15 R14t / R16 Exposure of Younger Students to Smoking EP171 / --::Teacher Role Modeling OR= 2.9 Smoking &Obs / EP23 Learning Number of > e.o PBl8l Friends Smoking EB22 ........__ Smoklg Sta'"' 1"819 t 1-l I = 0 4 1'(24 Direct & Obs Learning BE20 Academic Achievement The survey found that teacher smoking was associated with an increased risk of smoking, as was the number of close smoking friends. Interviews with students and educators described several influences at work in the school environment. An extension of mandatory elementary school education from 5 years to 8 years resulted in older and younger children being mixed together in a shared campus, exposing younger children to the smoking behavior of older children. In addition to the role modeling influence, teacher-smoking on-campus diluted the effectiveness of anti smoking messages received at school Any influence associated with 157

PAGE 172

increased student smoking would indirectly play into the influence of the number of close friends who smoke (an influence not exclusive to the school environment). Several of these influences are exemplars of reciprocal determinism. Table 6.2 explains the influences and relations in Figure 6.4. Table 6.2. Influences in the Partial Model for School Environment H1t-Bfl R13 Education Policy Mixing Age 1997 education policy change Groups extending mandatory education mixes age groups R14 Mixed ages Younger students Mix of students exposes younger exposed students to smoking behavior of older students R15 Younger ages Younger students Younger students are mixed in with exposed older students at school R16 Older ages Younger students Older students are mixed in with exposed younger students at school EP17 Younger students Role modeling Younger students model older exposed and obs learning students, learn smoking_ behaviors PB18 Role modeling and Smoking Status Learning influences increase in obs learning smoking behavior PB19 Academic Smoking Status Survey Measure: Academic achievement achievement is protective against smoking BE20 Smoking Status Younger students Increased smoking leads to more exposed exposure for _younger students BE21 Smoking Status Number of Increased smoking leads to more smoking friends close friends smoking EB22 Number of smoking Smoking Status Survey Measure: Number of close friends smoking friends associated with increased smoking EP23 Number of smoking Role modeling Number of friends smoking provides friends and obs learning opQQrtunity for learning smoking R24 Number of smoking Younger students Number of friends smoking exposes friends exposed some younger students to smoking EB25 Anti-smoking Smoking status Anti-smoking messages has messages influences non-smoking behavior PB26 Direct & Observat. Smoking Status Learned discounting of anti-smoking Learning messages influences smoking EB27 Teacher Smoking Smoking Status Survey Measure: Teacher smoking influences student smoking status EP28 Teacher Smoking Role modeling & Students role model on teacher obsv learning smoking behavior EP29 Teacher Smoking I Direct & obsv. Students learn to discount anti-Anti-Smoking mes. learning smoking messages 158 : .

PAGE 173

Limitations Sample Size. The relatively small sample size of this study limited the statistical analysis. Specifically, the sample size was not large enough to determine significant differences in regards to specific independent variables between female smokers and nonsmokers. However, the gender was still used as a control variable and logistic regression adjusted for the gender differences. Although this study is based on a relatively small sample of students, the major strength of the study is that a full census of the 1Oth graders was accomplished. Therefore, the study population is representative of 1Oth graders, but we can only generalize these findings to the glh and 11th grade students with caution. Similarly, we must also be cautious when generalizing these findings to adolescents who have dropped out of school or terminated their education after the compulsory education of eight years This would be particularly true of girls, who stop going to school in much larger numbers than boys. However, this study helps us to assess the smoking behavior and trends among these grades too. Self-Report Bias. As in all studies that gather data using questionnaires, the variables in this study were measured by the responses of the study subjects, rather than direct observation. Independent biochemical measures of nicotine or carbon monoxide were not used to confirm adolescents' reporting of their smoking status. This leaves open the possibility of underreporting or never reporting of smoking status. This limitation was not only true for self-reported behaviors, but also for the behavior of others, such as the subjects' friends, parents, siblings and teachers. 159

PAGE 174

This study divided adolescents into two groups: smokers and nonsmokers. A more detailed division among smokers (i.e. triers, experimenter, regular, and established smoker) should be made in future studies. In the current study, 83% of the students indicated that they have already tried smoking, but not all of these students have become regular smokers. We should investigate which types of adolescents become established smokers after their trying and experimentation phases. Measurement Issues. In retrospect a few of questions in the survey were problematic. For instance, it was difficult to interpret the question measuring the friends' smoking status. During the analysis I realized that the problem was caused by measurement error, due to allowing students to answer "not sure" to this question. Although survey design practice requires providing neutral options, I would not allow the "not sure" option in the future studies where the likelihood of not knowing is small. The question measuring the "Friend approval of smoking had the similar measurement problem, since students were allowed to answer "not sure." In the future studies, I would ask this question dichotomously, where students can answer either "yes" or "no." The question measuring "Weekly allowance" was categorized, but since the intervals were not equal, I was not able to treat weekly allowance as a continuous variable. In the future I would have the study participant fill out a blank space for the amount of weekly allowance they receive, and categorize it after the data collection. In addition, some of the students might have had some difficulties answering questions because of their lack of familiarity with surveys. This was raised by one of the school counselor as a potential barrier. 160

PAGE 175

Cross-Sectional Data. The findings of this study are based on cross sectional data, which makes this study an inappropriate instrument for sorting out causal pathways. Future studies will require longitudinal research to accomplish this goal. Furthermore, future studies should look into the different characteristics of smokers who are at the different stages of their smoking career (i.e., trier, experimenter, regular, and established smoker). Multiple Comparisons. I looked at many predictors in both by themselves and with other variables in the model. Each time one makes a comparison the probability of incorrectly rejecting the null hypothesis (Type I error) increases by .05. However, my study was testing a specific hypotheses which strengthens my approach and findings. Recall Bias. To avoid recall bias, in addition to asking an account of the past thirty days of smoking behavior, smoking in the past week would be more accurate in order to reflect youths' smoking trajectories. 161

PAGE 176

CHAPTER 7 CONCLUSION AND RECOMMENDATIONS A. Summary of the Findings The quantitative portion of this study found that the social cognitive model of social influences and sociodemographic variables explained 60% of the variance overall in smoking behavior; the model predicted 81% of the responses correctly. Only three demographic variables and two social environment variables predicted adolescent current smoking. Among the interpersonal variables, friend smoking provided the best prediction for current smoking. Teacher smoking also significantly predicted current smoking. In addition, among the sociodemographic variables gender, age, and academic performance significantly predicted current smoking. When control variables were added into the models, the leading risk factors for adolescent smoking were being older, male, having friends who smoke and teacher smoking. Low academic performance was an additional risk factor. In addition, the place of origin was approached statistical significance (.090). The only significant social influences supported by the survey are friends' smoking and exposure to teacher smoking (environment person interactions and person-behavior interactions). Both friend influences and teacher influences were strongly correlated with current smoking, with the risk of smoking increasing almost exponentially with the number of close friends who smoke. Interviews confirm the importance of teachers as role 162

PAGE 177

models, reinforced by an apparent education gap between youth and their parents, making identification with teachers easier than in past generations, and undercutting anti-smoking messages received at school when teachers smoke. The findings emphasize the importance of a smoke-free school environment by suggesting that exposure to teachers' smoking during school hours influences the smoking behavior of adolescents. The implication is that teachers epitomize the being powerful role model of the smoker, and smoking as a social norm. This study also emphasized the importance of gender-specific smoking prevention programs, especially through the knowledge gained from qualitative interviews. Support was also found for a relationship between smoking behavior and academic achievement (a person-behavior interaction). There was a weak correlation between place of origin and current smoking through the mediating process of academic achievement. Finally, person-behavior interactions were found between the biological personal factors of gender/age and smoking behavior. However, there was no support for role modeling of parents, siblings, or relatives influencing adolescent smoking behavior. Furthermore, there was no support for the relationship between adolescents' perception of their environment (environment/situation) and adolescents' behavior (current smoking). This finding was inconsistent with previous findings of a positive correlation between smoking and adolescents' perception of smoking rates among peers, and between smoking and the perceived support for smoking (Conrad et al., 1992; Flay & Phil, 1985). The relationship between school type (environment) and current smoking was not supported. The perceived economic situation and weekly allowances variables were used as proxies for socio-economic status (SES). There was no relationship between economic situation and smoking but 163

PAGE 178

there was a weak relationship between weekly allowances and adolescent current smoking behavior. Although smoking is a mainstream behavior in Turkey, especially among adult males, smoking among adolescent boys and girls is discouraged culturally. In general, however, there is a greater acceptance of boys' smoking than girls' smoking. There is a weak correlation between high smoking prevalence estimates among peers and adolescent smoking in the survey. This contrasts with interviews, where participants regularly stated that "everybody smokes." While social norms discourage young women from smoking in Dursunbey, they reinforce smoking as part of the male identity, established everywhere from the family home to coffee houses. For young women, study interviews and comparison of ever-smoking rates with past studies suggest that the protective influence of Turkish tradition might be disappearing. Recommendations For Future Studies and For Intervention The study has implications involving gender and age of uptake. The increasing ever-smoking rate among females in Turkey suggests that the regular smoking rate among young women may rise dramatically in the coming years. This is an urgent issue requiring further investigations into smoking among female high school students, college students and school dropouts. Many of the 1 olh grade students were already regular or established smokers, indicating a need for smoking prevention programs in elementary and middle school level. Although the qualitative interviews provided some evidence of family influences in smoking, the survey showed no significant family influence. Interviews showed that the education gap increased conflict between 164

PAGE 179

students, particularly girls, and their parents. Adults interviewed appeared to be resigned to their male children eventually smoking. A study dedicated to exploring family influences in greater detail might find other factors that my study did not reveal. Perhaps research employing other measures of family influences such as family bonding (Hirschi, 1969), and family functionality (Diana, 2000), and family communication patterns (Ritchie, 1991 and Tims & Masland, 1985) might be useful to use in future studies. In addition, offering classes to parents to teach how to effectively communicate with their children might be effective. Adolescent peer networks should also be investigated further for designing effective smoking prevention programs. We should understand the mechanism of peer influences. In future studies, close friends effects should be longitudinally assessed; studies should employ scales measuring the interactions between personal characteristics and the modeling domain. The adolescents in this study were adamant about not wanting younger ones to start smoking. We could use older teenagers' to help prevent younger students from smoking. This kind of method has been effectively used in the United States. As my study has found, greater academic achievement seems to have some protective effect against smoking. Schools should encourage parental involvement in their children's academic careers, and establish good connections between school and families. This has the potential of partially offsetting the disadvantages of growing up in a resource-poor rural school district. Furthermore, individual academic performance and its relation to other social factors should be investigated in future studies. Teacher influence has implications for future research design and youth smoking prevention strategies in schools in many countries. In my study, all the environmental and personal influences on behavior can be 165

PAGE 180

addressed as part of the school curriculum. School-based curricula and school environment changes are cost-effective prevention programs, particularly since they reach 95% of youth in schools, and teachers can be enlisted in the intervention effort. As a part of a comprehensive tobacco prevention program, cessation programs should be offered not only for students, but also for critical role models such as teachers, counselors, and other school staff. The study interviews showed that school staff members are not particularly knowledgeable about the deleterious health effects of smoking, often subscribing to some of the same smoking myths as their students. School based community outreach efforts should be used to extend cessation programs to school dropouts, who disproportionately represent the female population. Interventions should target the myths about smoking. The interviews revealed that many students believe that they can quit anytime. Adolescents must be warned about the power of nicotine addiction and difficulty of quitting, but not in a way that leaves adolescent smokers believing that they are powerless to quit. The study suggests that social learning has resulted in perceived unrealistically high smoking rates among peers. We should counter the social norms summarized as "everybody smokes," delivering the message that, "yes, almost half of the adult population smokes in Turkey, but other half don't." Since many students linked their smoking to stress caused by problems with family, friends, and school, school curricula should introduce stress and conflict management skills into their health-education programs, providing training in stress-reducing alternatives to tobacco use. However, teaching social skills in one program would not be enough to solve the tobacco use problem. These programs should be comprehensive and 166

PAGE 181

address the various intervening variables. Comprehensive tobacco prevention programs should also be combined with nutrition and physical activity advocacy. My study also shows that smoking is a widely practiced and socially accepted habit. It is also clear from the interviews that adolescent-specific anti-smoking media messages are seldom seen or heard in Turkey, since they are usually relegated to infrequent late-night time spots. As the experience in the United States has shown, large-scale mass media exposure can be a critical element in changing the social norms about smoking. High-quality media campaigns should be used to de-glamorize tobacco and counter common pro-tobacco norms. Tobacco control efforts should fund large-scale mass media campaigns, to include public figures in tobacco control such as soccer players, pop singers, and other role models that youth respect. Messages also should be delivered by youth themselves rather than authority figures. Serious health consequences of tobacco use, both short term as well as long term, should be publicized. Messages should also include warnings about the health effects of passive smoking, something that is absent from the few messages that are broadcast. Like the rest of the world, Turkish tobacco interventions and studies can benefit from experimenting with newly developed methods. Intervention methods successfully used in the US (e.g. Florida Truth Campaign and American Legacy Foundation's advertisement Campaigns) and elsewhere should be tailored to Turkish population and their effectiveness should be tested. Much can be gained by improvements in study design. The next nationwide study in Turkey should combine qualitative and quantitative methods to furnish the finer details of the influences on female smoking, resulting in better-designed anti-smoking interventions. As Flay et al. (1997) suggested, longitudinal studies will help establish causal lines of influence, 167

PAGE 182

and a greater integration of theory into research will add rigor to efforts predicting smoking outcomes. In my study, for example, I was not able to determine from the cross-sectional survey data if smoking was a result of friend influence or if smokers selected each other as friends. Policy Implications Some of the following recommendations are not directly drawn from my survey results, but are my observations reflecting the experience I gained during the qualitative study and my past four years' experience in tobacco control. In this study's interviews, school administrators and teachers expressed dismay over their loss of off-campus control of their students' smoking, a product of school-reform legislation; students can often purchase single cigarettes within sight of the school grounds. This suggests that a balance can be restored by strict enforcement of existing laws against single-cigarette sales and tobacco sales to minors, possibly through interventions designed to foster school and community influence over the police. This would be part of a strategy for delivering consistent anti-smoking messages to adolescents. The interviews revealed the sense of hopelessness for the future on the part of Dursunbey's 10111 graders. In part, this reflects the unequal distribution of educational resources between Turkey's urban centers and its rural districts. Although there are limited resources to go around, equal preparation for college would give the children of Dursunbey one less reason to abandon their educational aspirations, allowing some the additional protection of academic involvement. In this sense, tobacco reform and education reform have compatible interests. 168

PAGE 183

The interviews also showed that separation of the sexes for evening recreation led to male adolescents spending time in adults-only smoke-filled coffee houses. This partially reflects a lack of evening recreational alternatives for youth. This could be offset somewhat by after-school programs and community-organized evening recreation for youth. Like many other reforms, funding might be a major obstacle to bringing about changes of this nature. In 1996,Turkey passed comprehensive legislation banning the advertising and promotion of all tobacco products; the sale of tobacco products to minors; smoking in all public areas, public transport, in government offices, hospitals, schools in which five or more people work. The new legislation required public places to provide enclosed areas for smokers, and required national and private television running anti-smoking programs 90 minutes per month; health warning on cigarette packs must be printed in a manner so that can be easily read (Emri, 1997). After seven years, however, smoking is still highly prevalent in public places, cigarette sales to minors is still widespread, and the mass media buries anti-smoking messages into late-night timeslots where they are not seen. Moreover, to circumvent advertising restrictions, the transnational tobacco companies print "Price Announcements" to announce that prices have not changed. RJ Reynolds ran a TV advertisement for "Camel Trophy Watches" despite the government's ban on cigarette advertising on TV (Hammond & Purcell, 1999). In 1999, a severe earthquake hit six towns in Northwest Turkey. An estimated 45,000 people were killed and many hundreds of thousands were wounded. Millions were left homeless, jobless overnight. RJ Reynolds (now Japan Tobacco) helped this scene of catastrophe by distributing free cigarettes to earthquake victims (Dagli, 2002). RJ Reynolds was not alone in finding ways to ingratiate themselves 169

PAGE 184

to the Turkish population. In a controversial move, Philip Morris paid for restoration of Turkey's first post-Ottoman parliament building, which reopened in April, 2001 (Dagli, 2002). In December of 2001, a member of Turkish parliament submitted a law proposing to exempt Formula-1 races from the tobacco-advertising ban. This was the fifth attack on the comprehensive tobacco ad ban. Phillip Morris applied to Parliament, the Ministry of Sports and Youth, the Ministry of Tourism and the Constitutional Court to have the ban declared unconstitutional. Formula 1 races are scheduled in Turkey for the year 2005. Hundreds of faxes have been sent to Formula 1 to issue a statement that Formula 1 races in Turkey in 2005 will be free of tobacco sponsorship. So far, Formula 1 has ignored these demands, even though tobacco sponsorship is now illegal under Turkish law. Many tobacco control advocates suspect that Formula 1 will try to either weaken the law or pull entirely out of Turkish auto racing, as it already has in Belgium and Australia over those countries' advertising bans. Formula 1 gets about $350 million from tobacco industry sponsorship annually (Anna White, personal communication, 2003). In my own study (Dursunbey, Turkey), 54.4% of the respondents indicated that they have seen cigarettes adds on race cars (May 2002). Recently, the Turkish Government started construction of the racetrack for the 2005 Formula 1 races on a natural heritage site in Istanbul. This move created controversy among environmental and other organizations in Turkey. The Turkish government also agreed to pay the Formula 1 organizer $13.5 million a year over seven years for the rights to stage the race, as well as penalty of $25 million if the track is not ready by March 2005 (Financial Times, 2003). These examples show that passing anti-tobacco legislation is only a first step; the greatest challenge is enforcement and fostering community 170

