Citation
Social factors of cigarette smoking initiation among undergraduate college students

Material Information

Title:
Social factors of cigarette smoking initiation among undergraduate college students
Creator:
Emmeree, Jane F. ( Author, Primary )
Publication Date:
Copyright Date:
2003

Subjects

Subjects / Keywords:
Cigarette smoking ( jstor )
Cigarettes ( jstor )
College students ( jstor )
Colleges ( jstor )
High school students ( jstor )
School surveys ( jstor )
Sociability ( jstor )
Student surveys ( jstor )
Undergraduate students ( jstor )
Universities ( jstor )
City of Gainesville ( local )

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
Copyright Emmeree, Jane F.. Permission granted to the University of Florida to digitize, archive and distribute this item for non-profit research and educational purposes. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder.
Embargo Date:
9/9/1999
Resource Identifier:
53207934 ( OCLC )

Downloads

This item has the following downloads:

PDF ( 11 MBs ) ( .pdf )

emmeree_j_Page_157.txt

emmeree_j_Page_013.txt

emmeree_j_Page_037.txt

emmeree_j_Page_009.txt

emmeree_j_Page_040.txt

emmeree_j_Page_155.txt

emmeree_j_Page_123.txt

emmeree_j_Page_055.txt

emmeree_j_Page_093.txt

emmeree_j_Page_102.txt

emmeree_j_Page_107.txt

emmeree_j_Page_105.txt

emmeree_j_Page_068.txt

emmeree_j_Page_057.txt

emmeree_j_Page_133.txt

emmeree_j_Page_011.txt

emmeree_j_Page_080.txt

emmeree_j_Page_150.txt

emmeree_j_Page_053.txt

emmeree_j_Page_142.txt

emmeree_j_Page_160.txt

emmeree_j_Page_021.txt

emmeree_j_Page_016.txt

emmeree_j_Page_086.txt

emmeree_j_Page_110.txt

emmeree_j_Page_076.txt

emmeree_j_Page_128.txt

emmeree_j_Page_144.txt

emmeree_j_Page_029.txt

emmeree_j_Page_130.txt

emmeree_j_Page_161.txt

emmeree_j_Page_023.txt

emmeree_j_Page_041.txt

emmeree_j_Page_052.txt

emmeree_j_Page_095.txt

emmeree_j_Page_035.txt

emmeree_j_Page_082.txt

emmeree_j_Page_104.txt

emmeree_j_Page_075.txt

emmeree_j_Page_066.txt

emmeree_j_Page_090.txt

emmeree_j_Page_032.txt

emmeree_j_Page_097.txt

emmeree_j_Page_015.txt

emmeree_j_Page_138.txt

emmeree_j_Page_088.txt

emmeree_j_Page_106.txt

emmeree_j_Page_005.txt

emmeree_j_Page_085.txt

emmeree_j_Page_050.txt

emmeree_j_Page_125.txt

emmeree_j_Page_058.txt

emmeree_j_Page_061.txt

emmeree_j_Page_081.txt

emmeree_j_Page_036.txt

emmeree_j_Page_132.txt

emmeree_j_Page_111.txt

emmeree_j_Page_012.txt

emmeree_j_Page_002.txt

emmeree_j_Page_054.txt

emmeree_j_Page_137.txt

emmeree_j_Page_092.txt

emmeree_j_Page_120.txt

emmeree_j_Page_140.txt

emmeree_j_Page_030.txt

emmeree_j_Page_115.txt

emmeree_j_Page_164.txt

emmeree_j_Page_087.txt

emmeree_j_Page_122.txt

emmeree_j_Page_039.txt

emmeree_j_pdf.txt

emmeree_j_Page_108.txt

emmeree_j_Page_153.txt

emmeree_j_Page_027.txt

emmeree_j_Page_091.txt

emmeree_j_Page_014.txt

emmeree_j_Page_117.txt

emmeree_j_Page_065.txt

emmeree_j_Page_156.txt

emmeree_j_Page_070.txt

emmeree_j_Page_147.txt

emmeree_j_Page_069.txt

emmeree_j_Page_149.txt

emmeree_j_Page_073.txt

emmeree_j_Page_109.txt

emmeree_j_Page_022.txt

emmeree_j_Page_019.txt

emmeree_j_Page_113.txt

emmeree_j_Page_136.txt

emmeree_j_Page_067.txt

emmeree_j_Page_162.txt

emmeree_j_Page_008.txt

emmeree_j_Page_158.txt

emmeree_j_Page_124.txt

emmeree_j_Page_165.txt

emmeree_j_Page_003.txt

emmeree_j_Page_074.txt

emmeree_j_Page_046.txt

emmeree_j_Page_154.txt

emmeree_j_Page_024.txt

emmeree_j_Page_001.txt

emmeree_j_Page_004.txt

emmeree_j_Page_064.txt

emmeree_j_Page_072.txt

emmeree_j_Page_031.txt

emmeree_j_Page_148.txt

emmeree_j_Page_044.txt

emmeree_j_Page_103.txt

emmeree_j_Page_006.txt

emmeree_j_Page_025.txt

emmeree_j_Page_079.txt

emmeree_j_Page_026.txt

emmeree_j_Page_051.txt

emmeree_j_Page_089.txt

emmeree_j_Page_047.txt

emmeree_j_Page_143.txt

emmeree_j_Page_121.txt

emmeree_j_Page_033.txt

emmeree_j_Page_034.txt

emmeree_j_Page_127.txt

emmeree_j_Page_099.txt

emmeree_j_Page_112.txt

emmeree_j_Page_119.txt

emmeree_j_Page_139.txt

emmeree_j_Page_141.txt

emmeree_j_Page_007.txt

emmeree_j_Page_145.txt

emmeree_j_Page_096.txt

emmeree_j_Page_018.txt

emmeree_j_Page_100.txt

emmeree_j_Page_098.txt

emmeree_j_Page_042.txt

emmeree_j_Page_151.txt

emmeree_j_Page_134.txt

emmeree_j_Page_059.txt

emmeree_j_Page_083.txt

emmeree_j_Page_071.txt

emmeree_j_Page_101.txt

emmeree_j_Page_038.txt

emmeree_j_Page_048.txt

emmeree_j_Page_116.txt

emmeree_j_Page_118.txt

emmeree_j_Page_017.txt

emmeree_j_Page_045.txt

emmeree_j_Page_077.txt

emmeree_j_Page_094.txt

emmeree_j_Page_063.txt

emmeree_j_Page_062.txt

emmeree_j_Page_114.txt

emmeree_j_Page_135.txt

emmeree_j_Page_163.txt

emmeree_j_Page_084.txt

emmeree_j_Page_049.txt

emmeree_j_Page_043.txt

emmeree_j_Page_159.txt

emmeree_j_Page_010.txt

emmeree_j_Page_078.txt

emmeree_j_Page_028.txt

emmeree_j_Page_152.txt

emmeree_j_Page_126.txt

emmeree_j_Page_060.txt

emmeree_j_Page_020.txt

emmeree_j_Page_146.txt

emmeree_j_Page_129.txt

emmeree_j_Page_131.txt

emmeree_j_Page_056.txt


Full Text












SOCIAL FACTORS OF CIGARETTE SMOKING INITIATION AMONG
UNDERGRADUATE COLLEGE STUDENTS

















By

JANE F. EMMEREE


A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY

UNIVERSITY OF FLORIDA


2003


































Copyright 2003

by

Jane F. Emmeree
































In Dedication to



Jill Margaret Fameree

and

JoAnn McFarlin















ACKNOWLEDGMENTS

My first name, Jane, means "God is gracious." During my entire life, but

particularly during these past four years, I have been graced with the love and support of

so many outstanding people.

Over the past four years, I have had wonderful support from professors and

graduate students on campus. Dr. Jill Varnes is the best chairperson on the planet. I am

grateful for her competence, caring, wisdom, and flexibility. I also thank Dr. R. Morgan

Pigg for his calm and gentle spirit. I appreciate Dr. Bill Chen and Dr. Sadie Sanders for

their expertise in the area of smoking and your encouraging words. I would also like to

thank Dr. Randy Penfield for his patience and expertise with my survey development and

data analyses. I learned so many new things in his classes and during the dissertation

process.

I would also like to thank two professors who did not sit on my committee but had

a significance influence on me. They were especially kind, enthusiastic, and supportive:

Dr. Linda Behar-Horenstein and Dr. Kim Walsh-Childers. I am also indebted to my

office mate, Terri Mitchell, and other fellow graduate students who have become dear

friends: Teresa Lyles, Laura Hill, and Jamie Fisher. They are great friends.

I am also graced with loving and positive individuals in my personal life. Without

my mother JoAnn's help, I would not be where I am today. I thank her for taking care of

me during the most difficult times and celebrating with me during the most joyous ones.

My sister, Jill, is my best friend. I appreciate her prayers, optimism, and unfailing









encouragement, not only during the past four years but also for as long as I can

remember! My brother, Jody; his wife, Julie; and my niblings (a made-up word for

"niece and nephews)-Jenna, Jackson, and Jordan-were a great distraction. I thank all

of them for the camping trips, baseball and basketball games, rock-climbing, and

kayaking. I love hanging out with them. I thank my father, Gale, and my grandmother,

Minerva, for their prayers and wise advice.