PAGE 185

involvement. In sum, Turkey's experience indicates that transnational tobacco companies are relentless opponents of tobacco control legislation. The battle is not over when laws are passed. International norms will help countries like Turkey resist efforts to undermine the health of its citizens. Additional legislation also has its place. Turkey should pass legislation requiring that tobacco companies to put graphic health warnings on cigarette packs, a method that has been effectively used in Canada and Brazil. Smoke free work places should be implemented, a measure that has proven to be very effective. Smoke free public areas would help to de normalize smoking. Shops around school should not be allowed to sell cigarettes, and current laws prohibiting the sale of single cigarettes should be enforced. Tobacco taxes should be increased to an optimum level (optimum to avoid inducing a black market), increasing cigarette prices for the price-sensitive adolescent and adult populations. Since youth are price sensitive, population price increases have been very effective way of reducing smoking among youth. Unfortunately, the tobacco control community in Turkey is relatively small and without resources, and the government lacks the commitment to deal with tobacco control issues. This is where coordinated global tobacco control efforts can help, not only by sharing expertise, but also by joining the discussion of tobacco control issues with governments like Turkey. Furthermore, experience has shown that coordinated global efforts can result in policy changes on the part of Western governments, international agencies, and non-governmental agencies. The World Health Organization initiative Framework Convention on Tobacco Control (FCTC) is an example of what can be achieved through coordinated action. In particular, efforts should be aimed at increasing World Bank and other agencies' funding for tobacco control projects in developing countries. To have effective tobacco 171

PAGE 186

control programs, scientific findings from tobacco control studies should be combined with effective anti-tobacco advocacy. In sum, researchers and advocates should coordinate their efforts and become globally organized. The greatest enemy of Turkish tobacco control advocates is deeply ingrained social norms that favor smoking. As stated by one school counselor: Everyone should have a bad habit. Otherwise life is so boring. You should balance bad and good habits. If you don't drink and smoke, why do you live? Everybody will have to die. While we are living, we should enjoy the pleasures of life. Combating those norms requires a committed, multifaceted approach similar to those outlined in these recommendations. As in the United States and other countries whose smoking prevalence has dropped in the last half century, this is a decades-long, multi-generational project aimed at presenting youth with consistent anti-smoking messages. We will then have a chance of negating the well-known Turkish acceptance of smoking. 172

PAGE 187

APPENDIX A SURVEY INSTRUMENTS AND SEMI-STRUCTURED INTERVIEW GUIDE Preliminary Survey (English, June 2001) Preliminary Survey (Turkish, June 2001) Final Survey (English April 2002) Final Survey (Turkish, April 2002) Semi-Structured interview Guide (April 2002) 173

PAGE 188

Preliminary Survey (English, June 2001) Knowledge, Attitudes, and Behaviors about Smoking I Demographic questions 1. What grade are you in? ..... .............. ..... 2. How old are you? ......... 3. What is your sex? 1) Female 2) Male 4. What is your total household income? 1) 99 million and Less 3) 200-299 million 2) 100-199 million 4) 300-399 million 5. What is your weekly allowance? .......... .. I These questions are about your smoking 6. Have you ever tried cigarette smoking? (even one or two puffs) 1) Yes (How old were you) ............ .. 2) No (Please explain) ........................ 7. Have you ever been offered a cigarette? 1) Yes 2) No 5) 400-499 million 6) 500 million and above 8. If one of your best friends were to offer you a cigarette, would you smoke it? 9. Who encouraged you to try your first cigarette? .............................................................. 10. Where was this? ................................................................................................... 11. Who were you with? .................................................................. 12. How do you describe your cigarette smoking? 1) I don't smoke 3) I smoke regularly 2) I smoke once in a while 4) I quit smoking 13. Do you inhale cigarette smoke*? 1) Yes, I do 2) No, I don't 14. Have you tried to learn to inhale cigarette smoke*? 1 )Yes (Please explain) .............................................. 2) No 15. About how many cigarettes have you smoked in your entire life? 1) None 3) 1-10 cigarettes 5) More than 100 cigarettes 2) 1 or more puffs but never a 4) 11-100 cigarettes whole cigarette 174

PAGE 189

16. During the past 30 days, how many cigarettes did you smoke per day? 1) I don't smoke 5) 6 to 10 cigarettes per day 2) Less than one cigarette per day 6) 11 to 20 cigarettes 3) 1 cigarette per day 7) More than 20 cigarettes per day 4) 2 to 5 cigarettes per day 17. During the past30 days, what brand of cigarettes did you usually smoke? 1) I did not smoke 2) Turkish brands( please indicate .............. 3) Foreign brands (please indicate) .... .... ............ 18. Why do you prefer this brand? ..... ...... .............. .. 19. Would you prefer to smoke different brand than you smoke now (INhy please explain)? ........................ .. 20. Do you have a piece of clothing, like t-shirt or a hat that has a cigarette brand logo? 1) Yes 2) No 21. The following statements are about cigarette brand Mark the choice that best describes how much you agree or disagree with the statement. Strongly Agree Disagree Strongly Agree Cigarette brand shows how much money 1 2 3 4 people have Smoking foreign brands proves the 1 2 3 4 wealth of a person If people don't have money they smoke 1 2 3 4 Turkish brands The design and color of cigarette pack has influence on young people's brand 1 2 3 4 I preferences 22. During the past 30 days, how did you usually gel your own cigarettes? 1) I did not smoke 2) I bought them in a store, supermarket or TEKEL 3) I got them from my father' or mother's pack 4) I borrowed them from someone else 5) Other (Please indicate) .......................... 23. Have you ever purchased single cigarettes, sold out of a pack one at a time, at a store? 1) Yes 2) No 24. Where do you usually get the money fer cigarettes that you buy? 1) From parents 4) From your job 2) From brother or sister 5) Buy it with friends 3) From another relative 6) Other (please indicate) 25. How did you start smoking? 1) I do not smoke 3) Curiosity 2) Peer Influence 4) Other (Please describe) ....... .... ....... .. 175

PAGE 190

26 When d k ? o you smo e at most Strongly Agree Agree Disagree Strongly disagree When I am angry 1 2 3 4 When I am happy 1 2 3 4 When I am with friends 1 2 3 4 27. Why do you smoke now? (If you do not smoke you can skip to question 28) Strongly Agree Disagree Strongly A_gree disagree I want to show that I am 1 __ grown up 2 3 4 For pleasure 1 2 3 4 From stress 1 2 3 4 It helps me to find friends 1 2 3 4 I look up to my elder family 1 2 3 4 members I look up to my friends 1 2 3 4 28. What do you think about people who smoke? .... ......................... .... . .... ............ . ... ... 29 What are the main reasons for people of your age to start smoking? 30. Where do you usually smoke? 1) At home 3) At school 2) Outside 4) Friend meetings 31. Do your parents know that you smoke? 1) Yes 2) No 32 M k h h th b d ar t e c orce at est escrr es crgarette smo rng lY your parents b k' b ? Current Ex-smoker Never been a smoker smoker a) Your father 1 2 3 b) Your mother 1 2 3 33. Do any of your brothers or sisters smoke c i garettes? 0) I do not have any brothers or sisters 1) No, none of my brothers or sisters smokes 2) Yes one or more of my brothers or sisters smoke 3) I do not know whether any of my brothers or sisters smokes 34. H ow many o your nen s smo e f d k ? 1) None 4) Most of them 2) A few of them 5JAII of them 3) More than two 6) I do not know i f any of my friends smoke 35 On the average, out of 100 students in your grade how many do you think have tried smoking cigCI'ettes, that is, smoked at least a puff of a cigarette? 0) None 3) 21-30 6) 51-60 9) 81-90 1) 1-10 4)31-40 7) 61-70 10) 91-100 2) 11-20 5) 41-50 8) 71-80 11) All of them 176

PAGE 191

36 On the average, out of 100 students in your grade, how many do you think smoke cigarettes at least once a week? 0) None 3) 21-30 6) 51-60 9) 8190 1) 1-10 4) 31-40 7) 61-70 10) 91-100 2)11-20 5)41-50 8)71-80 11)AIIofthem 37 On the average, out of every 100 adults, how many do you think smoke cigarettes? 0) None 3) 21-30 6) 51-60 9) 81-90 1) 1-10 4) 31-40 7) 61-70 10) 91-100 2)1120 5)41-50 8)71-80 11)AIIofthem 38 What were you thinking about smoking before you started smoking? ... ... ......... .... ........ . 39 Do you think you can give up smok ing whenever you want? 1) I do not smoke 2) Yes 3) No 40 Do you want to completely stop smoking cigarettes? 1) I do not smoke now 2) Yes 3) No 41. During the past 12 months how many times have you tried to quit smoking? 1 ) I do not smoke 2) I have not smoked in the past 12 months 3) Never tr i ed 4) 1 times 5) 2 times 6) 3 to 5 times 7) 6 to 9 times 8) 1 0 or more times 42 How do you think cigarette-smoking affect your health? ..... . .............. . 43 Have you heard of any disease that may be caused by smoking cigarette (If your answer is yes, please list them)? ... ..... . ...... .... . . ................. . . .... ....... . .... .... ... .... .... ...... . .... . .... .... ...... .... 44 Have you ever thought to quit smoking because of future health worries? .................................. .......... . ... ........................... 45 Where do you usually get the information about harmful effects of smok i ng? I The next questions ask about your thoughts about tobacco. 46 Do you think that you will try a cigarette soon? 1 ) I have already tried smoking cigarettes 2) Yes 3) No 47. Do you think you will smoke a cigarette anytime during the next year? 1) Definitely yes 3) Probably not 2) Probably yes 4) Defin i tely not 48. If one of your best friends offered you a cigarette, would you smoke it?? 1) Definitely yes 3) Probably not 2) Probably yes 4) Definitely not 177

PAGE 192

49. Has either of your parents (or guardians) told you not to smoke cigarettes in the past 12 months? 1) Mother 3) Both 2) Father 4) Neither 50. Do you think young people who smoke cigarettes have more friends? 1) Definitely yes 3) Probably not 2) Probably yes 4) Definitely not 51. Do you think not smoking is a way to express your independence? 1) Definitely yes 3) Probably not 2) Probably yes 4) Definitely not 52. Do you think cigarette companies target teens to replace smokers who die? 1) Definitely yes 3) Probably not 2) Probably yes 4) Definitely not 53. Do you disapprove of people smoking one or more packs of cigarettes per day? 1) Definitely yes 3) Probably not 2) Probably yes 4) Definitely not 54 Do you think people risk harming themselves if they smoke one or more packs of cigarettes per day? 1) Definitely yes 3) Probably not 2) Probably yes 4) Definitely not 55 During this school year, were you taught in any of your classes the reasons why people your age smoke? 1) Yes 3) Probably not 2) Probably yes 4) Definitely not 56. During this school year, were you taught in any of your classes about the effects of smoking like it makes your teeth yellow, causes wrinkles, or makes you smell badly? 1) Yes 2)No 3) Not sure 57. How many teachers in your school do you think smoke cigarettes? 1) All of them 3) About half 5) Very few or none 2) Many 4)Some 58. How often do you actually see teachers at your school smoke on school property? 1) Every day 3) Occasionally 5) Never 2) Very often 4) Very seldom 59. What are your grades? 1) All 5's 4) Mostly 4's 7) Mostly 2's 2) Mostly 5's 5) Mostly 4 and 3 8) Mostly 2 or 1 3) All 4's 6) Mostly 3's 60 How many days were absent from school during this semester? 1) None 3) 10 or more days 2) 1-9 days 178

PAGE 193

61 M k th h th at' ec oice at describes you the best? Strongly Agree Disagree Strongly agree Disagree a) I am happy with the amount of control that I 1 2 3 4 have over decisions that affect my life b) My family looks for things to nag me about 1 2 3 4 c) Rules are created to be ignored 1 2 3 4 d) When I'm told to do something by a teacher, I 1 2 3 4 do it e) I have a lot of arguments with myfamily_ 1 2 3 4 f) When I make decision, I usually go by what 1 2 3 4 my parents taught me g) I can control mv decisions that affect my_life 1 2 3 4 h) If I don't like something I'm told to do, I often 1 2 3 4 put it off or just don't do it all i) If I don't like an order I have been given, I may 1 2 3 4 not do it or may do only part of it j) I feel guilty when I break a rule 1 2 3 4 kl_l can influence decisions that affect my friends 1 2 3 4 62. The following statements are about smoking cigarettes, Mark the choice that best describes how much you agree or disagree with the statement. Strongly Agree Disagree Strongly _agree a) Smoking makes people look tough 1 2 3 4 bl_l will not smoke in the future 1 2 3 4 c) Smoking helps people feel more comfortable 1 2 3 4 when they are with otherpeople d) Inhaling someone else's cigarette smoke can 1 2 3 4 cause lung cancer in nonsmokers e) Smoking is something you need to try before 1 2 3 4 deciding to do it or not f) Smoking helps people to forget their worries 1 2 3 4 g) Smoking helps people relax 1 2 3 4 h) If I don't smoke. I will be healthier 1 2 3 4 i) I strongly dislike being around people who are 1 2 3 4 smoking j) Smoking is enjoyable 1 2 3 4 k) If I smoke, it will cause problems with my family 1 2 3 4 and me I)Smokil}flhelps people look beautiful 1 2 3 4 m) Smoking makes you more likely to get heart 1 2 3 4 disease n) Smoking tastes good 1 2 3 4 o) Smoking cheers people up when they are in a 1 2 3 4 bad mood p) If I talk to my friends about the dangers of 1 2 3 4 smoking, it will help them decide not to smoke 179

PAGE 194

Preliminary Survey (Turkish, June 2001) Sigara ile ilgili Tutum, Bilgi Ve I Demografik sorular 1 Karymc s1mrtasm1z ? . . ........ ... ... ....... . 2. Kary ya!imda s1mz? 3. Cinsiyetiniz? 1) KIZ 2) Erkek 4. Ailenize giren ayllk para miktan? 1) 99 milyon ve a!iaQISI 3) 200-299 milyon 5) 400-499 mil yon 2) 100-199 milyon 4) 300-399 mil yon 6) 500 milyon ve ozeri 5 Haftallk harryhQiniZ ne kadar? .... I sorular sizin sigara ile ilgilidir 6. Hiry sigara irymeyi denediniz mi (sadece bir nefes bile)? 1) Evet (ilk kary ya!imda denediniz?) ........ . . ... 2) Hay1r (nedenini .................... .... . 7. Size hiry sigara ikram eden oldu mu? 1) Evet 2) Hay1r 8. Eger yakm arkada!ilanmz size sigara i kram else iryer misiniz? ...... ......... ............ .......................................... 9. ilk sigaramzl denemeniz iryin sizi kim cesaretlendirdi? 10. Ilk sigaramz1 nerede denediniz? .......... ........... ......... .......... ...... ................ ..... ... . . ......... .... ........ . 11. ilk sigaran1t1 denediginizde yammzda kimler vardl? . ..... . ....... ... ....... . ... ...... .................. . ..... . 12 Sigara iryiminizi nas1l tammlarsm1z? 1) irymiyorum 3) Arada bir sigara iryiyorum 5) Sigaray1 b1raktlm 2) DOzenli sigara iryiyorum 4) Hiry sigara irymiyorum 13 Sigaray1 iryinize ryekiyor musunuz? 1) Evet 2) Hay1r 14. Sigaray1 iryinize ryekmeyi ogrenmeye ryah!illniZ mi? 1) Evet (nas1l ac;:1klaym1Z) .................... . ...... ....... . .... ...... . ...... . . .......... ..... . ..... 2) Hay1r 15. Hayatm1z boyunca kac;: sigara ic;:tiniz? 1) Hiry denemedim 3) 1-10 sigara 5) 100 den fazla 2) Sadece bir nefes ryektim 4) 11-100 sigara 180

PAGE 195

16. 30 gun gOnde ortalama sigara 1) Sigara ic;:miyorum 5) 6 ile 10 sigara arasmda 2) Bir sigaradan az 6) 11 ile 20 sigara arasmda 3) Gunde bir sigara 7) 20 den fazla 4) 2 ile 5 sigara arasmda 17 Geytigimiz 30 gun ic;:inde hangi marka sigara 1) Sigara ic;:medim 2) Yerli marka belirtiniz) ....................... 3) Yabanc1 marka (belirtiniz) ....... ........ .. 18. Bu markay1 tercih edil? nedenlerinizi ac;:1klar miSiniZ? .... ....... .............. .. 19. Bal?ka bir markay1 tercih eder miydiniz? (Nedenini ................ .. 20. Ozerinde sigara markas1 olan hiC(bir k1yafetiniz var ml? 1) Evet 2) Hay1r 21. Sigara markas1 ile ilgili al?aQidaki soylemlere kat1hp katllmadiQiniZI belirtiniz. Kesinlikle Kat1hyorum Kat1lm1yorum Kesinlikle kallhy_orum kat1lm1yorum lvilen sigara markas1 1 2 3 4 _ara miktannm olcusi.idur Yabanc1 sigara ic;;mek 1 2 3 4 zenginli!'lin gostergesidir Paras1 olmayanlar yerli sigara 1 2 3 4 icer Sigara paketinin dizayn ve 1 2 3 4 rengi genc;:lerin marka seciminde etkendir 22. 30 gun sigaramz1 nas11 temin ettiniz? 1) Sigara ic;:medim 4) Odunc;; ald1m 2) Market, bakkal veya tekel bayilerinden kendim satm aldlm 5) Bal?ka (Belirtiniz) ...... ...... ..... .. 3) Anne veya babamm paketinden ald1m 23. Market, bakkal yada tekel bayisinden hie;; paket dll?tnda tek sigara satm ald1mz ml? 1) Evet 2) Hay1r 24 Sigara alacak paray1 nereden buluyorsunuz? 1) Ailemden 4) Kendi il?imden 2) Abim yada ablamdan 5) Arkadal?lanmla ortak ahyoruz 3) Bal?ka bir akrabamdan 6) Bal?ka (ac;;1klayrmz) 25. ilk sigara nasrl bal?ladrnrz? 1) Sigara ic;:miyorum 3) Merak 2) Arkadal? bask1s1 4) Bal?ka (ar;aklay1nrz) 26. Enc;;ok hangi zamanlarda sig_ara ic;:iy_orsunuz? Kesinlikle Kat1hyorum Kat1lm1yorum Kesinlikle kat1hyorum kat1lm1yorum Sinirli oldugumda 1 2 3 4 Mutlu oldu_9umda 1 2 3 4 1 2 3 4 birlikteyk_en 181

PAGE 196

27. $u anda niye sigara ic;:iyorsunuz? Kesinlikle Kat1hyorum Kat1im1yorum katlhyorum Sigara ic;:miyorum 1 2 3 Bi.iyi.idi.igi.imi.i kamtlamak ic;:in 1 2 3 Zevk ic;:in 1 2 3 Stresten 1 2 3 bulmama yard1mc1 1 2 3 oluvor Ozenti 1 2 3 28. Sigara ic;:enler hakkrndaki neler? ............................................. 29. Sizce sigara ic;:me nedenleri neler? ................ .... ... ....... ..... ......... 30. yerlerin hangisinde sigara ic;:iyorsunuz? 1) Evde 3) Okulda 2) 4) toplant1lannda 31. Aileniz sigara ic;:tiginizi biliyor mu? 1) Evet 2) Hay1r 32. Asag1dakilerden hangisi ailenizin siaara icimini en iyi tammhvor? Siaara icivor Siaarav1 b1rakt1 Hie icmedi a) Babamz 1 2 3 b) Anneniz 1 2 3 33. Erkek yada k1z sigara ic;:en var mi? 0) yok 1) Hay1r, k1z yada erkek hic;: biri ic;:miyor 2) Evet, bir yada daha fazla erkek ve kiZ ic;:iyor 3) sigara ic;:ip ic;:medigini bilmiyorum 34. Kac arkadasrnlz sigara ic;:iyor? 1) Hicbiri 4) Neredevse bi.iton arkadaslanm 2) Sadece birkac tanes1 5) Heosi Kesinlikle katllm1vorum 4 4 4 4 4 4 3) 2 den fazta 6) Arkadaslanmrn sigara ic;:ip ic;:mediaini bilmivorum 35. Sizce lise 2. stmfta 100 ogrenci ic;:inde ortalama kac;: tigrenci sigara ic;:meyi (bir nefes bile olsa) 0) Hic;:kimse 1) 1-10 2)11-20 3) 21-30 4)31-40 5)41-50 5) 51-60 6)61-70 7) 71-80 8) 81-90 9) 91-100 36. Sizce lise 2. srntfta 100 ogrenci ic;:inde ortalama kac;: ogrenci en az haftada bir sigara ic;:iyordur? 0) Hic;:kimse 3) 21-30 5) 51-60 8) 81-90 1) 1-10 4) 31-40 6) 61-70 9) 91-100 2)11-20 5)41-50 7)71-80 37. Sizce 100 ic;:inde ortalama kac;: tanesi sigara ic;:iyordur? 0) Hic;:kimse 3) 21-30 5) 51-60 8) 81-90 1) 1-10 4) 31-40 6) 61-70 9) 91-100 2) 11-20 5) 41-50 7) 71-80 182

PAGE 197

38. Sigara ir;meye once sigara hakkrnda ne ........... . . . .... ..... . 39. Sigaray1 isteginiz zaman blrakabileceginizi musunuz? 1) Sigara ir;medim 2) Evet 3) Hay1r 40. Sigaray1 tamam1 ile b1rakmay1 musunuz? 1) anda ir;miyorum 2) Evet 3) Hay1r 41. Ger;tigimiz 12 ay ir;inde kar; kere sigaray1 b1rakmay1 denediniz? 1) Sigara ir;miyorum 3) Hir; ir;medim 2) Ger;tigimiz 12 ay ir;inde sigara ic;:medim 4) 1 kere 5) 2 kere 7) 6-9 kere 6) 35 kere 8) 10 kereden fazla 42 Sigaranm saghgtmz Ozerinde ne gibi etkileri oldugunu .... . ........... ..... 43. Sigara tarafmdan kaynaklanan bir hastahk duydunuz mu? Cevabmtz evetse hastaltklarr S1ralar mtsmtz? ...... ............ . ..... ........................... ....................................... ..................... .. 44. Gelecekteki saghgtmz konusunda zamanlar oluyor mu? .... ......... 45. Sigaranm zararlan ile ilgili bilgiyi nereden ahyorsunuz? I AaQ1daki sorular sizin sigara ile ilgili diJiinceleriniz hakkmda 46 Sigaray1 yakm zamanda denemeyi musunuz? 1) $u and a sigara ir;iyorum 3) Hay1r 2)Evet 47. Sigaray1 gelecek y1l deneyeceginizi musunuz? 1) Kesinlikle denerim 3) Belki denemem 2) Belki denerim 4) Kesinlikle denemem 48 Eger yakln biri size sigara ikram else icter miydiniz? 1) Kesinlikle ir;mem 3) Belki ic;:mem 2) Belki ir;erim 4) Kesinlikle ir;erim 49 Geytigimiz 12 ay ir;inde aile Oyeleriniz size sigara ir;meme yonunde nasihat etti mi? 1) Annem 3) lkisi birden 2) Babam 4) Hir;biri 50. Sigara ir;en genr;lerin daha r;ok oldugunu musunuz? 1) Kesinlikle evet 3) Belki hay1r 2) Belki evet 4) Kesinlikle hay1r 51. Sigara ir;memenin bagtmstzhgrmzt ifadenin bir yolu oldugunu musunuz? 1) Kesinlikle evet 3) Belki hay1r 2) Belki evet 4) Kesinlikle hay1r 52. Sigara olen sigara tiryakilerinin yerine genr;leri hedef aldtgtm musunuz? 1) Kesinlikle evet 3) Belki haytr 2) Belki evet 4) Kesinlikle hayrr 183

PAGE 198

53 Gunde bir paket yada daha fazla sigara ir;enleri onayl1yor musunuz? 1) Kesinlikle evet 3) Belki hay1r 2) Belki evet 4) Kesinlikle hay1r 54. Bir yada iki paket ir;enlerin kendilerine zarar verdigini musunuz? 1) Kesinlikle evet 3) Belki hay1r 2) Belki evet 4) Kesinlikle hay1r 55. Bu ogretim y1flnda derslerinizde sigara ir;me nedenleri uzerinde mu? 1) Kesinlikle evet 3) Belki hay1r 2) Belki eve! 4) Kesinlikle hay1r 56 Bu ogretim y11inda sigaranm etkileri (ornegin : sararmas1, sebep olmas1, yada nefesinizin kotu kokmas1) anlatan bir dersiniz oldu mu? 1) Evet 3) Em in degilim 2) Hayrr 57. Okulunuzda kacr ogretmen sigara ir;iyor? 1) Hepsi 3) Yans1 5) Sadece birkay tanesi 2) Bir r;ogu 4)BazJian 58. Okulda iigretmenlerinizin sigara it;tigini hangi Slkhkta goruyorsunuz? 1) Her gun 3) Arada bir 5) Hicr garmuyorum 2) c;ok s1k 4) Nadiren 59. S1mf gecrme notlanmz nas1l? 1) Hepsi bell 4) c;ogu diirt 7) c;:ogu iki 2) c;ogu 5) c;ogu diirt ve ucr 8) c;:ogu iki yada bir 3) Hepsi dart 6) c;ogu ucr 60. Bu ogretim Yllmda kay gun devamSIZhk yaptlmz? 1) Hil;: yapmad1m 3) 10 gun ve uzeri 2) 1-9 gun 184

PAGE 199

61 A d k . f d a 1 1 a elere k k atiiiP atllmadiQiniZI belirtiniz? Kesinlikle Kat1hyorum Kat1im1yorum Kesinlikle kat1hyorum katlim1vorum a) Ozerinde sahip oldugum 1 2 3 4 kontrolden mutluY\.Jm b}_Ailem surekli saylenecek bulur 1 2 3 4 c) Yasaklar t;ignenmek it;indir 1 2 3 4 d) Ogretmenlerim bir sey soylediginde 1 2 3 4 _yaQanm e) Ailem ile sorekli 1 2 3 4 f) Kararlanm1 ailemin bana onerdigi 1 2 3 4 birtimde ahyorum g) kontrol eden kararlan ben 1 2 3 4 veririm h) Eger gitmeyen bir 1 2 3 4 vapmam istenirse __g_enelde bosveririm i) Eger gitmeyen bir yapmam 1 2 3 4 istenirse bir k1sm1n1 yap1p bir k1smm1 yapmam j) Kurallan rtignedigimde SU!fiU 1 2 3 4 hisssediyorum k) ile ilgili 1 2 3 4 kNarlarda elkili olurum 62. sigara irtimi ile ilgili ifadelere kat1hp kat1imad1gmrz belirtiniz. Kesinlikle Kat1hyorum Kat 1Im 1 yorum Kesinlikle kat1hyorum kahlm1yorum a) Sigara maco gozukurler 1 2 3 4 bl Gelecekte sigara icmeveceilim 1 2 3 4 c) Sigara insanlann grup irtinde kendilerini daha iyi 1 2 3 4 hissetmelerine vardrm eder d) sigara dumanmr ivine vekmek akciger 1 2 3 4 kanserine neden olabi/ir e) Sigara irtiP ic;;memeye kNar vermek ic;;in once 1 2 3 4 denemeniz gerekir f) Sigara insanlann s1krntrlannr unutmaya yarar 1 2 3 4 __g)_ SJg_ara insanlan rahatlatrr 1 2 3 4 h) Eger sigara irtmezsem daha saghkh olurum 1 2 3 4 i) Sigara insanlann etrafmda olmay1 sevmem 1 2 3 4 j) Sigara irtmek zevk almacak bir 1 2 3 4 k) Sigara ailem ile aramda sorun 1 2 3 4 .!l_ Sig_ara insanlann guzel gorunmesine neden olur 1 2 3 4 m) Sigara kalp hastahgma neden olur 1 2 3 4 n) Sigaranrn tad1 iyidir 1 2 3 4 o) Sigara insanlann canr s1kkm oldugunda onlarrn 1 2 3 4 yard1mc1 olur p) Eger sigaranm zararlanm 1 2 3 4 anlat1rsam, onlar slgar karan ahrlar 185

PAGE 200

Final Survey (English, April 2002) I Demographics Ncme of the high school : 1 How old are you? .......... .... 2 What is your gender? 1) Male 2) Female 3 What is your parent's total monthly household income ? 1) 50 million and Less 3) 151-300 million 5) 451 million+ 2) 51-150 million 4) 301-450 million 6) Don't know 4 According to you, what economical level are you in? 1) Very poor 2) Poor 3) Medium 4) Well 5) Verywell 5 What is your mother s & father s education ? Education Mother Father 1 Illiterate 2 Can read but no formal schooling 3 Drop out elementary school 4 Elementary school 5 Drop out secondary school 6 Secondary school 7 Drop out high school 8 High school 9 College drop out or 2 years college degree 10 Colleqe graduate 11 Masters I Ph D. 6 About how much money do you have each week to spend any way you want to? 1) None 2) Less than 500,000 3) 500,000-999,999 4) 1,000,000-2 999,999 5) 3,000,000-4 999,999 6) 5 000,000 or more 7. How is your school performance? 1) Excellent 2) Good 3) Average 4) Below average 8 Who do you live with? 1) At home w i th my family 2) At home with relatives 3) At home with friends or alone 4) Dormitory for students 5) Other 186

PAGE 201

9. Where do you come from? 1) Dursunbey 2)Village 3) Other towns 4) Other cities Smoking Status 10. Have you ever smoked a cigarette even lust one or two puffs? (WHO) 1) No (Skip to question 24) 2) Yes 11. How old were you when you smoked a whole cigarette for the first time? -----12. Do you smoke now? 1) No 2) Rarely less than once a month 3) Once in a while but less than once a week 4) Sometimes less than once a day 5) I smoke at least one cigarette a day 6) I smoke 5-6 cigarettes a day 7) I smoke close to one pack a day 8) I smoke more than one pack a day 13. With whom did you first try a cigarette? 1) Alone 2) School friends 3) Other friends 4) Relatives 5) Father/ mother 6) Other 14 If you h ave ever smok ed d i entify the main reasons you started smokinq: Yes No 1) It makes people took impressive 2) People in my house smoked 3) To imitate adults' behavior 4) To imitate friends' behavior 5) To see how it is 6) I was encouraged by friends 7) To qain acceptance by friends 8) I saw smoking on TV, videos, and the internet 9) To demonstrate anger or to protest against parents/teachers 10) Other (please indicate) 15. Have you continue smoking, after your first experience? 0) No (skip to question 16) 1) Yes 15a. What was the reason for continuing smoking, after your first experiences? (check all that apply) 1) To look powerful 2) To look grown up 3) To imitate the behavior of others 4) Encouraged by others 5) To gain acceptance by friends 6) It helped me to solve problems 7) It's not allowed (for protesting parents I teachers) 8) Addicted 9) To release tension 10) Enjoyment 187 No Opinion

PAGE 202

16. How many cigarettes have you smoked in your entire life? 1) 1-2 puffs only 2) 1-10 cigarettes 3) 11-100 cigcv-ettes 4) More than 100 17. During the past 30 days, on how many days did you smoke cigarettes? 1) 0 days 2) 1 to 2 days 3) 3 to 5 days 4)6 to 9 days 5) 10 to 19 days 6) 20 to 29 days 7) All30 days 18. How frequently have you smoked cigarettes during the past 30 days? (WHO) 1) I did not smoke in the last 30 days 2) Less than a one cigarette 3) 1-5 cigarettes per day 4) 6-10 cigarettes per day 5) 11 to 20 cigarettes per day 6) More than 20 cigarettes per day 19. How old were you when you started smoking at least once a week? 1 ) 7 years old 2) 8-9 years 3) 10-11 years 4)12-13years 5) 14-15 years 6) 15 years or older 7) I smoke less than once a week 8) I did not continue after I tried 9) Not applicable 20. How long can you go without smoking before you feel like you need a cigarette? (WHO) 1) Less than one hour 2) 1 to 3 hours 3) More than 3 hours but less than a day 4) A whole day 5) Several days 6) A week or more 7) I do not smoke 21. Have you ever tried to stop or cut down on your smoking 1) No, I smoke once in a while 2) No, I smoke regularly, but have not tried to stop or cut down 3) Yes, I was able to stop or cut down 4) Yes, but I was not able to stop or cut down 22. During the past 30 days, what brand of cigarettes did you usually smoke? (Choose only one answer) 1) Foreign 2) Turkish 3) Both brands 4) I did not smoke past 30 days 188

PAGE 203

23 Why do you like best that brand? (check all that apply) 1) My friends smoke the same brand 2) My parents smoke the same brand 3) It's the brand I can afford to buy 4) I like this brand because of hard taste 5) I like this brand because of soft taste 6) I believe this brand is Jess harmful to my health 7) I feel more elegant I cool with this brand 8) I enjoy the advertisement of this brand 9) My friends give it to me Attitudes toward smoking 24. Do you smoke now? 0) No 1) Yes 24a. I don't smoke because (check all that apply) 1) I don't like the smell and taste 2) It is to my health 3) It is too expensive 4) None of my friends smoke 5) None of my family members smoke 6) Other 25 Do you think you will be smoking five years from now? 1 ) I definitely will 2) I probably will 3) I probably will not 4) I definitely will not 26. Why do you think people your age stcrt to smoke? (check all that apply) 1) Their friends smoke 2) Their mother or father smokes 3) Their brother or sister smokes 4) They want to look important by this way 5) To look grown up 6) Their friends put pressure on them 7) For relaxing 7) Curiosity 8) Because it is not allowed (for protesting parents and teachers) 9) For something to do 11) to look cool 27 Does your father smoke cigarettes? 1) No 2) Yes 3) Don't know 28 Does your mother smoke cigarettes? 1) No 2) Yes 3) Don't know 29 Do any of your brothers or sisters smoke cigarettes? 0) I do not have any brothers or sisters. 1) No 2) Yes 3) Don't know 30. Are there other people smoking in your household other than your mother, father, and siblings? 1) At least one person 2) None 3) Don't know 189

PAGE 204

31. How often do you actually see teachers at your school smoke on schoo l property? (COMMIT) 1) Every day 2) Very often 3) Occasionally 4) Very seldom 5) Never 32 In your opinion what percentage of adult men do you think smoke cigarettes? 1) less than 10% 2) 10 to 25% 3) 26-50% 4) 51-75% 5) 76-90% 6) More than 90% 33 In your opinion what percentage of adult women do you think smoke cigarettes? (COMMIT) 1) less than 10% 2) 10 to 25% 3) 2650% 4) 51-75% 5) 76-90% 6) More than 90% 34 Do your parents know that you smoke? 1) I don't smoke 2) No 3) Yes 35. How does your father feel about your smoking? (AMC) 1) Don't approve 2)Approve 3) Not care 4) Not sure 36 How does your mother feel about your smoking ? (AMC) 1) Don't approve 2) Approve 3) Not care 4) Not sure 37. How wrong would most adults in your neighborhood, or the area a r ound where you live, th i n k smoking is for boys of your age: (CDHC) 1) Not wrong at all 2) A little bit wrong 3)Wrong 4) Very wrong 38. How wrong would most adults in your neighborhood, or the area around where you live thin k smoking is for girls of your age : (CDHS) 1 ) Not wrong at all 2) A little bit w r ong 3) Wrong 4) Very wrong 39 Think about people your age who smoke cigarettes. How many of them usually try to hide their smoking from adults? 1) None 2)Afew 3) About half 4) Most 5)AII 6) Don t know 190

PAGE 205

40. Do you smoke more often when you are alone or when you are with friends? (smoking is a social behavior) 0) I do not smoke 1) More often alone 2) More often with other people 3) About the same 41. H ow many of your four closest friends smoke cigarettes? (WHO) O)_None 3) Three 1)0ne 4) Four 2)Two 9) Not sure 42. How do you think your best friends might feel about you smoking on a daily basis? 1) Disapprove 2) Approve 3) Not care 4) Not sure 43. Has anybody offered you a cigarette? 1) Yes 2)No 44 Do you feel difficulty refusing a cigarette that offered by your friends? 1) Yes 2) No 3) Not sure 45 What happens if you refuse cigarette when offered by your friends? (check all that apply) 1) They may think I am like child 2) They do not accept me as a friend 3) I might lose my influence on group 4) I may be excluded from group 5) They do not care 6) They respect me for this behavior 7) Other 46 In your opinion, what percentage of students in your grade at this school have tried smoking cigarettes, that is, smoked at least a puff of a cigarette? (AMC) 1) less than 10% 2) 10 to 25% 3) 26-50% 4) 51-75% 5) 76-90% 6) More than 90% 47. In your opinion, what percentage of students in your grade at this school smoke cigarettes at least once a week? 1) less than 10% 2) 10 to 25% 3) 26-50% 4) 51-75% 5) 76-90% 6) More than 90% 191