My non-biological family also gave me great support and help. I thank John

Barbour, Stacey Hoffman, and Joy Asbell for their on-going friendship, for bringing me

meals-on-wheels, and for letting me borrow their children for kid-therapy. Their children

are a special blessing and helped me keep life in perspective. I would forget any

academic worries while searching for the moon during the daylight, walking dogs and

bathing cats, throwing stones in the pond, playing on the "patio," and answering

questions like "Aunt Jane, can I come and live with you for six and a half days?" It is

my hope that one day Kaitee, Luke, John William, and Joseph will read this and realize

how much joy they all gave me during a very challenging time in my life. I am also

indebted to my family at Faith Presbyterian Church for their love for God, on-going

prayers and petitions, sound counsel and teaching, and ministering to me at a crucial time

in my life.

Finally, and most significantly, I thank the Lord, for blessing me with His grace

and with these remarkable people.
















TABLE OF CONTENTS
page

A C K N O W L E D G M E N T S ....................................................................... .....................iv

LIST OF TABLES .................................. .. ... .... ................. .x

LIST OF FIGURES ................................................................. .... ..... xii

AB STRA CT....................... ... ............................................ ....... ..... xiii

CHAPTER

1 IN TR O D U CTION ............................................... .. ......................... ..

R research Problem ....................... .... .............. .............................. .2
P purpose of the Study ........................................ .................. .............. ..............
Rationale for the Study ..................................... .......... ...... ............... 2
The Negative Health Consequences of Smoking ...............................................3
Smoking Reduction as a Current Health Priority.......................................4
Smoking Cessation as a Difficult Goal ...................................... ............... 4
Lim ited R research on College Students............................................................... 5
Sm oking Initiation A m ong College Students ........................................ ....................5
R research Questions .................. .................................... .. ................ .8
D elim stations ............................................................... .... ..... ......... 9
L im itatio n s ........................................................... ................ .. 9
A ssu m option s .................................................. .......................... 9
D definition of Term s .......................... ......... .. .. ...... .. ............10
S u m m a ry ....................................................................... 1 1

2 REVIEW OF THE LITERATURE .................................... .......................... ........ 13

Cigarette Smoking Among College Students .................................. ............... 13
Smoking Prevalence Rates Among College Students.......................................13
Frequency and Quantity Sm oked ............................................. ............... 14
Sm oking R ates by G ender............................................................................. .... 15
Sm king R ates by R ace ............................................................ .............. 15
Smoking Trends Among College Students ............................... ............... .16
Smoking Initiation Among College Students .................................. ............... 17
Reasons Why College Students Smoke ....................................... ...............18
A Conceptual Framework for Smoking Initiation..........................................20
Incorporating Susceptibility into the Stages.................. .. .............. ....................21









R isk Factors of Sm king by Stage ........................................... .....................25
S u m m a ry ............................................................................................................... 2 5
Social Cognitive Theory ........................................... ........... ............... 27
Personal Determinants ............... .................. ............... 27
Environm mental D term inants ......................................................................28
B ehavioral D eterm inants .............................................................................. 29
O observational L earning ............................................... ............................ 30
R einforcem ent .................. .. ................................ ........... ..... 30
Su m m ary .................... .... ..................... .... ... ..................3 1
The Relationship between the Three Explanatory Variables and Smoking
Initiation ................ .... ... .. ..... ..... ........................... 31
Perceived Prevalence of Peers' Smoking Behavior ................. ............... 31
Smoking-Related Behavior of Friends ....................................................34
Sociability ............................................. 37
C conclusion ................................................. 38
S u m m a ry ............................................................................................................... 3 9

3 M E T H O D S ........................................................................................................... 4 0

R e se arch D e sig n ................................................................................................... 4 0
R research V a riab les ............................................................................................... 4 0
Instrum ent D evelopm ent .............................................................. 41
Item Selection and Development ................ ............ ........ ......... 41
Expert Review ..................................................................... .... ......... .......... ........ 42
Student Evaluation ............... ......... ......... ........ 42
P ilo t S tu d y ................................................................4 3
S ettin g .................................................................................................... 4 3
Participants ............................................................................... ..... .................. 43
D ata C collection Procedures ....................................................... 43
Data Analysis ....................................................................... .... ......... .......... ........ 44
Instrument Reliability ............................ ........................ 44
Final Survey Administration........................ .......... ......... 45
F in al Su rv ey F orm at...................................................................................... 4 5
Final Survey Description............................. ..............................45
S ettin g ............. ...................................................................................... 4 6
Participants ................. ....... .... ............ 46
Protection of Confidentiality .................. ............ ................ .... .........47
D ata C collection Procedures ....................................................... 47
Data Analysis .. ......... ... ....................... .... .................. 47
C controlling for B ias ......... ................ ........................................ .....................48
Sum m ary........................ .. ............... ................ 49

4 RESULTS ................................... .. ... .... ..................50

Sample Demographics and Characteristics .......................................................50
Participation by University, Gender, and Race ...........................................51
Participation by Age and Months in College ................................................ 51









Current Sm oking Behavior ............. ........................ ....................................... 52
Smoking Behavior by Gender and Race .................................. ............... 53
Frequency and Quantity Sm oked ............................................. ............... 54
Smoking Initiation Behavior......................................................... 55
First Ever Smoking Behavior by Age ...................................... ............... 56
First Ever Smoking Behavior by Time in College ...........................................57
Correlations Between the Covariates and Smoking Initiation............................... 63
R research Q question R results .............................................................. .....................65
R research Q question O ne ............................................... ............................ 65
R research Question Tw o ......................................................... .............. 66
R research Question Three ..................... ... ................................ ............. 67
Regression Analysis of All Three Explanatory Variables...............................68
S u m m a ry ............................................................................................................... 6 9

5 D ISC U S SIO N ............................................................................... 70

C current Sm oking B behavior .............................................................. .....................70
Sm king Initiation R results .................................................... ............................ 72
How Many Students Start Smoking in College? ...........................................72
When Do Most College Student Begin Smoking? ...........................................74
Correlations Between Covariates and Smoking Initiation..................................75
R research Q question R results .............................................................. .....................77
R research Q question O ne .............................................. ............................. 78
R research Q question Tw o ......... ................. .................................. ... ............ 79
R research Q question Three......... ................. ................................ .. ............. 80
Strengths and W weaknesses ........... ............... .............. ................. ............... 81
Conclusions .......................... ...................... ..... 83
R ecom m endations ........... .............. ...................................... ... ....... ...... .... 84
Recommendations for College Health Practice........................................84
Recommendations for Future Research................... ....... ..................85

APPENDIX

A LIST OF EXPERT PANEL M EM BERS ........................................ .....................87

B INSTITUTIONAL REVIEW BOARD APPLICATIONS AND APPROVAL
L E T T E R S .......................................................................... 8 8

C PARTICIPANT CONSENT LETTERS .................... ...........................119

D THE COLLEGE STUDent SMOKING SURVEY ...............................................127

E FINAL SURVEY SCALE ANALYSES ...................................... ............... 137

F LIST OF FACULTY ADVISORS ........................................ ....................... 142

G DATA COLLECTION PROTOCOL......................... ................................143









L IST O F R E FE R E N C E S ......................................................................... ...................145

B IO G R A P H IC A L SK E T C H ........................................... ...........................................151
















LIST OF TABLES


Tablege

2-1 Cigarette Smoking Prevalence Rates by Year and Source.................................14

2-2 Cigarette Sm oking Prevalence Rates by Race..................................... ..................16

2-3 College Students' M otivations for Smoking ............... ............ ....... .............19

2-4 Stage Models of Smoking Acquisition, 1987-1998 ..........................................22

2-5 Stage Models of Smoking Acquisition, 2000-2002 .............................................23

2-6 Actual Prevalence and Perceived Prevalence of Smoking Among College
S tu d en ts ............................................................................. 3 3

3-1 Instrument Specification Table .................... ......... ........ .......... 45

4-1 Distribution of Participants by University, Gender, and Race .............................. 51

4-2 Distribution of Participants by Age and Months in College ..................................52

4-3 Past 30-Day Smoking Prevalence Rates.......... ............ .... ..... .. ..... .......... 53

4-4 Past 30-Day Smoking Prevalence Rates by Gender and Race.............. ...............54

4-5 Number of Days and Quantity Smoked...... ....................... .............55

4-6 First Ever Smoking Behavior by Age for Sample and Current Smokers................56

4-7 First Ever Smoking Behavior by Time in College for Sample, Current Smokers,
and Ever Sm okers ............ .... ............. ............................................... 58

4-8 First Ever Smoking Behavior by Year in School for all Participants......................60

4-9 College-Initiated Smoking Behavior by Time in College .....................................62

4-10 College-Initiated Smoking Behavior by Year in School.................. ..................64

4-11 Correlations between Five Suspected Covariates and Smoking Initiation ..............65









4-12 Multiple Logistic Regression Analysis of Perceived Peer Smoking Behavior
and Sm oking Initiation .................................................. ............................... 66

4-13 Multiple Logistic Regression Analysis of Friends' and Roommates' Smoking-
Related Behavior and Smoking Initiation ........................................ ............67

4-14 Multiple Logistic Regression Analysis of Sociability and Smoking Initiation........68

4-15 Multiple Logistic Regression Analysis of Three Explanatory Variables of
Sm oking Initiation......... ............................................................ ...... .....68

E-1 Reliability Analysis for Perceived Prevalence of Peers' Smoking Behavior
S cale ..............................................................................13 7

E-2 Item-Total Statistics for Perceived Prevalence of Peers' Smoking Behavior
S cale ..............................................................................13 7

E-3 Reliability Analysis for Smoking-Related Behavior of Friends and
R oom m ates Scale ...................... .................... .. ............... ....... 138

E-4 Item-Total Statistics for Smoking-Related Behavior of Friends and
R oom m ates Scale ...................... .................... .. .... .......... ....... 138

E-5 Reliability Analysis for Sociability Scale ....... .... ............... ................ ......... 139

E-6 Item-Total Statistics for Sociability Scale......... .. ................ ............. .. ........ 139

E-7 Reliability Analysis for Susceptibility to Smoke Scale.......................................140

E-8 Item-Total Statistics for Susceptibility to Smoke Scale .............. .....................140

E-9 Reliability Analysis for Alcohol Consumption Scale ........................................141

E-10 Item-Total Statistics for Alcohol Consumption Scale ........................................141
















LIST OF FIGURES

Figure page

2-1 Psychosocial risk factors of cigarette smoking ....................................................... 26

2-2 A schematic of the explanatory variables and covariates within a Social
Cognitive Theory fram ew ork. ............................................................................. 28















Abstract of Dissertation Presented to the Graduate School of the University of Florida in
Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy



SOCIAL FACTORS OF CIGARETTE SMOKING INITIATION
AMONG UNDERGRADUATE COLLEGE STUDENTS


By

Jane F. Emmeree

August 2003

Chair: Jill Vames
Major Department: Health Science Education

This study examined whether three social factors affected cigarette smoking

initiation during the freshman and sophomore years of college: (1) perceived prevalence

of peers' smoking behavior, (2) smoking-related behavior of friends and roommates, and

(3) sociability. Moreover, this study determined the percentage of students who initiated

smoking during college, identified at-risk populations, and pinpointed when college

students are most vulnerable to begin smoking. A total of 1,772 undergraduate students

in intact classrooms from five universities completed a researcher-constructed self-report

survey. Questionnaires from 1,250 full-time, traditional-aged freshmen, sophomores, and

juniors were analyzed. A total of 26.8% of the sample had smoked cigarettes during the

30 days before the survey, and 9.3% of the sample had smoked for the first time on two

or more days during any one month during college. Nearly 57% of these "smoking

initiators" had started smoking during their first six months of college; 70.7%, by the end









of their freshman year; and, 90%, by the end of their sophomore year. Moreover, one out

of four participants who reported smoking at least one cigarette during the 30 days before

the study had started smoking in college.

Results of multiple logistic regression analyses showed that the smoking-related

behavior of friends and roommates was the only statistically significant explanatory

variable (p = .003). This relationship remained after controlling for the effects of

perceived prevalence of peers' smoking behavior, sociability, susceptibility to smoke

during freshman year, and alcohol consumption during freshman year. The coefficient

(b) for friends' and roommates' smoking-related behavior was 0.842. This value

suggests that the odds of initiating smoking during freshman or sophomore year of

college increases by e0842= 2.321-or more than doubles-for each unit increase in the

composite score of the smoking-related behavior of friends and roommates.

Findings will assist college health professionals (1) to identify students at risk to

begin smoking, (2) to develop and tailor anti-smoking initiatives, (3) to implement

interventions at the most strategic times during the academic year, and, (4) to help at-risk

students remain smoke-free during their college years.














CHAPTER 1
INTRODUCTION

Cigarette smoking has many negative consequences. Between 1995 and 1999,

smoking caused over 440,000 adult and over 1,000 infant deaths annually (Centers for

Disease Control and Prevention [CDC], 2002). Currently, smoking causes one out of

every five deaths, making it the leading preventable cause of death in the United States

(CDC, 1999). Furthermore, annual economic losses total over $157 billion-the

equivalent of $3,400 for every smoker in the United States (CDC, 2002). "If current

smoking patterns continue, an estimated 25 million people alive today will die of

smoking-related illnesses" (CDC, 1999, p. 41).

The Surgeon General first warned Americans about the health hazards of cigarette

smoking in 1964. At that time, one out of every two adults smoked. Since then, over 10

million people have died from smoking-related illnesses (CDC, 1997), and 56.3 million

Americans continue to jeopardize their health despite repeated warnings (Substance

Abuse and Mental Health Services Administration [SAMHSA], 2002).

Persons of all ages choose to ignore public health warnings related to smoking, but

adolescents and young adults have the highest smoking rates. Approximately 11% of

middle school students (CDC, 2001), and 34% of students in grades 9-12 have smoked at

least one cigarette in the past 30 days (United States Department of Health and Human

Services [USDHHS], 2000). Moreover, in 2001, approximately 33% of full-time,

traditional-aged college students have smoked in the past 30 days (SAMHSA, 2002).









Smoking rates among college students are comparable with those of high school students,

and therefore, pubic health efforts for smoking prevention and cessation are warranted.

Research Problem

This study determined the percentage of college students who began smoking in

college and examined the influence of three social factors on cigarette smoking initiation

during the freshman and sophomore years of college. These factors included (1)

perceived prevalence of peer smoking behavior, (2) smoking-related behavior of friends

and roommates, and (3) sociability.

Purpose of the Study

This study was conducted to accomplish three purposes. (1) This study determined

the percentage of college students who began smoking in college, identified at-risk

populations, and pinpointed when college students are most vulnerable to begin smoking.

This information is useful for targeting at-risk populations and for implementing these

anti-smoking initiatives at the most appropriate times. (2) The study examined the effect

of three social factors and smoking initiation during freshman and sophomore years of

college. Findings will help college health professionals develop tailored interventions to

help vulnerable students remain smoke-free during college. (3) The study helped to fill a

void in the professional literature regarding smoking initiation among college students.

Currently, a voluminous amount of research on smoking initiation has been conducted on

middle school and high school students, whereas only a limited amount of research exists

on smoking initiation among college students.

Rationale for the Study

Research on cigarette smoking initiation among college students is important for

four reasons: (1) Smoking negatively contributes to serious health consequences, and a









significant number of college students currently smoke, (2) reducing cigarette smoking is

both a national and a college health priority, (3) nicotine is a highly addictive substance,

and subsequently, smoking cessation is extremely difficult, and (4) research on smoking

initiation among college students is limited.

The Negative Health Consequences of Smoking

A significant number of college students are putting themselves and others at risk

for serious health consequences because of smoking, and the magnitude of these

consequences is enormous. Smoking is the number one preventable illness behavior in

the nation, resulting in more deaths each year than AIDS, alcohol, cocaine, heroine,

homicide, suicide, motor vehicle crashes and fires combined (USDHHS, 2000).

Smoking has been conclusively linked to lung, kidney, bladder, laryngeal, oral,

esophageal, and pancreatic cancer, cardiovascular disease, chronic obstructive pulmonary

disease, and stroke in both males and females (Samet, 2001). Moreover, female smokers

have an increased risk for infertility, delayed conception, ectopic pregnancy, and

spontaneous abortion when compared with nonsmokers (USDHHS, 2001 a).

Smoking not only negatively affects the health of smokers, but also harms

nonsmokers. Each year, 35,000 nonsmoking adults die of heart disease, and another

3,000 die of lung cancer, from passive smoke (CDC, n.d.) Furthermore, environmental

tobacco smoke has been linked to the exacerbation of asthma, reduced lung functioning,

and respiratory symptoms (Samet, 2001). Due to the enormous health hazards to

smokers and nonsmokers alike, helping nonsmoking college students refrain smoke-free,

and assisting those who currently smoke with cessation efforts, is critical.









Smoking Reduction as a Current Health Priority

Because smoking is the number one preventable cause of disease, health

professionals want to reduce adult smoking in the United States from 22% to 16% by the

year 2010 (USDHHS, 2000). Although the prevalence rate for smoking among college

students (26-33%) exceeds the national rate (22%), campus health professionals have set

the same goal as those who set the national agenda-a reduction of smoking to 12% or

less among college students by the year 2010 (Task Force on National Health Objectives

in Higher Education, 2002). To accomplish this ambitious goal, national and college

public health efforts must prevent the initial onset and progression of smoking as well as

encourage and help smokers to quit.

Smoking Cessation as a Difficult Goal

Reducing smoking to 12% among college students by the year 2010 can be

achieved in two ways-either nonsmoking college students must remain smoke-free, or

those who currently smoked must quit. Quitting smoking, however, proves to be very

difficult and is typically characterized by frequent relapses (CDC, 1998). For instance,

seven out of ten current smokers in the United States want to quit smoking, but because

of the highly addictive nature of nicotine, most smokers who try to quit repeatedly fail

(USDHHS, 2000). Annually, 45% of smokers who try to quit smoking succeed for at

least 24 hours every year, but fewer than 3% quit permanently (USDHHS, 2000).