PAGE 206

48 PI ease mar k th h th b d e C OICe at est "b h escn es owmuc h you agree or d" 1sagree wi th h f II t e o owing statements Strongly Agree Not Sure Disagree Strongly Agree Disagree Having an occasional cigarette is not risky (EC) Smoking helps relieve boredom (RI) Smoking helps peo_ple to relax (KAB) Cigarette is a need like eating and drinking It's safe to smoke for a year or two as long as you quit after that Cigarette is social ice breaker 49. Have you ever purchased single cigarettes, sold out of a pack one at a time? (PRELIMINARY STUDY) 1) I do not smoke 2) No 3) Yes 50. Where do you usually get your cigarettes? 1) I buy them at a small grocery store I corner store 2) I buy them at a supermarket 3) I buy them from a friend 4) I buy them on the street 5) My friends give them to me 6) A family gives them to me 7) I take them from home 8) I do not smoke 51. Do any of your teachers or family members ask you to buy cigarettes for them? (PRELIMINARY STUDY) 1) No 2) Yes Media Influences 52. When you watch TV, videos, or go to movies, how often do you see actors, cartoon characters or athletes smoking cigarettes or using other tobacco products? (WHO, AMC) 1) I do not watch TV, video, or go to movies 2) Most of the time 3) Sometimes 4) Hardly ever 5) Never 6) Not sure 53. During the past 30 days, how many times have you seen or heard anything on TV, the Internet, or radio about the DANGERS of cigarette smoking? 1) None 2) A few times in the past 30 days 3) 1-3 times per week 4) Daily or almost daily 5) More than once a day 54. Have you ever seen a cigarette or tobacco advertising? (check all that all that apply) 1) No, I have not seen any 2) Posters 3) Advertisements in public places 4) Logo on clothes 5) On racing cars 6) Music concerts or cinema films 7) Somewhere elseplease describe 192

PAGE 207

55. Do you have communication problems with your family? 1) No 2) Yes (below check all that apply) 1) I am under stress due to my parent's close control (pressure) 2) I cannot talk with my mom 3) I cannot talk with my dad 4) My parents does not trust me 5) My parents does not ask my opinion in family issues 6) I am not in a good communication with my brother(s)/sister(s) 7) My father and/or mother smaches me 8) My parents often fights with each other 9) I am living with my relatives and cannot communicate with them 10) My parents are not tolerant for my friendships 11) My parents does not accept my close friendship with other sex 56 Do you use internet? 1) No 2) Yes 57 Do you prepare for college exam? 1) No 2)Yes 193

PAGE 208

Final Survey (Turkish, April 2002) I A. Kiisel Bilgiler Okulun ad1: 1. Kar;: ____ 2. Cinsiyetiniz 1) Erkek 2) KIZ 3. Ailenize giren ayhk para miktan ne? 1) 50 milyon ve 3) 151-300 milyon 5) 451 milyon ve i.izeri 2) 51-150 milyon 4) 301-450 milyon 6) Bilmiyorum 4. Size gore ekonomik durumunuz nas11? 1) t;ok yoksul 4) iyi 2) Yoksul 5) t;ok iyi 3) Orta 5. Annenizin ve babamzm egitim durumunu, birer r;:arp1 koyarak belirtiniz. Emin degilseniz ola rak cevap verebilirsiniz. Egitim durumu Anne 1 Okur-yazar degil 2 Okumasl var ama okula 3 ilkokul terk 4 ilkokul mezunu 5 Orta okul terk 6 Orta okul mezunu 7 Lise terk 8 Lise mezunu 9 Oniversite terk va da 2 vllhk vi.iksekokul mezunu 10 4 y1lhk i.iniversite mezunu 11 Oniversite sonras1 y_i.iksek lisans veya doktora vapm1s I 6. Zorunlu harcamalar istediginiz harcayabileceginiz haftahk harr;:hgm1z ne kadar? 1) Hir;: yok 4) 1,000,000-2,999,999 aras1 2) 500 bin ve az1 5) 3,000,000-4,999,999 aras1 3) 500,000-999,999 aras1 6) 5,000,000'den fazla 7. Derslerinizin durumu nas1l? 1) yok iyi 2) iyi 8. Nerede 1 ) Evde ailemle 2) Evde akrabalanmla 3) Evde veya yalmz 9. Dursunbey'e nereden geldiniz? 1) Dursunbeyliyim 2) Koyden 3) bir ilr;:eden 4) bir B. Sigara Kullamm1 3) Orta 4) Kotu 10. Sigara ir;:meyi b!r nefes bile olsa denediniz mi? 4) Ogrenciler ir;:in yurt veya otelde 5) Diger 1) Hay1r (DiKKAT:Yamtlmz HAYIR ise liitfen 24. sorudan itibaren cevaplamaya devam edin!z) 2) Evet 11. Ilk bi.iti.in bir sigaray1 ir;:tiginizde kar;: ___ 194 Saba j

PAGE 209

12. !?imdi sigara irriyor musunuz? 1) Hayrr 2) Nadiren ayda birtaneden az 3) Arada srrada irrerim ama haftada bir taneden az 4) Bazen ic;erim ama gunde bir taneden az 5) Her gun en az bir sigara ic;erim 6) Olinde 5-6 sigara irrerim 7) Giinde 1 pakete yakrn sigara iyerim 13. ilk sigaranrzr kiminle birlikte denediniz? 1) YalntZ 2) Okul 4) Diger 5) Akrabalarrmla 6) Babamla veya annemle 7) Diger 14. Sigarayr ilk denemenizdeki sebepler nelerdi? (Cevaplarrnrzr her madde Evet/Hayrr/Fikrim yok birine fi'Srpr koyarak belirtiniz) Evet Hayrr Flkrim yok 1) lnsanlarr etkileyici gt;sterdigini d0$i.ini.iyordum 2) Ailemdeki insanlar ic;iyordu 3) Yetiskinlerin taklit etmek icin 4) Arkada$1anmm davrant$1annr taklit etmek il;'in 5) Nasrl oidugunu gormek ic;in 6) tarafrndan edildim 7) Arkada$1anmdan kabul gormek il;'in 8) Ofkemi gostermek ya da ailemi, ogretmenlerimi protesto etmek il;'in 9)_ Sigara ic;:enleri Teievizyonda ve videoda gordugum ic;:in 1 0) Diger (liitfen yazrnrz) : 15. Ilk denemenizden sonra, devam etmenizin sebepleri nelerdi?(Birden rrok cevap 1) Daha gurrlu gorunebilmek ic;in 2) Daha buyuk goriinebilmek irrin 3) <;evremdeki gibi davranmak icrin 4) tarafrndan edildim 5) tarafrndan kabul gormek iyin 6) Sorunlan criizmeme yardrmct oldugu irtin 7) izin verilmediginden (Anne babama/ogretmenierime kar'i' c;rkmak ic;in) 8) All'ikanhk yaptrgr irrin 9) Srkrnttmr giderebilmek iyin 10) Eglence amacryla (zevk akhgrm iyin) 11) Ilk denememden sonra sigara ic;meye devam etmedim 16. Bugune kadar toplam kac; sigara ic;tiniz? 1) Sadece 1-2 nefes ic;tim, butiin sigara ic;medim 2) 1-10 sigara 3) 11-100 sigara (Yanm paket ile 5 paket arasrnda) 4) 5 paketten crok 17. Gec;tiimiz ay ic;inde msYn sigara ic;tiniz? 1) o gun 2) 1-2 gun 3) 3-5 gun 4) 6-9 gun 5) 10-19 gun 6) 20-29 gun 7) Bir ay iyinde her gun 195

PAGE 210

18 ay, gQnde orta!ama kac sigara icrtiniz? 1) bir ay(30 gUn)icrinde sigara icrmedim 2) GUnde bir sigaradan az 3) GUnde 1 sigara 4) GUnde 2-5 sigara aras1 5) GUnde 6-10 sigara aras1 6) Giinde 10-20 sigara aras1 7) GUnde 20 sigaradan 19 En az haft ada bir kere sigara kacr 1) 7 civarmda 5) 14 15 2) 8 9 6) 15 yada sonra 3) 10 11 7) Haftada bir kereden az iyiyorum 4) 12-13 20. Sigaraya ihtiyay duymadan ne kadar sUre dayanabilirsiniz? 1 ) Sir saatten az 2) 1 ile 3 saat aras1 3) 3 saatten fazla ama bir gOnden az 4) 1 gUn 5) Sir ka9 giin 6) Sir hafta ya da daha fazla 8) Arada bir i9i n hi9 ilfmeden de durabilirim 21. Sigaray1 b1rakmay1 hi9 denediniz mi? 1) Arada bir i9tigim icrin denemedim 2) Evet, bir sUre ve sonra b1rakt1m 3) Evet b1rakmay1 denedim ama b1rakamad1m 4) Hay1r, b1rakmay1 denemedim 22. Son 30 giin i9 i nde hangi marka sigara iytiniz? (Sadece bir cevap seyiniz) 1) Yerli marka (belirtiniz) .............. .... .............. 2) Yabanc1 marka ............... ....... .. 3) 30 giin icrinde sigara i 9medim 23. Neden bu markay1 kullamyorsunuz? (Dikkat:Birden fazla cevap 1) ayn1 markay1 kuiian1yor 2) Ailem bu markay1 kuliamyor 3) Satm alabildigim marka 4) Bu markay1 sert oldugu i9in seviyorum 5) Su markay1 hafif oldugu i9in seviyorum 6) Su saghg1m iyin daha az zararh oldugundan 7) Su marka ile daha etkileyicil havah oldugumu hissettigim ir;;in 8) Bu markanm reklamlarm1 begeniyorum 9) ikram ediyor C. Sigara Konusunda 24. l;)u anda sigara icrmiyorum 90nkii (Birden fazla cevap i:?aretieyebilirsiniz) 1) Kokusunu ve tad101 sevmiyorum 2) Saghg1ma zararh 3) yok pahah 4) bir icrmi yor 5) Ailemde hi!f kimse icrmiyor 6) nedenlerden (liitfen yaziniz) : _______ 7) Sigara ir;;iyorum 25. Sundan 5 y1l sonra, sigara ir;;iyor oiacagJOJZI musunuz? 1 ) Kesinlikle i9erim 2) Belki iyerim 3) Selki 4) Kesinlikle iymem 196

PAGE 211

26. Sizin ya!jlmzdaki insanlar neden sigara ba!jlar? (Birden c;ok cevap i!jaretleyebilirsiniz) 1) Arkada!jlannm sigara ic;mesinden dolay1 2) Anne ya da babalan iytigi iyin 3) Karde!fleri iytigi i.yin 4) bnemli gor!inmek 5) Kendilerini biiyiik gostermek i.yin 5) Arkada!jlan zorlad1g1 6) Rahatlamak iyin 7) Merak ettilderi iyin 8) Aile ya da iigretmenlerine gelmek 9) yapml!f olmak iyin 10) Havah giiziikmek ic;in 11) Diger (liitfen yazm). _______ 27. Babamz sigara it;iyor mu (Arada bir bile olsa)? 1) Haytr 2) Evet 3) Bilmiyorum 28. Anneniz sigara it;iyor mu (Arada bir bile olsa)? 1) Haytr 2) Evet 3) Bilmiyorum 29. Erkek ya da k1z sigara ic;en var m1? 0) yok 1) Hay1r 2) Evet it;en var 3) Bilmiyorum 30. Anne, baba ve karde!jler dl!jlnda, sizle aym evde ve sigara i.yen kimse var m1? 1) Haytr, yok 2) Evet var 31. Ogretmenlerinizin okutda sigara iytigini hangi s1khkta gi:iriiyorsunuz? 1) Her giin 3) Arada bir 5) Hit; gormiiyorum 2) ;:ok s1k 4) Nadiren 32. Sizce yeti!jkin erkeklerin yiizde kac;1 sigara ic;iyordur? 1) %10 dan az1 3) %26-50 5) %76-90 2) %10-25 4) %51-75 6) %90 dan fazlas1 33. Sizce yeti!jkin kad1nlann yiizde kac;1 sigara ic;iyordur? 1) %10 dan az1 3) %26-50 5) %76-90 2) %10-25 4) %51-75 6) %90 dan fazlas1 34. Aileniz sigara iytiginizi biliyor mu? 0) Sigara iymiyorum 1) Haytr 2) Evet 35. Babamzm sizin sigara iymeniz konusundaki tutumu ne I ne olurdu? 1) Onaylamaz 3) Umursamaz 2) Onaylar 4) Emin degilim 36. Annenizin sizin sigara it;meniz konusundaki tutumu ne I ne olurdu? 1) Onaylamaz 3) Umursamaz 2) Onaylar 4) Emin degilim 197

PAGE 212

37. c;evrede sizin erkeklerin sigara if(mesi nasrl 1) Normal 2) Biraz oldugu 3) oldugu 4) c;:ok oldugu 38. c;evrede sizin krzlarrn sigara ic;mesi nasrl 1) Normal 2) Biraz oldugu 3) oldugu 4) c;:ok oldugu 39. Sizin sigara ic;en arkadaslarrnrzrn kac; tanesi sigara if(tiklerini bOyOklerden saklamaya 1) Hie; biri 4) Yarrdan fazlasr 2) Bir kac; tanesi 5) Hepsi 3) Yarrsr 6) Bilmiyorum 40. Genellikle yalnrz mr yoksa mr sigara ic;iyorsunuz? 1) Genellikle yalmz 2) Genellikle 3) Farketmiyor 41. DOrt yakrn kaf! tanesi sigara ic;iyor? 0) Hie; biri 3) Oc; 1) Bir 4) Dort 2) iki 9) Emin degilim 42. Yakrn sigara ic;meniz konusunda ne 1) Onay!amazlar 3) Umursamazlar 2) Onay!arlar 4) Emin degilim 43. Size hie; sigara ikram eden oldu mu? 1) Evet 2) Hayrr 44. tarafrndan sigara tutuldugunda reddetmek zor olur mu? 1) Hayrr, istemedigim rahatf(a reddederim 2) Evet reddetmekte zorlamrrm 3) Emin degilim 45. tuttugu sigarayr geri c;evirirseniz ne olur? (Birden c;ok cevap 1) c;:ocuk oldugumu 2) Beni olarak kabul etmezler 3) Gruptaki etkinligimi kaybederim 4) Gruptan 5) Kimse umursamaz 6) Bu iyin bana saygr duyarlar 46. Sizce okulunuzda 10. srnrflarda kay denemi!lltir (bir nefes bile olsa)? 1) %10 dan azr 4) %51-75 2) %10-25 5) %76-90 3) %26-50 6) %90 dan fazlasr 47. Sizce okulunuzda 10. srnrflarda kay en az haftada bir kere 1) %10 dan azr 4) %51-75 2) %10-25 5) %76-90 3) %26-50 6) %90 dan fazlasr 198

PAGE 213

48. ifadelere katlltp katllmadtQiniZI belirtiniz? (Cevaplanmzt her madde yarpt koyarak belirtiniz) Kesinllkle Kahhyorm Kararsz Katlmtyorum katthyorm Arada bir sigara iymek zararh degildir Sigara insanlara stktldtklannda yardtrnct olur Sigara insanlann rahatlamasma yardtmct olur Sigara yemek icmek gibi bir ihtiyactir Bir iki ytl iyip sigarayt biraktrsantz bir zaran olmaz Sigara insanlann birbirine yardtmct olur 49. Market ya da bakkaldan paket tek sigara satm aldtntz mt? 1 )Sigara iymiyorum 2)Haytr 3)Evet 50. SigarantZI en yak nereden ahyorsunuz? 1) Bakkaldan 5) veriyor 2) Si.ipermarketten 6) Bir aile i.iyesi veriyor 3) satm ahyorum 7) Evden ahyorum 4) Caddeden satm ahyorum 8) Sigara iymiyorum 51. Etraftntzdaki biiyiikleriniz sizi sigara almaya gonderiyor mu? 1) Haytr 2) Evet Keslnllkle katllmtvor 52. Televizyon, video izlediginizde, ne kadar stkhkla aktorleri, yizgi film kahramanlannt ya da sporcutan sigara iyerken goriiyorsunuz? 1) TV ve video izlerniyorum 2) <;:ogu Zaman 3)Bazen 4) Neredeyse hie; gormiiyorum 5) Hie; gormi.iyorum 6) Fikrim yok 53 Gec;tigimiz 30 giin ic;inde kac; kere TV, radyo, ya da lnternette sigaranm zarartan/tehlikeleri ije ilgili bir program gordiiniiz? 1) Hie; gormedim 2) Ay bir kac; kere 3) Haftada 1-3 kere 4) Neredeyse hergiln 5) Giinde birden fazla 54. Herhangi bir yerde sigara reklamt gordi.iniiZ mii? Nerede? (Birden yok cevap 1) Haytr gormedim 2) ve posterterde 3) Halka ac;tk yerlerdeki reklamtarda (Duvar veya ilan tabelalannda)gordiim 4) Bazt giysilerin iizerindeki resim ve yazllarda 5) Gazete veya dergilerde 6) arabalannm iistiinde 7) Konser yada filmlerde 7) Diger (lutfen yaziniz):. _________ 199

PAGE 214

cOmlelerden sizin it;:in uygun olanlan (Birden t;:ok cevap 1) Ailemin bask1s1 yOzOnden stres altmday1m 2) Annemle rahat 3) Babamla rahat 4) Ail em ban a gOvenmiyor 5} Ailem aile it;:i kararlarda bana sormuyor 6) iyi sa hip degilim 7} Annem veya babam beni dOv\iyor 8} Annemle babam s1k s1k kavga ediyorlar 9) Akrabalanmla ve onlarla kuram1yorum 10)Ailem esneklik gostermiyor 11)Ailem karsi cinsle yakm izin vermiyor 12)Ailemle hit;: bir sorunum yok 56. Do you use internet? 1} Hayir 2) Evet 57. Do you prepare for college exam? 1) Hayir 2} Evet NOT: Bu ile ilgili olarak ytizyOze yapmak istiyorum. Eger ilgi duyarsan1z ve tecrObelerinizi isterseniz IOtfen isminizi yazm sizinle temasa get;:ecegim 200