A large percentage of college smokers also want to quit smoking. Nearly two-

thirds have ever tried to quit smoking (CDC, 1997), and 50% to 80% have tried to quit

smoking during the year prior to being surveyed (Debernardo et al., 1999; Everett et al.,

1999). One study reported that one in five college students tries to quit five or more

times annually (Wechsler, Rigotti, Gledhill-Hoyt, & Lee, 1998). Although the majority









of college smokers want to quit, only one in four succeed (Everett et al., 1999). Because

the success rate of smoking cessation is so bleak, the "optimal public health strategy is to

prevent cigarette smoking completely" (CDC, 1998, p. 387).

Limited Research on College Students

To make progress toward the Healthy Campus 2010 goal of reducing smoking on

college campuses, the American College Health Association (ACHA) has recommended

that prevention and education initiatives address the risks of tobacco use and support non-

use (ACHA, 2000). Since the 1998 multi-billion dollar Master Tobacco Settlement

Agreement, funding for tobacco prevention and cessation programs has increased on

college campuses nationwide. To develop effective anti-smoking initiatives, however,

more than money is required. College health practitioners also must have relevant

information and tools. Currently, research on smoking initiation among college students

is limited and cannot adequately inform college health practitioners.

This study will help college health professionals move toward their 2010 goal of

reducing smoking among college students by providing useful information. Findings will

assist college health professionals in four ways: (1) to identify students at risk to begin

smoking, (2) to develop or refine tailored anti-smoking initiatives, (3) to implement

interventions at the most appropriate times during the academic year, and ultimately,

(4) to help at-risk students remain smoke-free during their college years. Furthermore,

college health practitioners can use the College Student Smoking Survey to determine the

nature and scope of cigarette smoking on their particular campus.

Smoking Initiation Among College Students

Primary prevention involves "actions and interventions designed to identify risks

and reduce susceptibility or exposure to health threats prior to disease onset" (American









Association for Health Education [AAHE], 2001, p. 101). In the case of smoking

initiation among college students, the goal is to identify nonsmoking students who are at

risk of becoming smokers in college, and then to create and deliver tailored interventions

to help them remain smoke-free. Although prevention of smoking among college students

is warranted, the preponderance of primary prevention research and interventions have

focused on middle and high school populations because four out of five smokers begin to

smoke before the age of 18 (USDHHS, 2000). Approximately 20% of smokers begin

smoking in their young adult years, representing around 11 million individuals who can

be categorized as "late-onset" smokers. Rates of smoking initiation among college

students cannot be determined from existing data.

Smoking initiation among college students is not well studied. Furthermore, all

national data collected from college students about smoking initiation have been

collected by age rather than by educational setting. Results from the 1995 National

College Health Risk Behavior Survey (NCHRBS) indicated that 22.6% of college

students who had ever smoked began smoking when they were either 17 or 18 years old.

A total of 11% began when they were age 19 years old or older (Everett et al.,1999;

Wechsler et al.,1998). Results from a statewide survey administered in California

showed that among students who had ever smoked, 14.3% smoked a whole cigarette for

the first time at the age of 19 or older (Patrick, Covin, Fulop, Calfas, & Lovato, 1997).

From these two studies, we can presume that at least 11% to 14.3% of college students

who had ever smoked began the behavior during college. To accurately assess how many

college students start smoking during college, however, data must be collected and

analyzed by educational setting rather than by age.









Only one study of college students has assessed smoking initiation by educational

setting. Researchers at Yale University determined whether college smokers began

smoking prior to or during college (Debernardo et al., 1999). They also assessed how

many nonsmoking students were at risk to become smokers in the future. Researchers

found that 5.7% of the participants began to smoke cigarettes in college. Among the

11.3% of the sample who considered themselves current smokers, half indicated that they

started smoking in college. Furthermore, 80% of participants who began smoking in

college started this behavior during their freshman or sophomore year. The researchers

also found that 19.4% of nonsmoking college students lacked a firm commitment to

refrain from smoking in the upcoming year. This finding was similar to national findings

from a high school sample-22.5% of students who had never smoked were considered

susceptible to trying cigarettes in the upcoming year (CDC, 2000).

The collective results from the study of Yale University students suggested that a

significant number are vulnerable to initiate smoking, particularly during their freshman

or sophomore year of college. Although intriguing, these findings need validation for

several reasons. First, college students from only one university participated in the study.

Second, only 13% of the sample smoked, which was significantly lower than the

corresponding national rate of 29% (CDC, 1997). Third, the definition for smoking was

not explicit. Participants were asked, "Do you smoke cigarettes now?" Current smoking

is typically measured using a specific time frame, usually the 30 days prior to the survey.

For example, many national surveys use the following question to measure current

smoking: "In the past 30 days, have you smoked cigarettes?" The inconsistency between

how these two ways measured smoking prevalence may explain the relatively low









prevalence rate in the Yale study. Fourth, and finally, the small sample size of college

smokers limits the conclusions of this study. Only 58 college smokers-and only 29

smoking initiators-participated in this study. Because of these limitations, findings

from this study are inconclusive and offer limited direction to health educators for

designing smoking prevention programs.

As mentioned previously, most individuals who smoke begin smoking during

adolescence (USDHHS, 2000; Everett et al., 1999). Subsequently, the preponderance of

research on cigarette smoking initiation investigates middle and high school populations.

However, this voluminous body of literature may shed light on why college students

begin to smoke. Rigorous studies of middle and high school students unequivocally

show that a variety of social factors increase the risk of starting to smoke. These social

risk factors include having friends, parents or older siblings who smoke; receiving offers

and encouragement to smoke; being socially competent; and having an inflated

perception of the norm for smoking (Carvajal, Wiatrek, Evans, Knee, & Nash, 2000;

Conrad, Flay, & Hill, 1992; Flay, Hu, & Richardson, 1998; USDHHS, 1994; West &

Michell, 1999; West, Sweeting, & Ecob, 1999).

Research Questions

The following questions were investigated in this study.

1. Does the perceived prevalence of peers' smoking behavior during freshman year of
college increase the risk of smoking initiation during the first two years of college?

2. Does the smoking-related behavior of friends and roommates during freshman year
of college increase the risk of smoking initiation during the first two years of
college?

3. Does sociability during freshman year of college increase the risk of smoking
initiation during the first two years of college?









Delimitations

The following delimitations were noted in this study.

1. Participants were selected from undergraduate students at one of five universities:
(1) Ball State University, (2) Texas A & M University, (3) University of Alabama,
(4) University of Florida, or (5) University of Georgia.

2. Participants were traditional undergraduate students.

3. Participants reported beginning college within 36 months of the survey.

4. Participants reported having full-time status during their first year of college and at
the time of the survey.

Limitations

The following limitations were identified for this study.

1. Participants selected from the five universities may not represent a nationwide
undergraduate student population.

2. Participation in the study was voluntary.

3. The diversity of the sample was limited to students who chose to participate in the
study.

4. The generalizability of the results was limited to the sample of participants in this
study.

Assumptions

The following assumptions were made for this study.

1. The five selected universities were representative of other universities in the United
States.

2. Individuals who voluntarily agreed to participate were representative of all
undergraduate students enrolled at their university.

3. The data collection instrument developed for this study was adequate for the
purpose of this study.

4. Participants were able to accurately recall their perceptions and behaviors during
their freshman year of college.









Definition of Terms

The following is a list of relevant terms for this study.

Extroversion. A preference for attending to the outer world of events by seeking

active stimulation and involvement in the environment (Morris, 1979).

First year of college. The first 12 months of attendance at a two- or four-year

college or university, after graduating from high school. Synonymous to "freshman year

of college."

Freshman year of college. The first 12 months that a student attends a two- or

four-year college or university, after graduating from high school.

Health promotion. Any planned combination of educational, political,

environmental, regulatory, or organizational mechanisms that support actions and

conditions of living conducive to the health of individuals, groups, and communities

(AAHE, 2001).

Perceived prevalence of peers' smoking behavior. An estimate of the

percentage of the participant's peers who engaged in a variety of smoking-related

behaviors during their freshman year of college.

Primary prevention. Actions or interventions designed to identify risk and

reduce susceptibility or exposure to health threats prior to disease onset (AAHE, 2001).

Second year of college. The second 12 months of attendance at a four-year

college or university, after graduating from high school. Synonymous to "sophomore

year of college."

Smoking. Actively inhaling smoke from regular or clove cigarettes.

Smoking initiation. The first time that a participant smoked on two or more days

during any one month.









Smoking-related behavior of friends. Smoking cigarettes, offering cigarettes to

others, or encouraging others to smoke.

Sociability. The tendency to prefer the presence of others to being alone and

seeking and being especially gratified by the presence, attention, sharing of activities, and

stimulation of others (Buss & Plomin, 1986).

Sophomore year of college. The second 12 months in attendance at a four-year

college or university, after graduating from high school. Same as "Second Year of

College."

Susceptibility. The lack of a firm commitment not to smoke in the future among

nonsmokers (Pierce, Choi, Gilpin, Farkas, & Merritt, 1996).

Traditional, undergraduate student. A student who began college between the

ages of 17 and 19 years old, and who were 18 to 22 years old at the time of the survey.

Summary

Each year, cigarette smoking costs thousands of premature deaths and millions of

dollars in lost productivity and health care costs in the United States. Consequently,

national and college health goals include reducing smoking prevalence rates to 12% or

less by the year 2010. To achieve this goal, aggressive efforts to prevent individuals

from starting to smoke, and to help smokers stop smoking, must occur nationwide

including on college campuses. Because most cigarette smoking begins during

adolescence, research has overwhelmingly focused on middle and high school students

rather than on college students. Nonetheless, over 46 million people in the United States

currently smoke, and one-fifth-or 9 million- began smoking as young adults. Recent

data suggest that a substantial number of young adults will begin to smoke during

college, particularly during their freshman or sophomore year (DeBernardo et al., 1999;









Everett et al., 1999; Wechsler et al., 1998). Therefore, research that examines potential

risk factors of smoking initiation among this population is important.

This chapter provided an overview of the study and included the research problem,

purpose, rationale, as well as a brief overview of smoking initiation among college

students. Additionally, research questions, delimitations, limitations, assumptions, and

definitions of key terms for the study were presented. Chapter 2 provides background

and context of cigarette smoking among college students, conceptual frameworks for the

study's variables, and a review of relevant literature on smoking initiation among college

students.














CHAPTER 2
REVIEW OF THE LITERATURE

This study determined the percentage of college students who began smoking in

college, identified at-risk populations, and pinpointed when college students are most

vulnerable to begin smoking. The effects between three social factors and smoking

initiation during freshman and sophomore years of college were also examined. The first

section of this chapter provides an overview of smoking and college students, including

prevalence rates and trends, smoking initiation among college students, and reasons why

college students smoke. The second section provides a conceptual framework for

smoking initiation. The third section provides a conceptual framework to explain the

relationship between the three explanatory variables and smoking initiation. The fourth

section reviews the empirical support between each of the three explanatory variables and

smoking initiation.

Cigarette Smoking Among College Students

The purpose of this section is to provide a background and context of cigarette

smoking among college students. First, prevalence rates and trends will be described,

and then, smoking initiation among college students will follow. This section will

conclude with reasons why college students smoke.

Smoking Prevalence Rates Among College Students

National surveys administered between 1993 and 2001 indicate 30-day smoking

prevalence rates for college students ranging from 22.3%-32.9%. (ACHA, 2002; CDC,









1997; Johnston, O'Malley, & Bachman, 2001; SAMHSA, 2002; Wechsler et al., 1998).

Table 2-1 shows various prevalence rates by year and source.

Table 2-1. Cigarette Smoking Prevalence Rates by Year and Source
Year of 30-Day Annual Life-time Source
Survey Prevalence Prevalence EVER
Prevalence
1993 22.3% 31.6% Wechsler et al., 1998
1995 29.0% 74.8% CDC, 1997
1997 28.5% 39.5% Wechsler et al., 1998
1999 40.4% 48.5% 69.0% SAMSHA, 2000
2000 28.2% 41.3 Johnston et al., 2001
32.0% ACHA, 2002
2001 25.0% ACHA, 2002
32.9% 69% SAMSHA, 2002
2002 26.5% ACHA, 2002


Frequency and Quantity Smoked

The 30-day prevalence rates reported above include any amount of cigarette use

during the month prior to the survey. Prevalence rates are also calculated based on how

frequently students smoke, and how many days they smoke per month. For example, in

1995, 16.5% students smoked on 20 or more days during the month prior to being

surveyed (CDC, 1997). In the fall of 2000 and the spring of 2002, these rates decreased

from 15% to 11.9% (ACHA, 2002). In 2000, daily smoking rates among college students

ranged between 10.6% and 18% (ACHA, 2002; Johnston et al., 2001). In the spring of

2002, 8.0% of college students smoked daily (ACHA, 2002).

In addition to tracking the number of days per month smoked, Wechsler and

colleagues (1998) assessed the number of cigarettes that college students smoked on the

days that they smoked. Among the 28.5% of participants who smoked during the month

prior to the survey, 43.6% smoked fewer than one cigarette on the days that they smoked.

Moreover, 24.2% smoked between one and nine cigarettes; 20.4% smoked 10-19









cigarettes; and 11.8% smoked 20 or more cigarettes, on average, on the days that they

smoked.

Smoking Rates by Gender

Findings from five national studies between 1993 and 2002 show similar smoking

rates for male and female college students (ACHA, 2002; CDC, 1997; Emmons,

Wechsler, Dowdall, & Abraham, 1998; Johnston et al., 2001; Wechsler et al., 1998).

Thirty-day prevalence rates measuring any amount of cigarette use, 30-day daily

prevalence rates, and annual prevalence rates were similar between males and females

(ACHA, 2002; Johnston et al., 2001). For example, National College Health Assessment

data collected in the spring of 2002 show that 27.6% of males and 25.8% of females

reported smoking at least once in the 30 days prior to the survey (ACHA, 2002). A total

of 12.1% of males, and 11.7% of females smoked on 20 or more days, and 8.5% of males

and 7.7% of females smoked on all 30 days prior to the survey.

Smoking Rates by Race

Although rates of smoking do not significantly differ by gender, they do

significantly differ by race. According to the 1995 National College Health Risk

Behavior Survey, (CDC, 1997) white students were significantly more likely than black

students or Hispanic students to report smoking: (1) any time in their lifetime, (2) any

time during the month prior to the survey, (3) on 11 or more days during the month prior

to the survey, (4) on 20 or more days during the month prior to the survey, and (5) 30

days during any month during their lifetime. Additionally, Hispanic students were

significantly more likely to report lifetime cigarette use and current use as were black

students. Rates for other racial groups were not available from this study.









Although more white students smoke than do blacks or Hispanics, researchers at

Harvard University found significant increases in smoking among all racial groups during

between 1993 and 1997 (Wechsler et al., 1998). Table 2-2 shows these data. Although

black students had the lowest 30-day prevalence rate (13.7%), they had the highest

increase in smoking between 1993 and 1997 (47%). The rates for Hispanic students also

increased (12.0%), but this group had the lowest rate of increase compared to other racial

groups.

Table 2-2. Cigarette Smoking Prevalence Rates by Race
Race 30-day prevalence 30-day prevalence % Increase
in 1993 in 1997
African American 9.6 13.7 42.7
Asian/Pacific Islander 18.3 22.4 22.5
Hispanic 22.7 28.8 12.0
White 23.2 30.4 31.2


Smoking Trends Among College Students

During the 1990s, overall smoking rates among college students steadily increased.

Wechsler and his colleagues (1998) used data from the Harvard School of Public Health

Alcohol Study to compare smoking rates between 1993 and 1997. In 1993, 22.3% of

students smoked during the month prior to the survey; however, 28.5% smoked in 1997.

This represented a 27.8% increase in smoking over this 4-year time period. During that

time, rates also significantly increased for all student subgroups including gender,

race/ethnicity, and year in school, as well as among students attending both private and

public universities.

Findings from the Monitoring the Future Study verified the Harvard researchers'

results (Johnston et al., 2001). Thirty-day prevalence rates in 1991 and 1999 were 23%

and 31%, respectively, representing a 30% increase. Most recent data from this on-going









study, however, suggest that this upward trend may be changing. Findings from the year

2000 indicate that the 30-day smoking prevalence rate among college students slightly

decreased for the first time since 1991. This 2.4% decline (from 31% to 28.2%) was not

a statistically significant change. Results from the 2001 National Household Survey on

Drug Abuse, however, show a higher 30-day smoking prevalence rate than that found in

the 2000 Monitoring the Future Study (SAMHSA, 2002). Nearly 33% of full-time

college students, and 44.6% of part-time students, aged 18-22 years had smoked some

time during the month prior to the survey.

Data from the Monitoring the Future Study and from the National College Health

Assessment suggest that daily smoking may be leveling off after a steady increase

throughout the 1990s (ACHA, 2002; Johnston et al., 2000). Findings from the former

study show that the number of college students who smoked daily increased between

1991 and 1999 by 401%-from 14% to 19%, respectively. Between 1999 and 2000,

however, the rate dropped from 19% to 17.8%. Similarly, results from three annual

administrations of the National College Health Assessment also showed a downward

trend in daily smoking among college students. These data indicated that 10.6%, 8.6%,

and 8.0% of college students smoked daily in the 2000, 2001, and 2002, respectively

(ACHA, 2002).