PAGE 215

Semi-structured Interview Guides Student Interview Guide (English) Introduction: Hi! My name is Hulya Yuksel. I am interviewing with students in your school today. I am not interested in reporting your names to the school management. am interested in your ideas, beliefs and knowledge about smoking. When did you start smoking? 1. Tell me about how you started smoking? I For nonsmokers, why do they think their peers smoke? 1 a: What do you think influenced you to begin smoking/not smoking? 1 b: What do you think influences you to keep smoking/not smoking? 2. What are the reasons you smoke? Probe: "What do you see as the benefits of smoking? What about disadvantages? 3. What people around you smoke? How does that influence you? (parents /friends) 4. Can you tell me about a typical day of smoking for you? (when, where and with whom you smoke?) During the interviews the following questions were added to the interview script: 5 Do you see smoking as a problem? 6. Do you talk about cigarettes in friend groups? Can you give examples? 7. When do people smoke cigarettes more? 8. What does smoking mean in your social environment? 9 Do you know songs with the cigarette lyrics? 10 Do you think seeing smoking in movies/on TV cause smoking among students? 11. Which TV shows do you watch? Do you see smoking in them? 12. Are you aware of any anti-tobacco ads? Student Interview Guide (Turkish) lsmim Hulya Yuksel. Bugun sizin okulunuzda ogrencilerle gorusmeler yapiyorum. Sizin isimlerinizi okula rapor etmekle ilgilenmiyorum. Sizin sigara ile ilgili bilgi ve dusuncelerinziden yararlanmak istiyorum. Sigara icrneye ne zaman basladiniz? 1. Sigaraya nasil basladiniz? 1a: Baslamanizda/ baslamamanizda sizi ne etkiledi? 1b: Sigara devam etmenizdeki sebep ne? 2. Sigara icme nedenleriniz neler? Probe: "Sigaranin faydasi/ sagladigi avantajlar ne? Sigaranin size zarari ne? 3. Anne babaniz iciyor mu? Baska kimler iciyor ailenizde? Bunun sigara icmenizdeki etkisi ne? Arkadaslarinizin sigara icrnenizdeki etkisi ne? Baslamaniz icin arkadaslarinizin etkisi ne oldu? 201

PAGE 216

4. Siz en cok ne zaman sigara iciyorsunuz? Nerede? Sigara icmeyi en cok ne zaman seviyorsunuz? En cok kimlerle iciyorsunuz? 5. Sigarayi problem olarak goruyor musunuz? 6. Arkadas gruplarinda sigara ile ilgili konularda konusuyor musunuz? Ornek verebilir misiniz? 7. lnsanlar ne zaman en cok sigara icer? 8 Sosyal cevrenizde sigaranin anlami ne? 9 lcinde sigara sozu gecen sarkilar biliyor musunuz? 10. Sigara ile ilgili yayinlari TV 'de goruyor musunuz? Bu tur yayinlari izlemek yasitlarinizin sigara icrnesine neden olabilecegini dusunuyor musunuz? 11. TV de izlenen filmier? Bu filimlerde sigara icenleri goruyor musunuz? 12. Sigaraya karsi yapilan reklamlari goruyor musunuz? Principal Interview Guide (English) 1 What kind of rules do you apply in your school? 2 Can students smoke cigarettes in school? Do they get punished if they smoke? 3. Do you think the rules you have in your school enough to prevent students from smoking? 4. Do you perceive any increase in smoking rates compare to 10 years ago? 5. Do you think the health dasses your school offers satisfactory? 6 What are the social and cultural factors affect smoking uptake? 7. Do you think it is an acceptable behavior to smoke by high school youth? Principal Interview Guide {Turkish) 1. Okulda sigara ile ilgili ne gibi kurallar uyguluyorsunuz? 2 Ogrenciler okulda sigara icebiliyor mu? Sigara icenler cezalandiriliyor mu? 3. Sigara icme oranlarinda sizce 10 yil oncesine gore bir artis var mi? 4. Sizce okulda sigarayi kontrol etmek icin konulan kurallar ogrencileri sigaradan uzaklastirmada etkili oluyor mu? 5 Sizce saglik derslerinde verilen egitim yeterli mi? 6. Sizce ogrencilerin sigaraya baslamasinda etkili olan sosyal ve kulturel factorler neler? 7 Sizce lise ogrencilerinin sigara icmesi toplum tarafindan Kabul edilen bir davranis mi? 202

PAGE 217

APPENDIX B FIELD SEASON ONE QUALITATIVE AND QUANTITATIVE FINDINGS, SUMMER 2001 203

PAGE 218

Field Season One Study Methods There were three phases of the preliminary data collection. In the first phase, a focus group was formed to get a sense of teenager's attitudes and beliefs about smoking, as well as their smoking/nonsmoking behaviors. This phase guided the construction of the survey questionnaire. In the second phase, the sixty-two item survey questionnaire was administered in Dursunbey and lzmir, Turkey to total of 247 high school students. In the third phase, semi-structured interviews conducted with 32 of these high school students in the urban area of lzmir. Pilot Study. After the completion of the questionnaire a pilot study of the questionnaire was administrated to five students. This pilot study helped the researcher understand whether the subjects understand the questions differently or do not understand particular questions. Moreover, the researcher tried to understand how people interpret the questionnaire's questions and concepts, and on what understandings they base their responses. According to subjects' explanations and interpretations of the questions, the necessary modifications was be made to the questionnaire. The administration time was between 25-30 minutes. Sample. All the high schools in Dursunbey were surveyed. Total of 17 4 student were surveyed in Dursunbey: Religious high school (n=37), Health Vocational School (n=21 ), Regular High School Industrial Vocational High School (n=70). Four high schools, all the 10 graders who are available during the administration of the survey were surveyed. In addition, one class of gth graders from Industrial Vocational High School and one gth graders in Regular High School, and one Class of gth and 11th graders from Religious high school were surveyed. For the survey administered in lzmir, students (n=43) from two 1Oth grade classes of the Karsiyaka Regular High School and 32 students from various EVKA-4 high schools were surveyed. The administration procedures were identical to those followed in Dursunbey. All the semi-structured interviews were conducted in lzmir. The set of subjects for the interviews was a subset of the subjects in the second phase, where the parental consent forms covered the contingencies for both study phases. Students participating in the first phase were chosen by convenience sampling for participation in the second phase. Thirty-two 204

PAGE 219

teenagers from various grades (91 h, 1Oth, 11th) were interviewed, with the majority of them 1Oth graders. These teenagers were from EVKA-4 Regular High School (n=16), lzmir Industrial Vocational High School and high school (n=6), and high school drop outs (n=B). Karsiyaka Regular High School (n=2). Smokers were represented by higher numbers than nonsmokers in the semi-structured interviews. Females and males segments of the sample were at rough parity (14 girls and 18 boys). Qualitative Findings Semi-structured interviews were coded by key themes and concepts. Emerging themes in the data were coded based on the literature review. The following categories defined for reasons for smoking initiation and maintenance: (1) social or interpersonal factors (Parental smoking, peer influence, emulating parent and friends, social status symbol--brand preferences-, impressing others, problems with family, parental control, symbol of friendship, minority access, perceived prevalence of smoking) ; (2) intra personal factors (problems with school, self-medication as medium of reducing stress, pleasure, weight control); (3) attitudes and beliefs about outcome expectancies (perception of health risks). (4) Cessation. Reasons for not smoking also coded (it's bad for my health, it's gross, my friends don't smoke). Outcome measure: Smoking status Although the qualitative analysis captured the subject's formally defined smoking status (smoking even one cigarette within thirty days of the measurement), teenagers often did not see themselves as smokers. Various reasons were given: occasionally smoking did not constitute smoking in their view, or if they do no buy the packs themselves. They also do not see themselves as smokers if they do not inhale. Many say that the practice of inhaling qualifies as real smoker, "if you don't inhale you are faking it." Some of the teenagers said that "if you don't inhale you will get throat cancer''. 205

PAGE 220

Reasons Given for Smoking. The teenagers in this study gave a wide variety of reasons for smoking: as an antidote for boredom; to satisfy curiosity, to alleviating stress, to aid in concentration, to impress girls and flirt with them, to look rich and tough, to impress others, to prove that they are grown ups. For girls, weight control was given as a practical reason for smoking. Other reasons given were peer influence, family influence, emulating friends and adults, and problems with family and school. The following quote is mentioned by a male smoker during the interview as a common Turkish say i ng : "The real man's mouth smells of alcohol and cigarettes." Social or Interpersonal Factors. Parental smoking and monitoring. Parental involvement in their adolescent's lives is an important factor for keeping them off cigarettes. The subjects indicated that parental talks about the harms of smoking were important influences on their decision to smoke or not to smoke, but this message was often undercut if the parents themselves smoked, as the teenagers often found this hypocritical Actual punishment for smoking appeared to be counterproductive. Peer Influence, Emulation (Ozenti). Many smokers mentioned that they smoke more when they are socializing with friends. Also many mentioned peer influence or pressure as a reason for initiation. Usually, smokers within a group put pressure (or encouragement) on non-smoker friends by stressing that their peers should "light one up." Sharing a pack of cigarette is also a symbol of friendship. Social status symbol impressing others. Many subjects indicated that preferred brands were symbols of affluence and discriminating taste, particularly foreign brands. The Marlboro brand was frequently cited for having this fashionableness. Here is one example about symbolic meaning of smoking: "Foreign brands use better quality tobacco with a nicer and fuller taste. I wouldn't even touch Turkish cigarettes. I would break them. I have no use for them Foreigners know their business, their machines are clean, Tekel (Turkish State Tobacco Monopoly) has dirty machines." 206

PAGE 221

Minors' Access. There are small markets near high schools that sell single cigarettes. Since teens have limited resources this makes smoking more accessible (although they also pool their resources to buy a pack). In these small markets, no one asks minors for IDs even though the minors are clearly underage and sales to minors is illegal. Part of the explanation for this is that older, majority-aged youth often send youngsters to buy cigarettes for them, as do parents, and, as one subject reported, even teachers. Perceived prevalence of smoking. As one subject exclaimed, "Everybody smokes!" Some subjects in the semi-structured interview claimed that girls smoke more than boys. In the literature, perceived prevalence of smoking is used as predictor of youth smoking onset and increase in smoking. Adolescents overestimate the prevalence of cigarette smoking among adolescents and adults compared with the actual statistics (Nichter et al., 1997: Flay et al., 1998; Carvajal et al., 2000). Control. Many girls indicate as a reason for smoking is that their family puts too much pressure on them. Traditionally boys have more freedom in Turkish culture, so I did not hear this complaint from them. For males, the data indicates that it is more of a manhood issue, indicated by the dominance of males in the smoking population (63% male, 20% female population in Turkey) and the male subject's emphasis on acting "grown up." For both sexes smoking might mean that they have control over their life. Although their families live in very small rooms limiting their personal freedom, they might at least be enjoying the freedom of smoking as being in charge of their lives. lntrapersonal Factors Self-medication: Cigarettes are perceived as stress reduction tools. Many of the teenagers said, "cigarettes are number one medication for sinir"(it means "nerve" in English, so this means nervousness, irritability). Turks use "being sinirli" as a characteristic of a person in Turkish culture. It is usually used to rationalize one's behavior. This suggests that alternate methods of dealing with stress might have some effect on smoking maintenance or even uptake. Yoga might be a candidate for such as alternative, especially since it can be socially marketed in much the same terms as cigarettes: something fashionable that gives one personal power. 207

PAGE 222

Stress related to family problems Family pressure was especially felt by girls. They experience lack of freedom for spending time with friends and going out. Stress related to school work. Many students mentioned school and work as stressors, and smoking as a way of reducing this stress. Addiction. Occasional smokers claimed that they are not addicted. This reflects that they are not informed about the concept of gradual addiction to nicotine. Regular smokers who already tired quitting mentioned the addictive qualities of smoking. They seemed to readily identify with addictive personalities. Biological differences. Biological factors seemed to play a role in youth smoking. Some of the teenagers stated that although they wanted to smoke, they cannot because they become dizzy, nauseated, or have headaches. This might reflect a lower tolerance for nicotine. Attitudes and Beliefs About Outcome Expectancies: Self-exempting beliefs about tobacco use Perception of health risk. Many teenagers had some awareness that smoking is unhealthy, although most subscribed to one or more mythologies about smoking and health. Some claimed that smoking is harmless as long as you don't smoke too much. Others claimed that the "lite" or menthol cigarettes are not as bad as regular cigarettes: "Menthol cigarettes open your throat, they don't smell and they are not harmful to you. They are ideal for using to quit smoking. There are also cigarettes with chocolate, banana, and strawberry flavors." Above statement echoes the Tobacco firms' misleading the public message about cigarette brands that are lite, low tar, mild or low nicotine. David Kessler ( 1994 }, in a statement to a U.S. district court, wrote that smokers of mild or lite cigarettes did not realize these caused just as much nicotine to be absorbed by their body as regular cigarettes. Some of the teenagers expressed their feelings about smoking as follows: "They write smoking is harmful to your health on cigarette packs. If it is harmful why do they even produce it?" Others said that they would eventually die of something anyway, so why not smoke? 208

PAGE 223

Others claimed that foreign brands were safer than Turkish brands, for example: "Turkish brand Samsun is not good. It causes throat cancer, foreign brands are less dangerous. They have good quality products." Two subjects reported that medical professionals advocated a limited form of smoking. One teenager who had quit smoking reported that a psychiatrist recommended that she resume smoking one cigarette a day to manage stress. Another reported that the same advice was given her mother by a psychiatrist. "If you don't inhale the smoke deeply into your lungs you would get a throat cancer." "They say that cigarettes cause cancer but they've found a cure for it and I'll just pay money for it and be cured." Cessation Cessation. Stress was often cited as a barrier to quitting. Many smokers already tried quitting but relapsed. The ubiquitous smoking environment is also seen as a barrier:" It is hard when everybody smokes around you." Perception of self in the future as smoker. Many smokers mentioned that they would like to quit in the future but right now it is so stressful to do it. Some of them are not planning to quit at all. Reasons For Not Smoking. Nonsmokers indicate the bad smell and harmful effects of smoking as reasons for not smoking. Nonsmokers seemed better informed about the health effects of smoking compare to smokers. 209

PAGE 224

Quantitative Findings Table 1. Characteristics Of Sample By Smoking Status (Self-Identification Of Smokin Status 14.3 98 100 30.4 148 100 246 100 s tOO ... 7 26 9 .26 .... /100 27 24. 6 109 100 19 24.1 79 100 4 17.4 23 100 Grade Grade9 35 57.4 3 4.9 9 14. 8 14 23.0 61 100 Grade10 73 43.7 9 5.4 42 25. 1 43 25.7 167 100 Grade11 14 77.8 1 5.6 1 5.6 2 11.1 18 100 12.2 41 3\).6" 62 ... 20 ; 0 65 43.3 30 15.8 19 500+:. ' 22.2 18 Weekly Allowance <=999,999 15 65 2 1 4.3 2 8.7 5 21.7 23 100 1000,000-2999,999 53 57.6 4 4.3 14 15.2 21 22.8 92 100 3000,000-4999,999 11 47 8 1 4 3 6 26.1 5 21. 7 23 100 5000,000-9999,999 18 39 1 3 6.5 13 28. 3 12 26 1 46 100 >=1 0,000,000 5 26.3 2 10.5 8 42.1 4 21.1 19 100 203 100 73 100 :. 173 100 246 100 Type of School Religious 27 75.0 4 11.1 2 5 6 3 8.3 36 100 Health Voc 12 57.1 1 2 2 3 14.3 6 28.6 21 100 Dursunbey 22 47.8 4 5.7 12 26.1 11 23.9 46 100 Reg HS 28 40.0 4 9.3 14 20. 0 24 34.3 70 100 Industrial 24 55. 8 8 18. 6 7 16.3 43 100 lzmir Reg. 6 37.5 7 43.8 3 18.8 16 100 EVKA-4 2 33.3 1 16. 7 3 50. 0 6 100 lzmir Vocational 210

PAGE 225

Table 2. Smoking Behavior By Current Smoking Status [Smokers Only] 2 22. 2 7 77.8 9 100 2 28. 6 5 71.4 7 100 6 35.3 11 64.7 17 100 7 41. 2 10 58.8 17 100 5 33.3 10 66.7 15 100 20+ 28 68.3 13 31.7 41 100 Brand name Turkish 10 71.4 4 28.6 14 100 Foreign 18 50 18 50 36 100 American Blend Turkish 13 40 6 19 59 4 32 100 Cigarette Mix Brand 10 58 8 7 41.2 17 100 Reasons for preference No preference 4 100 4 100 Price 16 64 9 36 25 100 Quality/taste 18 45 22 55 40 100 Appearance/ brand 6 66.7 3 33.3 9 100 Emulate others 1 12.5 7 87.5 8 100 Would you prefer another brand? No 30 49.2 31 50 8 61 100 Yes 19 59.4 13 40.6 32 100 211

PAGE 226

Table 3. Smoking History And Smoking Behavior By Gender ,,:,,;-!' ',, i'r ;; ,,'''' 'Female ,,. MaJe:r ... TotcJI . . N ... % . N % N 0/11 . ........... ....... .... '' : Current smoking status Non-smoker 65 53. 3 57 46 7 122 100 Ex-smoker 5 38.5 8 61.5 13 100 Occasional smoker 14 26 9 38 73.1 52 100 Regular smoker 14 23.7 45 76.3 59 100 Ever tried Yes 59 32.6 122 67.4 181 100 No 39 60 26 40 65 100 Been offered a cigarette Yes 74 100 130 63.7 204 100 No 25 58.1 18 41. 9 43 100 [SMOKERS ONL Y-N=111] Who encouraged you to smoke 1st cig? 8 28.6 22 26.5 30 100 No one 1 3.6 9 10.8 10 100 Family member or relative 19 67.9 52 62.7 71 100 Friend or friends [SMOKERS ONLYN=11 0) Place tried 1st cig? At home 17 60 7 15 18. 3 32 100 Outside home 11 39. 3 67 81.7 78 100 [SMOKERS ONLYN=110} Who were you with for 1 cig? Noone 6 21.4 11 13.4 17 100 Family member or relative 1 3.6 9 11 10 100 Friend or friends 21 75 62 75.7 83 100 212