Smoking Initiation Among College Students

Research has been conducted that consistently assesses how many college students

currently smoke, how much, and how often, however, research is lacking that assesses

how many college students begin to smoke during college. Only two national studies

have collected smoking initiation data, but unfortunately, these data were collected by

age rather than educational setting. Nonetheless, these age-specific data can shed some









light on smoking initiation behavior among college students. For example, results from

the 1995 National College Health Risk Behavior Survey indicate that among students

who had ever smoked, nearly nine out of ten smoked a whole cigarette for the first time at

the age of 18 or younger (Everett et al., 1999). More specifically, 18.5% smoked a whole

cigarette for the first time when they were 12 years old or younger; 24.5% when they

were 13 or 14 years old; 23.7% at the age of 15 or 16 years, and; 22.6% when they were

either 17 or 18 years old. Approximately 11% smoked their first cigarette at age 19 years

or older. Furthermore, a national study conducted in 1997 found that 11% of student

smokers smoked their first whole cigarette at the age of 19 or older (Wechsler et al.,

1998). From these findings, we can assume that at least 11% of college smokers may

begin smoking during college. However, to precisely assess smoking initiation during

college, students must be asked whether they began smoking before or during college.

A random study conducted among Yale University students assessed how many

current smokers began smoking in college (Debernardo et al., 1999). Although current

smokers comprised only 13.0% of the sample, half of them indicated that they started

smoking in college. Furthermore, 80% of these smokers who began smoking in college

started smoking during their freshman or sophomore year. These results suggest that a

significant number of college students are vulnerable to start smoking in college,

particularly during their freshman or sophomore year.

Reas ons Why College Students Smoke

Two studies of college students have investigated reasons why college students

smoke. One national study used multivariate logistic regression to assess predictors of

smoking among current college smokers (Emmons et al., 1998). Their sample included

students from 140 randomly selected 4-year colleges in the United States. The found that









seven variables increased the likelihood of being a smoker in college: binge drinking

prior to college; binge drinking during college; using marijuana during college; having

multiple sex partners in college; participation in leisure activities during college;

dissatisfaction with college education; and choosing not to participate in athletics during

college. Moreover, females who engaged in three high-risk behaviors-binge drinking,

using marijuana, and having multiple sex partners during college-were more likely to

smoke than their male counterparts. However, researchers of this study did not assess

how many smokers in their sample began to smoke in college, nor did they investigate

whether these seven variables predicted smoking initiation during college. Nevertheless,

their findings provide insight into possible correlates of smoking behavior among college

students.

The second study took place at Yale University. This small-scale study examined

reasons why college students smoked and also why those who were contemplating

smoking would begin to smoke (Debernardo et al., 1999). Descriptive results are shown

in Table 2-3. According to these findings, large differences exist between smokers and

Table 2-3. College Students' Motivations for Smoking
Motivation Smokers (%) Nonsmokers (%)
Stress 49.3 6.5
Image smokers project 39.1 2.5
Expression of independence 36.2 1.8
Friends smoke 11.6 11.1
Depression 31.9 2.7

nonsmokers' motivations for smoking. For example, nearly one out of two smokers

reported that stress motivated them to smoke. Conversely, less than one in ten

susceptible nonsmokers thought that stress would motivate them to begin to smoke. A

similar percentage of smokers and nonsmokers reported friends' smoking behavior as









motivator for smoking. Approximately 11.6% of the smokers claimed that friends who

smoked influenced their current smoking behavior. Similarly, 11.1% of the non-smokers

implicated friends who smoke as potentially influencing their future smoking behavior.

A Conceptual Framework for Smoking Initiation

Cigarette smoking has been conceptualized as a gradual progression through a

series of stages. In general, researchers have divided the process into four to six stages

beginning with not smoking and ending with established smoking and addiction (Flay,

1993, Mayhew, Flay, & Mott, 2000; Pallonen, Prochaska, Velicer, Prokhorov, & Smith,

1998; Stem, Prochaska, Velicer, & Elder, 1987). Two similar, yet distinct, types of these

stage models have emerged (See Table 2-4). The first type, based on the

Transtheoretical Model of smoking cessation, classifies smoking into four or five stages:

precontemplation, contemplation, decision-making (or preparation), action, and

maintenance (Stern et al., 1987; Pallonen et al., 1998). The second type of stage model

also divides the process of smoking into five stages, but uses different labels for each

stage: preparatory, trial, experimentation, regular use, and nicotine dependence/addiction

(Flay, 1993) [See Table 2-4].

Recently, scholars have merged these two types of stage models into yet another

five-stage framework ranging from non-smoking to established/daily use (Mayhew et al.,

2000). Their conceptualization is depicted in Table 2-5. The researchers collapsed the

precontemplation, contemplation, and preparation stages of the Transtheoretical Model

into one, non-smoking stage with two sub-stages based on smoking intention The four

stages that follow the two nonsmoking stages use Flay's (1993) labels and descriptions:

tried, experimentation, regular use, and established/daily use.









Incorporating Susceptibility into the Stages

The concept of susceptibility is a recent addition to the stages of smoking

acquisition (Prokhorov et al., 2002). Susceptibility as a measure for smoking initiation

was first introduced in 1993 in a report of tobacco use in California. Since then, it has

been found to be a strong predictor of smoking initiation among children and adolescents

(Choi, Pierce, Gilpin, Farkas, & Berry, 1997; Distefan, Gilpin, Choi, & Pierce, 1998;

Jackson, 1998; Pierce et al., 1996; Pierce et al., 1993; Unger, Johnson, Stoddard, Nezami,

& Chou, 1997). Consequently, susceptibility has been incorporated into the Youth

Tobacco Survey (USDHHS, 2001).

Pierce and colleagues (1993) define susceptibility as the "absence of a determined

decision not to smoke in the future" (p. 43), and measure this construct with three

questions: (1) Do you think you will smoke a cigarette soon? (2) Do you think you will

smoke a cigarette in the next year? and (3) Do you think you would smoke if your closest

friend offered you a cigarette? Those who respond "definitely not" to all of three items

are classified as nonsusceptible, whereas, all others are considered susceptible to future

smoking behavior (Pierce et al., 1996).

Prokhorov and colleagues (2002) have recently suggested that integrating

susceptibility into the Transtheoretical Stage Model would improve its validity. More

specifically, they hypothesized that an adolescent's response to the question about a

closest friend's offer to smoke would further determine who was susceptible among those

who also had no intention of smoking in the future. Their integrated model proved to

have better predictive and concurrent validity than did either of the two individual












Table 2-4. Stage Models of Smoking Acquisition, 1987-1998
Source Stage 1 Stage 2 Stage 3 Stage 4 Stage 5


Stern et al., 1987


Precontemplation
Has not yet
considered smoking
and reports no desire
to do so in the future.
Perceives few
positive
consequences of
smoking.
Precontemplation
Has not tried
smoking and has no
intentions of smoking
in the next 6 months.

Preparatory
Involves the
formation of
knowledge, beliefs
and expectations
about smoking and
the functions that it
can serve.


Pallonen et al.,
1998





Flay, 1993


Contemplation
Thinking about
starting to smoke.
Perceives
relatively more
positive
consequences
than in Stage 1.

Contemplation
Never smokers
who are thinking
about trying to
smoke in the next
6 months.
Trying
The first two or
three times a
person smokes.
Peers are usually
involved.


Decision-Making
Pros and cons of
smoking are closely
balanced.






Preparation
Never smokers who
are thinking about
trying to smoke in
the next 30 days.

Experimentation
Repeated, but
irregular smoking;
often situation
specific (like at a
party).


Action
Experiments with
cigarettes, but not
totally committed
to smoking in the
future. Receives
minimal pleasure
from smoking.

Recent Action
Smoked
cigarettes
regularly, but for
less than 6
months.
Regular Use
Weekly smoking
across situations
and personal
interactions.


Maintenance
Smokes on a
regular basis.
Committed to
smoking now and
in the future. Has
no desire to stop.
Receives pleasure
from smoking.


Nicotine
Dependence and
Addiction.
Physiological need
for nicotine.
Established
tolerance and
withdrawal
symptoms with quit
attempts.












Table 2-5. Stage Models of Smoking Acquisition, 2000-2002


Source
Mayhew
et al., 2000








Prokhorov,
et al., 2002


Stage 1
Non-smoking
Preparation
Does not intend to
smoke.






Precontemplation
Nonsusceptible
Never smokers who
have no intention of
smoking in the next
6 months and state
that they would
"definitely not"
smoke if a friend
offered them a
cigarette.


Stage 3
Tried
Has smoked, but
no more than one
or two cigarettes.
Has not smoked
in the past year.



Contemplation
Never smokers
who are thinking
about trying to
smoke in the next
6 months.


Stage 4
Experimentation
Smokes
occasionally on
an experimental
basis. Does not
intend to become
a permanent
smoker.

Preparation
Never smokers
who are thinking
about smoking in
the next 30 days.


Stage 5
Regular Use
Smokes at
least monthly,
but not as
frequently as
daily.


Stage 6
Established/
Daily Use
Smokes daily or
almost daily.
May smoke
heavily on
occasion.
Nicotine
dependence.


Stage 2
Non-smoking
Contemplation
Intends to smoke.
Belief and attitude
formation;
susceptible to peer
pressure.


Precontemplation
Susceptible
Never smokers
who have no
intention of
smoking in the
next 6 months, but
fail to state that
they would
"definitely not"
smoke if a friend
offered them a
cigarette.









models. Furthermore, their integrated model predicted susceptibility among

adolescents who reported no intentions to smoke in the future. Based on these findings,

they divided the precontemplation stage of smoking acquisition into two sub-stages based

on their definition of susceptibility. They categorized individuals with strong convictions

to refuse a friend's offer to smoke as Nonsusceptible, and those with weak convictions as

susceptible (See Table 2-5).

Since the introduction of susceptibility as a relevant construct in smoking behavior,

researchers have validated ways to measure it among adolescents (Pierce et al., 1996;

Unger et al., 1997). Furthermore, the relationship between susceptibility and smoking

acquisition has been examined primarily among adolescent populations (Choi et al.,

1997, Jackson, 1997, 1998; Pierce et al., 1996; Prokhorov et al., 2002; Unger et al., 1997;

Unger, Rohrbach, Howard-Pitney, Ritt-Olsen, & Mouttapa, 2001). Little investigation of

this construct, however, has occurred among college populations.

Only one study has examined how many nonsmoking college students maybe

susceptible to begin smoking. DeBernardo and colleagues (1999) measured

susceptibility by asking students whether they "would try cigarette smoking in the next

12 months" (p. 63). They found that nearly one out of five nonsmoking participants were

susceptible to start smoking in the upcoming year. More specifically, their findings

indicated that 11.5% of non-smokers were considering smoking, and an additional 7.9%

were unsure whether they would continue to abstain from smoking, during the next year.

This study, however, did not examine the relationship between susceptibility and

smoking initiation among college students.









Risk Factors of Smoking by Stage

In addition to susceptibility to smoke in the future, a variety of other psychosocial

risk factors have emerged for smoking initiation and progression through the stages of

smoking. Conrad, Flay, and Hill (1992) reviewed 27 prospective studies on smoking

initiation published since 1980 and identified risk factors for each stage of smoking.

Their findings indicated that some type of social influence-like parent, adult, sibling,

and/or peer's smoking behavior and attitudes-plays a role at every stage of smoking

acquisition. Their findings are summarized in Figure 2-1.

Summary

This section provided a conceptual framework for smoking initiation and

progression to addiction. The process of becoming a cigarette smoker was depicted on a

continuum represented by four to six stages beginning with receptivity to smoking and

ending with nicotine dependence and addiction. Within the last 10 years, the concept of

susceptibility has been incorporated into two stage models of smoking acquisition.

Prospective studies of smoking initiation among adolescents have identified a variety of

psychosocial risk factors for each stage of the smoking process. Now that a conceptual

framework for the outcome variable of the study has been described, a conceptual

framework for the three explanatory variables, and the five covariates, will be provided.









1. PREPARATORY STAGE


Psychosocial risk factors include
advertising, adult/sibling role
models who smoke cigarettes, and
susceptibility to smoke.


Adolescent forms attitudes and
beliefs about smoking


Never smokes
2. TRYING STAGE


Risk factors include peer influences to
smoke--friends' smoking, approval,
and offers to smoke, the perception
that smoking is normative, the
availability of cigarettes, alcohol use,
and marijuana use.


Adolescent smokes first few
cigarettes.


No longer smokes


3.EXPERIMENTAL STAGE


Risk factors include social situations
and peers that support smoking,
parental smoking, self-efficacy in
ability to refuse offers to smoke, and
the availability of cigarettes.


Adolescent smokes repeatedly,
but irregularly.

No longer smokes


4. REGULAR USE


Risk factors include peers who
smoke, the perception that smoking
has personal utility, and few
restrictions on smoking in school,
home, and community settings


Adolescent smokes at least
weekly across a variety of
situations and personal


Quits smoking


5. ADDICTION/DEPENDENT
SMOKER
Adolescent has developed the
physiological need for nicotine.


Figure 2-1. Psychosocial risk factors of cigarette smoking.

Adapted from Flay (1993), Flay et al. (1998), Prokhorov et al. (2002), USDHHS (1991),
and USDHHS (1994, p. 92),









Social Cognitive Theory

Psychosocial risk factors have been implicated at every stage of the smoking

acquisition process, and therefore, a psychosocial theory is necessary to help explain the

progression of smoking from cognitive susceptibility to nicotine dependence. One of the

most promising models applied to smoking initiation is social cognitive theory (Wang,

Fitzhugh, Eddy, Fu, & Turner, 1997). Social cognitive theory explains behavior as a

process involving the dynamic interactions between personal, behavioral, and

environmental factors (Bandura, 2002). This on-going process is known as reciprocal

determinism. Figure 2-2 diagrams the study's three explanatory variables and five

covariates as a schematic showing reciprocal determinism between each pair of the three

types of determinants in the model. Each of these three determinants as well as two

concepts that are relevant to smoking initiation, observational learning and reinforcement,

are depicted and will be briefly described.

Personal Determinants

Personal determinants within social cognitive theory include thoughts, perceptions,

and emotions, as well as biological factors and genetic predisposition. More specific

examples of personal determinants include beliefs, expectations, self-perceptions, gender,

ethnicity, and temperament. In this study, two explanatory variables, sociability and

perceptions of peers' smoking behaviors fit within this category. Furthermore, two of

the five covariates in the study, susceptibility to smoke and grade point average, can be

considered personal determinants of smoking initiation.












Perceived prevalence of peers' smoking behavior
Sociability
Susceptibility
Grade point average



Personal Determinants





Observational Learning
Direct Reinforcement
Vicarious Reinforcement




Behavioral Determinants Environmental
Determinants


Cigarette smoking Smoking-related behavior of friends
Alcohol consumption Parents' smoking status
Living situation
Parents smoking status
Living situation

Figure 2-2. A schematic ofthe explanatory variables and covariates within a Social
Cognitive Theory framework.

Adapted from: Bandura (2002)

Environmental Determinants

In contrast to personal determinants that are internal to an individual,

environmental determinants are external factors that influence an individual's thoughts

and behavior. With few exceptions, smoking initiation occurs within social

environments. According to Bandura (2002), "self-development, adaptation, and change









are embedded in social systems" (p. 121). Traditional-aged college students exist in a

very dynamic and diverse social environment. Undergraduate students typically interact

with roommates, housemates, dorm-mates, classmates, co-workers, instructors, friends,

and/or romantic partners on a regular basis. Moreover, college students also tend to have

many group affiliations, like being members of academic clubs, intramural sports teams,

sororities or fraternities, or community groups.

Because of the dynamic and diverse nature of relationships within college students'

lives, the social environment plays a crucial role in this study. In particular, the

relationship of smoking-related behavior and interactions between friends, roommates,

romantic partners, and sorority sisters and fraternity brothers (if applicable) and smoking

initiation was examined. In addition, parents' smoking status and whether participants

lived alone or with other students was assessed.

Behavioral Determinants

Behavioral determinants interact with both personal and environmental

determinants. Social cognitive theory posits that a reciprocal relationship exists between

a person's actions and his or her environment. In the case of cigarette smoking, for

example, individuals who smoke may choose friends or roommates who also smoke.

Conversely, friends or roommates who smoke may influence a nonsmoker to start

smoking.

Another behavioral determinant that is relevant to smoking behavior is alcohol

consumption. As mentioned earlier in this chapter, binge drinking both before high

school and during college was predictive of smoking behavior in college. Therefore, this

study measured alcohol consumption during freshman year among its participants.









Observational Learning

Two other constructs within social cognitive theory are also very relevant social

influences of smoking initiation behavior. The first, observational learning, is defined as

"behavioral acquisition that occurs by watching the actions and outcomes of others'

behavior" (Baranowski, Perry, & Parcel, 1997, p. 157). A key tenant of social cognitive

theory is that human beings learn not only by direct experiences, but also vicariously

through observing others (Bandura, 2002). Moreover, vicarious learning can occur both

inside and outside of one's immediate environment. An example of vicarious learning

within one's environment would be observing a group of college students smoking

outside one's residence hall or sorority house, and creating interpretations about that

situation. An example of observational learning occurring outside one's environment

would be seeing an anti-smoking commercial on TV or seeing a character in a movie

smoking a cigarette, and subsequently forming thoughts, feelings, and interpretations

about what was seen.

Reinforcement

In addition to observational learning, social cognitive theory suggests that two

types of reinforcement influence thoughts and behavior. The first, known as vicarious

reinforcement, is related to observational learning. In the previous example of the group

of students smoking outside of a residence hall, vicarious reinforcement could occur if

the observer admired, liked, or was attracted to one or more of the individuals who were

smoking. In this instance, smoking may be positively reinforced in the observer. The

second type of reinforcement-direct reinforcement-is inherent in the explanatory

variable "Smoking-Related Behavior of Friends" as measured in this study. Respondents

indicated how many of their friends, roommates, romantic partners, and sorority sisters or









fraternity brothers smoked, offered them cigarettes, or encouraged them to smoke during

their freshman year of college. The composite score of these three behaviors for these

four types of relationships provided a measure of the level of positive reinforcement

received for initiating smoking.

Summary

This section provided a conceptual framework for the study's three explanatory

variables and for five suspected covariates. Social cognitive theory posits that substance

use is learned through observation and experiences within the social environment.