PAGE 227

Table 4. Smoking-Related Attitudes By Smoking Status Brand shows how much money people have Strongly agree Agree 33 43.4 3 3 9 19 25 0 21 Disagree 31 50. 8 4 6.6 14 23.0 12 Strongly disagree 19 42 2 4 8.9 12 26.7 10 19 52 8 1 2 8 6 16. 7 10 Foreign brands demonstrate a person's wealth Strongly agree 18 58 1 1 3 2 3 9.7 9 Agree 15 35.7 4 9 5 9 21. 4 14 Disagree 39 50 6 2 2.6 20 26 16 Strongly disagree 28 43 8 4 6 3 20 31. 3 12 If people don't have money they smoke Turkish brands Strongly agree 15 35.7 1 24 13 31 13 Agree 33 56 9 4 6 9 6 10 3 15 Disagree 27 43 5 4 6 5 18 29 13 Strongly disagree 25 49 2 3.9 14 27.5 10 Design of cigarette pack has influence on young people brand preference Strongly agree 18 43 9 3 7.3 10 24.4 10 Agree 28 47 5 1 1 7 18 30 5 12 Disagree 33 50 6 9.1 9 13.6 18 Strongly disagree 19 43 2 1 2 3 14 31. 8 10 Table 5. When Do You Smoke The Most? [Smokers Only] When I am happy When I am with friends Di-Sagree N $ 14 14.5 41 45.5 14 14. 6 213 % 27 6 76 100 19.7 61 100 22 2 45 100 27 8 36 100 29 31 100 33 3 42 100 20.8 77 100 18 8 64 100 31 42 100 25 9 58 100 21 62 100 19.6 51 100 24.4 41 100 20 3 59 100 27.3 66 100 22 7 44 100

PAGE 228

Table 6. Why Do You Smoke Now? [Smokers Only] am grown up For pleasure 14 18 7 30 40 0 14 18.7 17 22.7 From Stress 37 43.0 34 39.5 8 9.3 7 8.1 It helps me to find 3 4.5 4 6.0 27 40.3 33 49.3 friends I look up to my elder 6 8.5 16 22.5 20 28.2 29 40.8 family members I look up to my 15 20.3 20 27.0 19 25 7 20 27 0 friends Table 7. What Are The Main Reasons For People Of Your Age To Start Smoking? [Smokers Only] Don't know Emulating others Peer influence Stress, social, academic, family problem Pleasure 2 To gain social 3 acceptance Curiosity 9 Addiction Any combination of 72 above 28.6 75 60 62 1 25 5 33.3 1 50 18 15 5 214 Jl1!9tl19'r totaL. .,smbtet N 6 46.2 4 23.5 6 23.1 7 30.4 5 71. 4 1 6 7 1 50 26 22.4 7 100 2 100 15 100 2 100 116 100

PAGE 229

Table 8. Self-Identification Of Smoking Status Vs Who's Identification Of Smoking Status N % Non-smoker 113 84 3 11 8 2 10 7.5 134 100 Occasional smoker 2* 3.8 2 3 8 48 92 3 52 100 Daily smoker 3* 5.1 7 11.9 49 83 1 59 100 Total 118 100 20 100 107 100 245 100 *Five respondents provided contradictory statements in two different questions (question 12 and question 16) regarding their smoking status. 215

PAGE 230

,----,-------------------------------APPENDIX C TURKISH SONGS REGARDING SMOKING TWO POEMS BY ONE OF THE STUDY SUBJECTS (English Translation) 216

PAGE 231

Some Turkish Songs Regarding Smoking One Last Cigarette Resembling a stranger Touching with sadness I knew the moment had arrived To say goodbye and depart. Still feeling the lightness of your touch on me, No trace of you left with me The moment had arrived To say goodbye and depart. No farewells, no promises What were we holding on to? Why were we in such a hurry? Let's smoke one last cigarette Go afterwards, if you must go. Smoke it slowly and let it last If you are not coming back. (Singer: Yesim Salkim) * Inhaling the smoke, with every breath I remember you Hair turned gray Eyes filled with tears Looking to the path you took Feeling the endless night Inhaling the smoke, With every breath I remember you. Taking in from the goblet With every sip I remember you. Feeling your hands on mine Feeling your gaze on me How did those years went by? Oh, my beloved, where did you disappear to? 217

PAGE 232

Two Poems by One of the Subjects (English Translation) If See the stream of my endless tears Understand in this vast world No place to rest my head No one to tell of my agony No one to trust whose words No friend to rely on See the stream of my endless tears I feel the burden of loneliness When my never-ending pain intensifies I put my feelings into words And defy my destiny. My beloved stabbed me in the back The times I feel the immense loneliness My cigarette keeps me company It is my friend It is my companion in pain. Understand my endless cries They tell you of a deceitful lover Or they recount friends who have stabbed me in the back * How Hard It Is To Quit! Do you know how helplessness feels? Drowning in my tears Yet I cannot cry. Griping in vain, my hope will not rise again. Thoughts race through my head Driving me to madness I strive to forget But my thoughts don't loosen their grip. Inside me, unfinished worries. In the darkness of the night 218

PAGE 233

I await the dawn. In the lightness of the day I await the dusk. Waiting is a way to distract my mind. No matter what I do, I cannot break the chain of my thoughts They wouldn't let go, no matter what I do I cannot control my desires They are worse than a dagger's pain It's no a pain due to love Nor is it a wound to the heart. It's a battle within me And long ago, I lost the war. I, again, have it in my hand Dammit! It's hard to break Or, throw it away Again I light it up While inhaling the smoke, I think How hard it is to quit! (--Poems by S B., one of the subjects in the study) 219

PAGE 234

APPENDIX D CONSENTS AND HUMAN SUBJECTS APPROVALS Parental Consent Form Student Assent Farms Human Subjects Approvals 220

PAGE 235

PARENTAL CONSENT AND STUDENT ASSENT FORMS Parental Consent Form Dear Parent: I am a Ph.D. student in the US. I am doing a research for my doctoral thesis about youth and smoking. The purpose of this study is to examine knowledge, attitudes, and behaviors of young people about smoking. I am interested in their opinions and experiences about smoking. The results of this study will be used to develop better education programs in the future for young people like themselves. This project will not include any education about tobacco. I hope that you will allow your child to participate to this study. If you agree, please keep this form with you otherwise please sign and give back to your child to bring it back to school. Thank you, I do approve my child's participation to this study _____ I do not approve my child's participation to this study ___ Student Assent form for the Self-Administrated Survey Dear Participant: Thank you for volunteering to partidpate in this study about smoking for my doctoral thesis. The purpose of this study is to examine knowledge, attitudes, and behaviors of young people about smoking. I am interested in your opinions and experiences about smoking. The results of this study will be used to develop better education programs in the future for young people like yourself. This project will not include any education about tobacco. I hope that you will answer all of the questions honestly and thoughtfully. If you agree, you will receive a copy of this form. Your participation consists of the following: The survey will last about 25-30 minutes. Please remember that your participation is completely voluntary and you can decide to cease participation at any time. The benefit of your participation is in your contribution to understanding of smoking behavior among youth that could lead to successful tobacco control efforts. It is possible in exploring your experiences that negative feelings and distress could surface. This is an anonymous survey, so please do not put your name on the questionnaire. No information will be shared with parents, teachers, school officials, friends or others not directly involved in this research project. Please work quietly and by yourself. If you have any questions about this research study, please feel free to ask me or your teacher. My home number is 312-4427. Also, if you have questions about your rights as a partidpant in this research, please call the Office of Academic Affairs in the University of 221

PAGE 236

Colorado Denver Building {Lawrence and 141 \ Suite 700. The phone number is 303-5562550. If you'd like, I 'll be more than happy to share my thesis with you when it s done Thank you for your willingness to participate. Sincerely, Hulya Yuksel Student Assent form for the Self-Administrated Survey (Turkish Translation) Degerli Katthmct : Doktora tezim iyin yapttgtm bu kattllmmtz ic;in Bu c;alt!?manm amact lise ogrencilerinin sigara iyimi ile ilgili bilgi, tutum ve davrant!?lannt olc;mek ve sizin sigara ile ilgili fikirferinizden ve deneyimlerinizden yararlanmakttr BUtOn sorulan ictenlikle cevaplayacagmtzt umut ediyorum. Eger bu kattlmayt kabul ederseniz bu onaylama formunun kopyast size verilecektir kattlmayt kabul ettiginiz takdirde: Soru formunu cevaplamak olarak 25-30 dakika sOrecek Bu kattltm goni.iiiUdi.ir, ve istediginiz anda katthmdan vazgec;ebilirsiniz Bu ogrencilerin isimleri gizli tutulmaktadtr. isminiz hie; bir !?ekilde c;ah!?mada yer almayacakttr. sonunda elde edilen bilgiler aile Oyeleri, ogretmenler yada diger ogrenciler ile payla!?tlmayacakttr Eger herhangi b i r sorunuz varsa, gori.i!?me strasmda c;ekinmeden bana sorabilirsiniz. Bunun yanmda bu kattlmantz ile ilgili haklanntz hakkmda Colorado Oniversitesinden of Academic Affairs in the University of Colorado Denver, US Lawrence and 141 st. Suite 700. Telefon numarast: 303-556-2550. Bu c;ah!?ma tamamlandtgmda sonuc;lan sizinle mutluluk duyacagtmt bildirmek isterim Bu katlhmtntz iyin te!?ekki.ir ederim LUtfen sorulan dikkatle okuyarak cevaplaylntz Size uygun olan cevaplan yuvarlak ahmz gereken yerlerde boJiuklan doldurunuz. Saygtlanmla, HOiya YOksel 222

PAGE 237

Student Assent form for Semi-Structured Interviews Dear Participant: Thank you for volunteering to participate in this study about smoking for my doctoral thesis. The purpose of this study is to examine knowledge, attitudes and behaviors of young people about smoking I am interested in your opinions and experiences about smoking. The results of this study will be used to develop better education programs in the future for young people like yourself This project will not include any education about tobacco. I hope that you will answer all of the questions honestly and thoughtfully If you agree, you will receive a copy of this form. Your participation consists of the following : I anticipate the interview will last approximately 35-45 minutes The interview will be audiotaped for future transcription During this time, I will be taking notes to jot down my observations and comments. If there is anything you prefer I don't record (on tape or in writing), please let me know and this will be completely respected. I might be asking some questions of a personal nature, so I understand you may have reservations about sharing this information with me. Please remember that your participation is completely voluntary and you can decide to cease participation at any time. The benefit of your participation is in your contribution to understanding of smoking behavior among youth that could lead to successful tobacco control efforts. It is possible in exploring your experiences that negative feelings and distress could surface. This is an anonymous semi-structured interview, so I will not put your name on any of the interview pages. No information will be shared with parents, teachers, school officials, friends or others not directly involved in this research project. If you have any questions about this study, during the interview you can ask me directly will be glad to answer your questions Also, if you have questions about your rights as a participant in this research, please call the Office of Academic Affairs in the University of Colorado Denver Building (Lawrence and 141h), Suite 700. The phone number is 303-5562550. If you'd like, I'll be more than happy to share my thesis with you when it's done. Thank you for your willingness to participate. Sincerely, Hulya Yuksel 223

PAGE 238

Student Assent form for Semi-Structured Interviews (Turkish Translation) Degerli Katilimci : Doktora tezim icin yaptigim bu calismaya katiliminiz icin tesekkurler. Bu calismanin amaci lise ogrencilerinin sigara icimi ile ilgili bilgi tutum ve davranislarini olcmek. Bu catismanin amaci sizin sigara ile ilgili fikirlerinizi ve deneyimlerinizinden yararlanmaktir Butun sorulari ictenlikle cevaplayaciginizi umut ediyorum Eger bu calismaya katilmayi Kabul ederseniz bu onaylama formunun kopyasi size verilecektir Calismaya katilmayi Kabul ettiginiz takdirde: Gorusme yaklasik olarak 35-45 dakika surecek. Gorusmeyi teype alacagim. Bu arada gorusme sirasinda notlar alacagim Eger kayit etmemi istemediginiz yerler olursa kayit cihazini durduracagim. Bu calismaya katilim gonulludur ve istediginiz anda katilimdan vazgecebilirsiniz. Bu calismaya katilmaniz sigara iciminin azaltilmasi ile ilgili programlarin gelistiritmesine yardimci olacaktir Bu calismada ogrencilerin isimleri gizli tutulmaktadir. tsminiz hie bir sekilde calismada yer almiyacaktir. Gorusme sonunda elde edilen bilgiler aile uyeleri, ogretmenler yada diger ogrenciler ile paylasilmayacaktir. Eger herhangi bir sorunuz varsa, gorusme sirasinda cekinmeden bana sorabilirsiniz. Bunun yaninda bu calismaya katilmaniz ile ilgili haklariniz hakkinda Colorado Universitesinden (Office of Academic Affairs in the University of Colorado Denver, US Lawrence and 141 h St, Suite 700 Telefon numarasi: 303-556-2550 Bu calisma tamamlandiginda sonuclari sizinle paylasmaktan mutluluk duyacagimi bildirmek isterim Bu calismaya katiliminiz icin tesekkur ederim Saygilarimla, Hulya Yuksel 224

PAGE 239

UCD Human Subject<; Research Committee 1\. COVERSHEET -------------------------------HR( Prmocol u Principu/lnve&tigator: .J:!k\L'-If.l E-mail .hu\.'{0. z.Qba+rna.\\. c oro --ob-r (please answer que.$lJOns he/ow) Funding agency: ---------Dote funding propo"1l submlltcd: ---------------PrinC1p"f 11arnc and title of project 1f different from above : ------------in .wi>ntiuing this pr.oux:ol and sj.gning below, l anify that I v.ill conduct the rc:seareh imnh'ing._ hunlny ur to' [)ate I r t h-is fonn. the faculty advio;ur is : 1greeing thai has. approved rhis r.eSearcb t:ilch nf the fotto.w1ng COfli"Jrt1-0n:s \.\-ere mer: Rc.>earch rles1gn is cie<1r and ro the d1 l<' krw they were Rec.ru!lmem procedures hdr lhat )t.:bje(l_ 1!-. Voluntary is explicitly assured Informed consent procedures arc appropnate l<.l Protection \>f privacv and/or co n lidenrmlit) is odequate Potential risks >cx:ial. eccomrmrt, are Jdtntitied ,rnd Bene/irs of researrh outweigh risks fornV s talcmenl.aJld nr dt::Sf .. :ript1un.;, 1lf rese:.1r(h ihstrtlmCrtl'\ i !l t.:. ;:LWChcd Human s PrCl((h.:oJ '"' fu1!: J:ld COMMf','ITS: 3C --zotJI ------r-225

PAGE 240

Human Subject Approval of Ministry of Education in Turkey l'/!NU ..:.:<::...: 'f:.LJ: :..ur _::>1J..l. ::-=-} r i.:.. :;. : . .,;a ::; ''"' ,.,.::;-.._ 4 , .,. <;' 4'911 /fl ?Cl ;...!, ' ..J, 't .......... <. ....... J. '1.J I Sigara ile Anket. Acil ge -;Bni n -be -:; .t -.):iiJYar: ..:-!: ..:.n Sigara ile bilg. i tut.\.itJ ve .. 0 r vo? a;-a:;;s.rrma Y.?-!-ma s:..r.1 arz ,.. ........ .J._..... '-' ..... _.;_,.,, -.. ..... .) :_. !.J I \ .)"8-' '"(; 226

PAGE 241

HUlVIANSUBJECTSRESEARCH COMMTTTEE tJnivi!.rsity of Colorado at Demer Campus Box 129, P.O. Box 173364 Denver, CO DATE: TO; FROM: SLTBJECT: MEMORANDUM July6. 2001 Hulya Ynksd Dorot hy 'r'a:es. DSPA Directnr Human Research Protocol #707: Turk isn Youtb Know icdg:::. Attil\.:de5. and Behaviors About Smoking Your pro[ocol .hllsbeen approved as Non-exempt. This approval is good for up to on" year from this dare H you have any questicn s please contact me ar ( 303) 556-4060. 227

PAGE 242

HUMAN SUBJECTS RESEARCH COMMirfEE University of Colorado at Denver Campus Box 129, P.O. Box 173364 Denver, CO 80217-3364 DATE: TO: FROM: SUBJECT: MEMORANDUM March 26 2002 Hulya Yuksel Dorothy Yates DSPA Director Human Research Protocol #707: Turkish Youth's Knowledge, Attitudes, and Behaviors About Smoking Your reques t for a continuation with modifications has been approved This approval is good for up tv one year from this date If you have any questions please contact me at (303) 556-4060. 228

PAGE 243

BIBLIOGRAPHY Abernathy, T., Massad, L., Dwyer, R. (1995). The Relationship between Smoking and Self-Esteem. Adolescence, Winter 30(120), 899-908. Abercrombie, N Hill S., Turner, B. (2000). The Penguin Dictionary of Sociology. New York: Penguin Books. Aday, L. (1996). Designing and Conducting Health Surveys. San Francisco, CA: Josey-Bass Inc. Ahlgren, A., Norem, A., Hochhauser, M., Garvin, J. (1982). Antecedents of smoking among pre-adolescents. Journal of Drug Education. 12, 325-340. Aker, R.L. (1998). Social Learning and Social Structure: A General Theory of Crime and Deviance. Boston: Northeastern University Press. Alexander, C. Piazza., M., Mekos, D., Valente, T. (2001 ). Peers, schools, and adolescent cigarette smoking. Journal of Adolescent Health, 29,22-30. Alexander, H., Callcott, R., Dobson, A., Hardes, G., Lloyd, D., O'Connell, D. Leeder, S. (1983). Cigarette smoking and drug use in schoolchildren: IV factors associated with changes in smoking behavior. International Journal of Epidemiology. 12, 59-66. Alikasifoglu, M., Erginoz, E., Ercan, 0. (2002). Cigarette smoking among Turkish high school students (letter to the editor). Journal of Adolescent and Health, 30, 7-8 Allbutt, H., Amos, A., Cunningham-Burley, S. (1995). The social image of smoking among young people in Scotland. Health Education Research: Theory & Practice, 10 (4), 443-454. Ary, D., Biglan, A. (1988). Longitudinal changes in adolescent cigarette smoking behavior: onset and cessation. Journal of Behavioral Medicine. 11, 361-382. 229

PAGE 244

Asut, 0. (1993). Hekim ve Sigara. Ankara: Turk Tabibler Birligi Yayinlari. Baer, D., Katkin, J. (1981 ). Imitation of smoking by sons and daughters who smoke and smoking behavior of parents. Journal of Genetic Psychology, 118, 293-296. Bailey, S., Hubbard, R.L. (1990). Developmental variation in the context of marijuana initiation among adolescents. Journal of Health and Social Behavior, 31, 58-70. Bandura, A. (1977). Social Learning Theory. Englewood Cliffs, New Jersey: Prentice Hall, Inc., pp. vii, 194-195 Bandura, A. (1986). Social Foundation of Thought and Action: A Social Cognitive Theory. Englewood Cliffs, New Jersey: Prentice-Hall, Inc. ISBN 013-815614-x Bandura, A. (1991 ). Social Cognitive Theory of Self-Regulation. Organizational Behavior and Human Decision Processes. 50, 248-285. Bandura, A. (1997). Self-Efficacy: The Exercise of Control. New York: W.H. Freeman and Company. Bandura, A., Walters, R.H. (1963). Social Learning and Personality Development. Austin, Texas: Holt, Rinehart and Winston. Baranowski, T., Perry C.L., Parcel, G.S. (1997). How Individuals, Environments, and Health Behavior Interact: Social Cognitive Theory. In: Glanz, K., Lewis, F., Marcus & Rimer, B.K., eds. Health Behavior and Health Education: Theory. Research, and Practice. 2"d Ed. San Francisco: Jessey Bass. Barber, J., Bolitho, F., Bertrant, L.D. (1999). The predictors of adolescent smoking. Journal of Social Service Research, 26(1 ), 51-66. Baris, E., Brigden, L.W., Prindiville, J., Silva, V. L.C. (2000). Research priorities for tobacco control in developing countries: a regional approach to a global consultative process. Tobacco Control, 9, 217-223 (Summer). 230

PAGE 245

Bauman, K., Fisher, L., Bryan, E., Chenoweth, R. (1984). Antecedents, subjective expected utility, and behavior: A panel study of adolescent cigarette smoking. Addictive Behaviors. 9, 121-136. Benowitz, N.L. (1988). Drug therapy: pharmacologic aspects of cigarette smoking and nicotine. The New England Journal of Medicine, 319 (20), 1318-1331. Bernardo, R., Aldinger, C. E., Dawood, O.R., Hanson, R. E., Lee, S. J., Rinaldi, S .R. (1999) An e-mail assessment of undergraduates' attitudes toward smoking. Journal of American College Health. 48, 61-66. Biglan, A. Duncan, T., Ary D., Smolkowski, K. (1995). Peer and parental influences on adolescent tobacco use. Journal of Behavioral Medicine, 18(4), 315-330. Bilir, N., Dogan B. G., Yildiz A. N. (1997). Smoking Behavior and Attitudes in Ankara Turkey, RITC-Research for International Tobacco Control. Bilir, N., Onder, Z. (2000). Impact of the Ban on Smoking in Public Places in Turkey, 7th International Conference on System Science in Health Care, Budapest, Hungary, 29 May-2 June 2000. Brislin, R. W., Lonner, W. J., Thorndike, R. M. (1973). Cross-cultural Research Methods. New York: John Wiley. Brunswick, A., Messeri, P. (1984). Causal factors in onset of adolescents' cigarette smoking: a prospective study of urban black youth. Advances in Alcohol and Substance Abuse. 3, 35-52. Brynin, M. (1999). Smoking behavior: predisposition or adaptation? Journal of Adolescence, 22, 635-646. Camp, D., Klesges, R., Relyea, G. (1993). The relationship between body weight concerns and adolescent smoking. Health Psychology. 12(1 ):24-32. Campaign for Tobacco Free Kids. (1999). Turkey Case Study. http://tobaccofreekids.org/campaign/global/casestudies/turkey.pdf 231

PAGE 246

Carvajal, S., Wiatrek, D.E .. Evans, R., Knee, R., Nash, S. (2000). Psychosocial determinants of the onset and escalation of smoking: cross sectional and prospective findings in multiethnic middle school samples. Journal of Adolescent Health, 27, 255-265. Celik, A., Esen, A., Yorgancioglu, A., Topcu, F. (2000). Attitudes of High School Students Toward Smoking. European Respiratory Society Congress, September, 2000. Center for Disease Control and Prevention. (1992). Comparison of the cigarette brand preferences of adult and teenaged smokers-Untied States, 1989, and 10U.S. communities, 1988 and 1990. MMWR (Morb. Mortal. Wkly.Rep.), 41(1 0), 179-181. Chandler,W. (1986). Worldwatch Paper 68: Banishing Tobacco. Washington DC: Worldwatch Institute. Charlton, A., Blair, V. (1989). Predicting the onset of smoking in boys and girls. Social Science and Medicine. 29, 813-818. Chassin, L., Presson, C., Sherman, S., Corty, E., Olshavsky, R. (1984). Predicting the onset of cigarette smoking in adolescents: A longitudinal study. Journal of Applied Social Psychology. 14, 224-243. Chassin, L.., Presson, C.C., Sherman, S.J.(1990). Social psychological contributions to the understanding and prevention of adolescent cigarette smoking. Person Soc Psycho! Bull, 16, 133-51. Chassin, L .. Pression, C.C., Steven, J.S., Edwards, D.A .. (1992). The natural history of cigarette smoking and young adult social roles. Journal of Health and Social Behavior, 33(4), 328-347. Chassin, L .. Presson, C.C., Sherman, S.J., Montello, D., Mcgrew, J. (1986). Changes in peer and parent influence during adolescence: longitudinal versus cross-sectional perspectives on smoking initiation. Developmental Psychology, 22(3), 327-334. Choi, W., Pierce, J., Gilpin, E., Farkas, A., Berry, C. (1997). Which adolescent experimenters progress to established smoking in the United States. American Journal of Preventive Medicine, 13, 385-391. 232

PAGE 247

Clarke, V., White, V., Hill, D., Borland, R. (1994). School structural and policy variables associated with student smoking. Tobacco Control, 3, 339346. Clayton, S. (1991). Gender differences in psychological determinants of adolescent smoking. Journal of School Health. 61{3):115-120. Coleman, J. (1990). Foundations of Social Theory. Boston: Harvard University Press. Collins, L., Sussman, S., Rauch, J., Dent, C., Johnson, C., Hansen, W., Flay, B. (1987). Psychosocial predictors of young adolescent cigarette smoking: a sixteen-month, three-wave longitudinal study. Journal of Applied Social Psychology. 17, 554-573. Conrad K., Flay B., Hill D. (1992). Why children start smoking cigarettes: predictors of onset. British Journal of Addiction, 87, 1711-24 Corbett, K. (2001 ). Susceptibility of youth to tobacco: a social ecological framework for prevention. Respiration Physiology, 128, 103-118. Crabtree, B.F., Miller, W.L. (1999). Doing Qualitative Research. 2"d Edition, Thousand Oaks, California: Sage Publication. Crisp, A., Sedwick, P., Halek, C., Joughin, N., Humphrey, H. (1999). Why may teenage girls persist in smoking? Journal of Adolescents. 22, 657-672. Cummings et al. (1997). Comparison of recent trends in adolescent and adult cigarette smoking behaviour and brand preferences. Tobacco Control, 6{S2) 2, S31-7. Cunningham, R. (1996). Smoke & Mirrors: The Canadian Tobacco War. Ontario, Canada: International Development Research Centre. Czaja, R., Blair, J. (1996). Designing Surveys: A Guide to Decisions and Procedures. Pine Forge Press. Dagli, E. (1999). Are low-income countries targets of the tobacco industry? International Journal of Tuberculosis and Lung Disease, 3(2), 113-118. 233

PAGE 248

Dagli, Elif (2002). Philip Morris and British American Tobacco's "Charitable" Donations Around the World. Available at: http :1/www. essentia lacti on. org/toba ceo/ q ofm/0 1 08a. html Dedeoglu, N. (1994). Tobacco use among health personnel (physicians and nurses) in Antalya. Turkish Journal of Smoking and Health, 1, 7-11. De Moor, C. Cookson, K., Elder, J.P., Young, R., Melgaard, C.A., Wildey, M. (1992). The association between teacher attitudes, behavioral intentions, and smoking and the prevalence of smoking among seventh-grade students. Adolescence, 27(1 07), 565-78. Denscombe, M. (2001 ). Uncertain identities and health-risking behaviour: the case of young people and smoking in late modernity. British Journal of Sociology, 52(1), 157-177. Denscombe, M. (2001 ). Peer group pressure, young people and smoking: new developments and policy implications. Drugs: Education. Prevention and Policy, 8( 1 ), 7-32. Derman, U., Demir, G., & Akan, P. (1995). Is awareness of its risk enough to stop people from smoking? Journal of Cancer Education, 10(2), 68-70. DeVries, H., Kok, G., Dijkstra, M, 1990, Self-efficacy as a determinant of the onset of smoking and interventions to prevent smoking in adolescents. European Perspectives in Psychology, 2, 209-222. Diamond, J. (1992). The Third Chimpanzee: The Evolution and Future of The Human Animal. New York: Harper Collins. Diana, A. (2001 ). Youth at Play: Preventing Youth Problem Behavior through Sport and Recreation, International Institute for Sport and Human Performance, University of Oregon, Oregon. Distefan, J. M., Gilpin, E. A., Choi, W. S., Pierce, J. (1998). Parental influences predict adolescent smoking in the United States, 1989-1993. Journal of Adolescent Health, 22, 466-474. DuPont, R. (1994). Nicotine addiction: principles and management (book reviews). JAMA, 272 (15), 1221-1222. 234

PAGE 249

Eckert, P. (1983) Beyond the stat i stics of adolescent smoking. American Journal of Public Health, 73(4), 439-441. Elders M. J., et al. (1994). The Report of the Surgeon General: preventing tobacco use among young people. American Journal of Public Health, 84, 543-547. Emri, S. (1997). New legislation in Turkey, Tobacco Control 6(1), 61. Emri, S. (1997). Letter to the Editor: Cigarette smoking habits among primary school children in Turkey. Chest, 112(3), 863. Emri, S., Bagci, T Karakoca, Y., Baris, E. (1998). Recognition of cigarette brand names and logos by primary school children in Ankara, Turkey. Tobacco Control, 7, 386-392. Erbaydar, T., Dagli, E., Hayran, 0., Ayla, F., Lawrence, S., Co lli shaw, Ne il E., Gallard, C. (2002) Smoking Patterns of Youth in Turkey, Final Report Marmara University, Physicians for a Smoke-Free Canada. RITC-Research for International Tobacco Control, Project No : 001726-022 Evans, R.I., Dratt, L.M., Raines, B.E., Rosenberg, S. (1988). Social influences on smoking initiation: importance of distinguishing descriptive vs. mediating process variables. Journal of Applied Social Psychology, 18, 925-43 Farkas, A. J., Distefan, J. M., Choi, W. S., Gilpin, E. A., Pierce, J.P. (1999). Does parental smoking cessation discourage adolescent smoking? Preventive Medicine, 28, 213-218. Fearnow, M Chassin, L., Presson, C., Sherman, S. (1998) Determinants of parental attempts to deter their children's cigarette smoking. Journal of Applied Developmental Psychology, 19(3), 453-468. Financial Times (2003). Formula 1 and Turkey http://search.ft.com/search/article.html?id=030829000090, 29 August. Fiore, M., Novotny, T., Piece, J., Hatziandreu, E., Patel, K., Davis, R. (1998). Trends in cigarette smoking in the United States. The changing influence of gender and race. Journal of the American Medical Associate. 261(1 ):49-55. 235

PAGE 250

Firat, D (1996). Tobacco and cancer in Turkey. Journal of Environmental Pathology. Toxicology and Oncology, 15(2-4), 155-160. Fisher, J. (1976). Sex differences in smoking dynamics Journal of Health and Social Behavior. 17(2), 155-162. Fisher, L.A. Bauman, K.E., (1988) Influence and selection in the friend adolescent relationship: Findings from studies of adolescent smoking and drinking. Journal of Applied Social Psychology 18, 289-314. Flay, B. R. (1993) Youth Tobacco Use: Risks, Patterns, and Control. Nicotine Addiction: Principles and Managements. C.T., Orleans & Slade, J. (eds .), Oxford, England : Oxford University Press. Flay, B.R., d'Avernas, J.R., Best, J.A., Kersell, M.W., Ryan, K.B. (1983). Cigarette smoking: Why young people do it and ways of preventing it. In McGrath, P.J., Firestone, P. (Eds.), Pediatric and Adolescent Behavioral Medicine (pp. 132-183). New York: Springer. Flay, B.R., Hu, F.N., Siddiqu i 0., Day, L.E., Hedeker, D Petraitis, J., Richardson, J., & Sussman, S. (1994). Differential influence of parental smoking and friends' smoking on adolescent initiation and escalation and smoking. Journal of Health and Social Behavior, 35(3), 248-265. Flay, B. R., Petraitis, J. (1991 ). Methodological i ssues in drug use prevention research: Theoretical foundations. National Institute on Drug Abuse research monographs. 107:81-109 Rockville, MD. Flay, B.R., Petraitis, J. (1994). The theory of triadic influence: A new theory of health behavior with implications for preventive interventions. Advances in Medical Sociology, 4, 19-44. Flay, B.R., Petraitis, J., Hu, F .B. {1997). Psychosocial Risk Factors For Adolescent Tobacco Use. Robert Wood Johnson Foundation Conference: New Partnerships and Paradigms for Tobacco Prevention Research Sundance Utah. Flay, B., Petraitis J., Hu, F.B (1999) Psychological risk and protective factors for adolescent tobacco use. Nicotine & Tobacco Research, 1, S59S65 236

PAGE 251

Flay, B.R., Phil, D. (1985). What we know about the social influences approach to smoking prevention: Review and recommendations. NIDA Research Monograph, 63, 67-112. Flay, B.R., Phil, D., Hu, F., Richardson, J. (1998). Psychosocial predictors of different stages of cigarette smoking among high school students. Preventive Medicine. 27, A9-A 18. Glanz, K., Lewis F.M., Rimer B.K. (1997). Health Behavior and Health Education: Theory, Research and Practice. San Francisco: Jossey-Bass Publishers. Governor's Office Report (2001) Dursunbey. Turkiye Cumhuriyeti Dursunbey Kayrnakamligi. Gray, D., Amos, A., Currie, C. (1996). Exploring young people's perceptions of smoking images in youth magazines. Health Education and Research: Theory and Practice. 11(2):215-230. Green, L.W., Richard, L., Potvin, L. (1996}. Ecological foundations of health promotion. American Journal of Health Promotion, 10(4), 270-281. Griesbach, D., lnchley, J., Currie, C. (2002) More than words? The status and impact of smoking policies in Scottish schools. Health Promotion International, 17( 1 ), 31-41 Gritz, E., Crane, L. (1991 ). Use of diet pills and amphetamines to lose weight among smoking and nonsmoking high school seniors. Health Psychology. 1 0(5}:330-335. Guneri, O.Y, Hatipoglu, Z.S, Yildirim, A. (1999}. Sources of self-identity among Turkish adolescents Adolescence, 34(135), 535-546. Guarnaccia, P.J. (1996). Anthropological perspectives: The importance of culture in the assessment of quality of life. In: Quality of Life and Pharmacoeconomics in clinical trials (2"d edition). Spilker, B., ed. Philadelphia, PA: Lippincott-Raven Publishers. Harrell, J., Faan, R., Bangdiwala, S., Deng, S., Webb, J., Bradley, C. (1998). Smoking initiation in youth: The roles of gender, race, socioeconomics, and development status. Journal of Adolescent Health, 23, 271-279. 237

PAGE 252

Hammond, R, Purcell M. Country Case Studies: Turkey. Tobacco Free Project of the San Francisco Department of Public Health. June 1999. Available at http://tobaccofreekids.org/campaign/global/casestudies/turkey.pdf. Headden, S. (1998). The Marlboro man lives! Restrained at home, tobacco firms step up their marketing overseas. U S. News & World Report, 125(11 ), 58-59. Henriksen, L., Jakson, C. (1998). Anti-smoking socialization: Relationship to parent and child smoking status. Health Communication, 1 0( 1 ), 87-1 001. Heishman, S.J., Kozlowski, L.T., Henningfield, J.E. (1997). Nicotine addiction: implications for public health policy. Journal of Social Issues, 53 (1 ), 13-21. Hirschi, T. (1969). Causes of Delinquency, Berkeley: University of California Press. Huba, G.J., Wingard, J.A., Bentler, P .M. (1980b) Longitudinal analysis of the role of peer support, adult models, and peer subcultures in beginning adolescent substance use: An application of stepwise canonical correlation method. Multivariate Behavioral Research, 15, 259-279. Hwang, S. L. (1998). Drawn to Marlboro's Buzz, Taste, The Turkish Shun the Local Leaf, The Wall Street Journal, September 11 Jackson, C., Henriksen, L. (1997). Do as I say: Parent smoking, antismoking socialization, and smoking onset among children. Addictive Behaviors, 22(1), 107-114. Jarallah, J., AI-Rubeaan, K., AI-Nuaim, A., AI-Ruhaily, A., Kalantan, K. (1999). Prevalence and determinants of smoking in three regions of Saudi Arabia. Tobacco Control, 8, 53-56. Jenks, R.J. (1994). Smoking and satisfaction and motivations: a comparison of men and women. The Journal of Social Psychology, 134(6), 847-850. Jessor, R., Donovan J.E., Costa, F.M. (1991 ). Beyond adolescence: Problem behavior and young adult development. Cambridge, England: Cambridge University Press. 238

PAGE 253

Jones, C. (2000). TobaccoUp in Smoke: Philip Morris uses well-oiled plan to slice into Turkey market. http://www.qatewavva.com/rtd/special/ tobacco/chip0921.shtml Kandel, D.B. (1978). Longitudinal Research on Drug Use: Empirical Findings and Methodological Issues, pp. 73-99. Washington, DC: Hemisphere. Kandel, D. B. (1985) On the process of peer influence in adolescent drug use: A developmental perspective. In Advances in Alcohol and Substance Abuse. (pp. 139-163). New York: Haworth Press. Kandel, D.B. (1980). Drug and drinking behavior among youth. Annual Review of Sociology, 6, 235-85. Kandel, D.B., Kessler, R.C., Margulies, R.Z. (1978). Antecedents of adolescent initiation into stages of drug use: A developmental analysis. In Kaplan, H.B., Martin, S.S., Robbins, C. (1984). Pathways to adolescent drug use: Self-derogation, peer influence, weakening of social controls, and early substance use. Journal of Health and Social Behavior. 25, 270-289. Karababa, A. 0., Bati, H., Kocabiyik, N. (1997). lzmir-Aitindag ciraklik egitim merkezindeki ciraklarin sigara icmelerinde ailerinin ve is ortaminin etkisi. Ege Universitesi Tip Fakultesi Halk Sagligi Anabilim Dali, Bornova, lzmir. Karanci, N. A., Rustemli, A. (1987). Smoking typology in a Turkish sample. The International Journal of the Addictions, 22(3), 289-299. Kellam, S., Ensminger, M. (1980). Mental health in first grade and teenage drug, alcohol, and cigarette use. Drug and Alcohol Dependence. 5, 273-304. Kessler, D. A. (1995). Nicotine addiction in young people, The New England of Journal of Medicine, 333(3), 186-190. Klebnikov, P. (1994). Opiate of the masses. Forbes, 153(8), 74. Kocabas, A., Burgut, R., Bozdemir, N., Akkoclu, A., Cildag, 0., Dagli, E., Erkan, L., lsik, R., Turktas, H. (1994). Smoking patterns at different medical schools in Turkey. Tobacco Control, 3, 228-235. 239

PAGE 254

Kottke T.E., Battista R.N. et al. (1998). Attributes of successful smoking cessation interventions in medical practice. A meta-analysis of 39 controlled trials. JAMA, 259(19), 2883-2889. Laurier, E., McKie, L., Goodwin, N. (2000). Daily and lifecourse contexts of smoking. Sociology of Health & Illness, 22(3), 289-309. Lee, K. (1997). A study on male high school students' smoking patterns. Taehan Kanho, 36(5), 91-100. Leo, J. (1997). Boyz to (Marlboro) men. U.S. News & World Report, June 2, 1997 Lewis, F. (1997). Models of Interpersonal Health Behavior. In: Glanz, K., Lewis, F., Marcus, Rimer, B.K., eds. Health Behavior and Health Education: Theory. Research. and Practice. 2"d Ed. San Francisco: Jossey-Bass. Lloyd-Richardson, 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 and Clinical Psychology, 70(4), 998-1009. Lucas, K., Lloyd, B. (1999). Starting smoking: girls' explanations of the influence of peers. Journal of Adolescence. 22, 647-655. Mackay, J., & Crofton, J. (1996). Tobacco and the developing world. British Medical Bulletin, 52, 206-221. Makary, M. (1998). The international tobacco strategy. Journal of the American Medical Association, 280(13), 1194. Malcon, M., Menezes, A., Maia, F., Chatkin, M., Victora, C. (2003). Prevalence of and risk factors for cigarette smoking among adolescents in South America: a systematic literature review. Rev Panam Salud Publica. 13( 4 ):222-228. Marshall, G. (1998}. Oxford Dictionary of Sociology. New York: Oxford University Press. 240

PAGE 255

Marshall, M. (2000). Social and cultural patterns of tobacco use: an international and historical perspective. Journal of Substance Abuse, 5, S155-158. Maziak, W. (2002) Smoking in Syria: profile of a developing Arab country. The International Journal of Tuberculosis and Lung Disease 6(3), 183-191. Maziak, W., Mzayek, F. (2000a). The dynamics of tobacco smoking among male educated youths in Aleppo-Syria. European Journal of Epidemiology, 16, 769-772. Maziak, W., Mzayek, F (2000b). Characterization of the smoking habit among high school students in Syria. European Journal of Epidemiology, 16(12), 1169-1176. McAlister, A. (1980). Pilot study of smoking, alcohol, and drug abuse preventing. American Journal of Public Health, 70, 719-721. McGee, R., Stanton, W. R. (1993). A longitudinal study of reasons for smoking in adolescence. Addiction, 88, 265-271. McNeill, A., Jarvis, M. Stapleton, J., Russell, M., Eiser, J., Gammage, P., Gray, E. (1988). Prospective study of factors predicting uptake of smoking in adolescents Journal of Epidemiology and Community Health, 43, 72-78. Mechanic, D. (1999). Issues in promoting health. Social Science and Medicine, 48, 711-718. Miles, S. (1995). Towards an understanding of the relationship between youth identities and consumer culture. Youth and Policy, 50(1 ), 35-45. Miles, S., Cliff, D., Burr V. (1998). 'Fitting in and sticking out': consumption, consumer meanings and the construction of young people's identities. Journal of Youth Studies, 1(1 ), 81-96 Miller, N., Dollard, J. (1941). Social Learning and Imitation. New Haven, CT: Yale University Press. Mitchell, L. (1997). Pressure Groups: young people's accounts of peer pressure to smoke. Social Sciences in Health, 3(1) 3-17. 241

PAGE 256

Mitchell, L., Amos, A. (1997). Girls, pecking order and smoking. Health Education and Research: Theory and Practice. 11(2), 1861-1869. Mittelmark, M., Murry, D., Luepker, R., Pechacek, T., Pirie, P., Pallonen, U. (1987). Predicting experimentation with cigarettes: the childhood antecedents of smoking study (CASS) American Journal of Public Health. 72, 206-208. Moore, L., Roberts, C., Tudor-Smith, C. (2001 ). School smoking policies and smoking prevalence among adolescents: multilevel analysis of cross sectional data from Wales. Tobacco Control, 10(2), 117-123. Moos, R.H. (1976). The Human Context: Environmental Determinants of Behavior. New York: Wiley. Morse, J.M. (1992). Qualitative Health Research. Newbury Park: Sage Publications. Motulsky, Harvey (1995). Intuitive Biostatistics. New York: Oxford University Press. Murray, C., Lopez, A. D. (1996). The Global Burden of Disease. Harvard School of Public Health. Boston, MA. Murray, M., Swan, A., Johnson, M., Bewley, B. (1983). Some factors associated with increased risk of smoking by children. British Journal of Psychiatry. 24, 223-232. Nagelkerke, N.J. D. (1991 ). A note on a genderal definition of the coefficient of determination. Biometrica, 78, 691-692. National Institute on Drug Abuse Research Report Series (1999). Nicotine Addiction Http :1/www .nida .nih .gov/researchreports/ni co tine Newcomb, M., McCarthy, W., Bentler, P. (1989). Cigarette smoking, academic lifestyle, and social impact efficacy: an eight-year study from early adolescence to young adulthood. Journal of Applied Social Psychology. 19, 251-281. Nichter, M., Cartwright, E. (1991 ). Saving the Children for the Tobacco Industry. Medical Anthropology Quarterly, 5, 422-433. 242

PAGE 257

Nichter, M., Nichter, M., Thompson, P.J., Shiffman, S., Moscicki, A. (2002). Using qualitative research to inform survey development on nicotine dependence among adolescents. Drug and Alcohol Dependence. 68(1 ):4156. Nichter, M., Nichter, M., Vuckovic, N., Quintero, G., & Ritenbaugh, C. (1997). Smoking experimentation and initiation among adolescent girls: qualitative and quantitative findings. Tobacco Control, 6, 285-295. Norman, N.M., Tedeschi, J. T. (1989). Self-presentation, reasoned action, and adolescents decisions to smoke cigarettes. Journal of Applied Social Psychology, 19(7), 543-558. Nutbeam D., Smith C., Moore L., Bauman, A. (1993). Warning! Schools can damage your heath: alienation from school and its impact on health behaviour. Journal of Pediatric Child Health 29(supplement 1 ):25-30 O'Connell, D., Alexander, D., Dobson, H., Lloyd, A., Hardes, G., Springthorpe, H., Leeder, S (1981 ). Cigarette smoking and drug use in schoolchildren: Eleven factors associated with smoking. International Journal of Epidemiology, 10, 223-231. Ozcan, Y. Z., Ozcan, M. 0. (2002). Determinants of youth smoking-evidence from Turkey. Substance Use & Misuse. 37(3), 313-336. Pandina, R.J. Schuele, J.A. (1983). Psychosocial correlates of alcohol and drug use o adolescent students and adolescents in treatment. Journal of Studies on Alcohol, 15(4), 519-526. Petraitis, J., Flay, B., Miller, T. (1995). Reviewing theories of adolescent substance use: organizing pieces in the puzzle. Psychological Bulletin. 117( 1 ), 167-86 Pool, C. (1998) Firing Up Latin America Latin Trade, 6(10), 46-50 Poulsen, L.H., Osler, M., Roberts, C., Due, P., Damsgaard, M.T., Holstein, B.E. (2002). Exposure to teacher smoking and adolescent smoking behaviour: analysis of cross sectional data from Denmark. Tobacco Control 11, 246-251. 243

PAGE 258

Prochaska, J.O., DiClemente, C.C. (1983). "Stages and process of self change of smoking toward an integrative model of change. Journal of Consulting and Clinical Psychology, 51, 390-395. Report of the Surgeon General. ( 1994 ). Preventing Tobacco Use among Young People. Atlanta: Centers for Disease Control and Prevention, Public Health Service, Department of Human Service. Rissel, C., Mclellan, L., Bauman, A. (2000). Factors associated with delayed tobacco uptake among Vietnamese/Asian and Arabic youth in Sydney, NSW. Australian New Zealand Journal of Public Health. 24(1), 2228. Ritchie, L.D. (1991 ). Family communication patterns: an epistemic analysis and conceptual reinterpretation. Communication Research, 18, 548-565. Rose, J.S., Chassin, L., Presson, C., Sherman, S. (1999). Peer influences on adolescent cigarette smoking: A prospective sibling analysis. Meriii Palmer Quarterly, 45-1, 62-71. Rowe, D.C., Plomin, R. (1981 ). The importance of nonshared environmental (E1) influences in behavioral development. Developmental Psychology, 17, 517-531. Rowe, D.C., Woulbroun, E.J., Gulley, B.L. (1994). Peers and friends as nonshared environmental influences. In Hetherington, E.M., Reiss, D., Plomin, R. (Eds.), Separate Social Worlds of Siblings: The impact of Nonshared Environment on Development, Hillsdale, NJ: Erlbaum. Sarason, G., Mankowski, E. S., Peterson, A.V., Dinh, K. T. (1992). Adolescents' reasons for smoking. Journal of School Health, 62(5), 185191. Sargent, J., Beach, M., Dalton, M., Mott, L., Tickle, J., Ahrens, M., Heatherton, T. (2001 ). Effect of seeing tobacco use in films on trying smoking among adolescents: cross sectional study. British Medical Journal. 323, 1-6. 244

PAGE 259

Satman, 1., Yilmaz, T., Sengul, A., Salman, S., Salma, F., Uygur, S., Bastar, 1., Tutuncu, Y., Sargin, M., Dinccag, N., Karsidag, K., Kalaca, S., Ozcan, C. & King, H. (2002). Population-based study of diabetes and risk characteristics in Turkey: results of the Turkish Diabetes Epidemiology Study (TURDEP). Diabetes Care, 25(9), 1551-1556. Semmer, N., Dwyer, J., Lippert, P., Fuchs, R., Cleary, P., Schindler, A. (1987). Adolescent smoking from a functional perspective: the Berlin Bremen Study. European Journal of Psychology of Education. 11, 387-402. Shadel, W.G., Shiffman, S., Niaura, R., Nichter, M., Abrams, D.B. (2000). Current models of nicotine dependence: what is known and what is needed to advance understanding of tobacco etiology among youth. Drug and Alcohol Dependence, 59 Suppl. 1, S9-S21. Simmel, G. (1971 ). On Individuality and Social Forms Chicago: University of Chicago Press. Skinner, W., Massey, J., Krohn, M., Lauer, R. (1985). Social influences and constraints on the initiation and cessation of adolescent tobacco use. Journal of Behavior Medicine, 8(4), 353-376. Smet, B., Maes, L., DeClercq, L., Haryanti, K., Winarno, R. (1999). Determinants of smoking behavior among adolescents in Semarang, Indonesia. Tobacco Control. 8, 186-191. Smith, K.H. & Stutts, M.A. (1999). Factors that influence adolescents to smoke. The Journal of Consumer Affairs, 33(2), 321-357. Spear, S. F., Akers, R. L. (1988). Social learning variables and the risk of habitual smoking among adolescents: The Muscatine Study. American Journal of Preventive Medicine, 4(6), 336-342. Squier C.A. & Dreher, M.C. (2000). Editorial: Social and cultural patterns of tobacco use: an international and historic perspective. Journal of Substance Use, 5, 153-154. Stanton, W., Oeio, T., Silva, P. (1994). Sociodemographic characteristics of adolescent smokers. The International Journal of the Addictions, 29(7), 913925. 245

PAGE 260

Stebbins, K. R. (1990). Tobacco, politics and economics: implications for global health. Social Science Medicine, 33(12), 1317-1326. Stern R.A., Prochaska, J.O., VelicerW.F., & Elder J.P. (1987). Stages of adolescent cigarette smoking acquisition: measurement and sample profiles. Addiction Behavior, 12, 319-29 Stone, Danice, 1998, Social Cognitive Theory, University of South Florida web page: http://hsc.usf.edu/-kmbrown/Social_ Cognitive_ Theory_ Overview .htm Strecher, V.J., Rosenstock I,M. (1997). Health belief model. In Glanz, K., Lewis, FM., and Rimer, BK., (eds.), Health Behavior and Health Education: Theory, Research. and Practice. 2"d Ed. {pp.41-59), San Francisco: Jessey Bass. Sussman, M.S ., Dent, C., Flay, B., Hansen, W., Johnson, C. (1987). Psychosocial predictors of cigarette onset by White, Hispanic, and Asian adolescents in Southern California. Morbidity and Mortality Weekly Report. 36, 11-16. Tabachnick, B ., Fidell, L. (2001 ). Using Multivariate Statistics Needham, MA: Allyn & Bacon. Taha, A., Ball, K. (1985). Tobacco and the Third World: the growing threat. East African Medical Journal, 62(1 0), 735-7 41. Tansel, A. (1993). Cigarette demand, health scares and education in Turkey. Applied Economics, 25, 521-529. Thorton, W., Douglas, G. A., Houghton, S. J. (1999). Transition through states of smoking: the effect of gender and self-concept on adolescent smoking behavior. Journal of Adolescent Health, 25, 284-289. Tilly, C. (1990). The World is Their Ashtray. Dollars & Sense, 160, 12-15. Tims, A. R., Masland, J.L. (1985). Measurement of family communication patterns. Communication Research, 12, 35-37. Tobacco or Health. (2000). A Global Status ReportCountry Profiles by Region: http://www.cdc.gov/tobacco/who/turkey.htm 246

PAGE 261

Tobacco Reporter. Tekel privatization moves ahead slowly News Link: http://www. tobaccoreporter .com/currenUstory2 .asp Trevathan, W. R., Smith, E. 0., McKenna, J. J. (1999). Evolutionary Medicine. Oxford, NY: Oxford University Press. U.S. Department of Health and Human Services, Preventing Tobacco Use Among Young People A Report of the Surgeon General. Atlanta, Georgia: U.S. Department of Health and Human Services, Public Health Service, Centers for Disease Control and Prevention, National Center for Chronic Disease Prevention and Health Promotion, Office on Smoking and Health, 1994. Warren, C.W., Riley, L., Asma, S., Eriksen, M.P., Green, L., Blanton, C., Loa, C., Batchelor, S., Yach, D (2000). Tobacco use by youth: a surveillance report from the Global Youth Tobacco Survey project. Bulletin of World Health Organization, 78(7), 868-874. Weissman, R., White, A. (2002). Needles Harm: International Monetary Fund Support for Tobacco Privatization and for Tobacco Tax and Tariff Reduction. and the Cost to Public Health, An Essential Action Report, www .essentialaction .org/tobacco West, P., Sweeting, H., Ecob, R., (1999). Family and friends' influences on the uptake of regular smoking from mid-adolescence to early adulthood. Addiction, 94(9), 1397-1412. World Bank. (1996). Curbing the Epidemic: Governments and the Economics of Tobacco Control. Washington: World Bank. World Bank. (2000). Curbing the Epidemic: Governments and the Economics of Tobacco Control. Washington: World Bank. The World Factbook (2002). Turkey. Available at: http://www. cia .gov/cia/publicati ons/factbooklgeos/tu. html World Health Organization. (1997). Tobacco or Health: Global Status Report. Geneva: World Health Organization. World Health Organization. (1998) Guidelines for Controlling and Monitoring the Tobacco Epidemic. Geneva, p.76 247

PAGE 262

World Health Organization. (1999). World Health Report Washington, D.C: World Health Organization Press. World Health Organization, American Cancer Society, International Union Against Cancer (2003). Tobacco Control Country Profiles, 2"d Edition. 12th World Conference on Tobacco Or Health, Helsinki, Finland. Available at: http://www globalink org/tccpllntroduction.pdf Yach, D (1986). The impact of smoking in developing countries witb special reference to Africa. International Journal of Health Services, 16(2), 279-292. Zhang, L., Wang, W., Zhao, Q., Vartiainen, E. (2000) Psychosocial predictors of smoking among secondary school students in Henan, China. Health Education Research, 15(4 ), 415-422 Zimmerman R. S, Vernberg, D. (1994). Models of Preventive Health behavior: comparison, critique and meta-analysis. Advances in Medical Sociology 4, 45-67. 248