Moreover, behavioral, personal, and environmental determinants interact collectively and

dynamically to influence an individual's thoughts and actions. Observational learning,

vicarious reinforcement, and direct reinforcement were also described as they relate to

the study's variables. A diagram was created to integrate within a social cognitive theory

framework all three of the explanatory variables, all five of the covariates, and pertinent

concepts for this study. The next, and final section of this chapter, will review evidence

of the relationship between each of the three explanatory variables and cigarette smoking

initiation.

The Relationship between the Three Explanatory Variables and Smoking Initiation

Perceived Prevalence of Peers' Smoking Behavior

Social cognitive theory suggests that personal perceptions, whether real or

imagined, influence behavior. Considerable evidence indicates that perceived prevalence

of peers' smoking behavior influences the perceiver's smoking behavior. In particular,

children and adolescents are at an increased risk of becoming smokers if they

overestimate the number of their peers who smoke (Bauman, Botvin, Botvin, & Baker,









1992; Carvajal et al., 2000; Flay et al., 1998; Jackson, 1997; Unger & Rohrbach, 2002;

USDHHS, 1994).

Flay and his associates (1998), for example, conducted a longitudinal study of the

psychosocial predictors of the different stages of smoking among high school students.

Perceived smoking behavior of peers and adults was among one of the many variables

they measured among 7th graders. Cigarette smoking stages were assessed 6 years later,

when participants were in the 12th grade. Students who never smoked, not even a puff or

two, were classified as "never smokers." Those who had smoked only one whole

cigarette or less in their entire life were labeled triers. Those who had smoked more than

one whole cigarette in their lifetime, but had not smoked in the week preceding the

assessment, were classified as experimenters. Finally, students who smoked during the

week before they had completed the survey were categorized as regular smokers.

The relationship between perceptions of smoking and all three active smoking

stages, as well as, transitions between trying and experimentation, and experimentation

and regular use were assessed. The composite score of a 3-item scale served as the

measure for perceived prevalence of smoking. Two items assessed perceived prevalence

of peers' smoking behavior, and one measured perceived prevalence of adult smoking

behavior. The scale's internal consistency was .71 as measured by Cronbach's reliability

coefficient.

Results indicated that perceived prevalence of peers' smoking significantly

predicted experimentation and regular use. Prevalence estimates also predicted one of

the two transitions measured between smoking stages. The transition between trying

cigarettes to experimental use was significantly predicted by the participants' perceptions









of prevalence, however, this relationship was not found for the transition between

experimental smoking and regular use.

The study just described served as an example of the quality of many other

longitudinal studies of perceived prevalence and smoking among children and

adolescents (Bauman et al., 1992; Carvajal et al., 2000; Flay et al., 1998; Jackson, 1997;

Unger & Rohrbach, 2002). Collectively, findings provide strong evidence that perceived

prevalence of peers' smoking behavior influences smoking behavior. However, no

published study currently exists that has examined the relationship between perceived

peers' smoking prevalence and smoking initiation among college students. Only

descriptive data are available on prevalence estimates of peer smoking among college

students.

Descriptive data show that college students greatly overestimate the number of

peers' who have never smoked, as well as those who currently smoke (as measured as

any cigarette use during the 30 days before data collection). Furthermore, the magnitude

of these misperceptions has been fairly consistent over the past three years (see Table

2-6).

Table 2-6. Actual Prevalence and Perceived Prevalence of Smoking Among College
Students
Spring 2000 Spring 2001 Spring 2002
Actual Perceived Actual Perceived Actual Perceived
Never smoked 70.0% 9.0% 71.0% 8.0% 68.2% 6.1%

Smoked one or 30.0% 91.0% 25.0% 93.0% 26.6% 93.9%
more days in
past month
Source: American College Health Association, 2002.

Evidence also suggests that college students overestimate daily smoking behavior

among their peers (Page, 1998). In a study of 775 college students from one university,









female students, on average, estimated that 42.3% of male, and 42.2% of female peers,

smoked daily. Men also overestimated these daily prevalence rates but by a smaller

margin. They perceived that 36.9% of men, and 33.9% of women on their campus

smoked daily. Actual rates of daily smoking proved to be much lower-13.4% for men

and 17.8% for women.

Smoking-Related Behavior of Friends

Evidence from studies of elementary school students (Jackson, 1997), middle

school students (Carvajal et al., 2000; Conrad et al., 1992), and high school students

(Duncan, Tildesley, Duncan, & Hops, 1995; Flay et al., 1998; Killen et al., 1997; Wang

et al., 1997) clearly indicate that friends play a major role in smoking initiation and

escalation of use. For example, the first time an individual smokes most commonly

occurs with a friend who already smokes, and reinforcement from friends has been

implicated in further experimentation and progression of smoking (USDHHS, 1994).

Furthermore, over 90% of high school students who smoke report that one or more of

their closest friends smoke, compared to 33% of nonsmokers reporting having one or

more close friends who smoke (CDC, 2001).

Prospective studies of high school students provide strong evidence that friends'

smoking status and reinforcement are more significant than many other psychosocial risk

factors for smoking initiation and progression (Duncan et al., 1995; Flay et al., 1998;

Killen et al., 1997; Wang et al., 1997). For example, in a longitudinal study described

above (Flay et al., 1998), friends' smoking behavior significantly predicted trying,

experimenting, and regular smoking. Moreover, this relationship became stronger as the

stage of smoking progressed and as the number of friends who smoked increased.

Friend's smoking behavior also predicted movement from trying cigarettes to









experimentation for both genders, but not progression from experimentation to regular

use.

Killen and associates (1997) also examined the relationship between friends'

smoking behavior and smoking initiation among high school students. For several years,

they annually assessed these behaviors among participants in 2 consecutive cohorts of 9th

graders. Among girls and boys with no history of smoking at baseline, those with friends'

who smoked at baseline were significantly more likely to try smoking over the study's

duration. The researchers concluded that friends' smoking "was the most important

predictor of smoking" (p. 1013) for both genders when compared to other factors

including temperament, depression, and drive for thinness.

In another study of secondary school students, smoking initiation and progression

were assessed over a 3-year time span (Wang et al., 1997). The relationship between

parents', siblings' and best friends' smoking status and smoking initiation and

progression was examined. Best friends' smoking status proved to be the only consistent

and significant predictor of smoking initiation and progression for males and females.

A 4-year prospective study of five age cohorts who began the study between the

ages of 11 and 15 years olds also found evidence that peer reinforcement was predictive

of cigarette use. Results indicated that encouragement to smoke by a best friend and lack

of family cohesion predicted initial cigarette use. Encouragement-but not family

cohesion-also predicted progression to more advanced stages of smoking (Duncan et

al., 1995).

In contrast to the numerous prospective studies that examine friends' smoking

behavior and smoking initiation among high school students, only a limited amount of









cross-sectional data link friends' smoking behavior to smoking initiation during college.

In a study of African-American students, current residence, parents and friends who

smoked during the participant's childhood, and current friends who smoked were

significantly correlated with smoking at least one cigarette but less than one hundred total

in one's lifetime (Hestick, Perrino, Rhodes, & Sydnor, 2001). When these variables were

subsequently entered into a prediction model, however, only parents and friends who

smoked during the participant's childhood proved to be statistically significant in

predicting trial smoking behavior in college. Current friends' smoking status was not a

significant predictor of trial smoking in this study, however, it did predict lifetime

smoking behavior-defined as smoking 100 or more cigarettes in one's lifetime.

Furthermore, the risk of lifetime smoking was reduced when neither friends nor parents

smoked and the student viewed spirituality as important.

An email assessment of undergraduate students at Yale University also provided a

limited amount of cross-sectional data on friends' behavior and smoking status

(Debernardo et al., 1999). Nonsmoking students who were susceptible to start smoking

within a year after the survey were asked about motivations for smoking. Approximately

11% indicated that they would be motivated to begin smoking because of friends who

smoked. The most frequently selected motivator for future smoking initiation behavior

was friends' behavior (11%), followed by stress (6.5%), depression (2.7%), image that

smokers project (2.5%), and expression of independence (1.8%). Obviously, these

descriptive data cannot assess whether having friends' who smoke actually increased the

risk of smoking among their susceptible friends.









Sociability

Sociability is defined as the tendency to prefer the presence of others to being alone

and seeking and being especially gratified by the presence, attention, and sharing of

activities (Buss & Plomin, 1986). Psychologists generally agree that sociability is one

component of extroversion (Morris, 1979; Sipps & Alexander, 1987). Extroversion is

defined as the preference for attending to the outer world of events by seeking active

stimulation and involvement in the environment (Morris, 1979). Extroverts inherently

"value, enjoy, and engage in a wide range of social and affiliative activities and [are]

socially confident" (Morris, 1979, p. 89). Not surprisingly, similar items that measure

both constructs are included in the International Personality Item Pool (2001).

Compared to the amount of research on the relationship between friends' behavior

and smoking initiation, extroversion-and its related construct of sociability-has been

studied less extensively. Yet overall, a limited number of longitudinal and cross-

sectional studies have found an association between sociability (or extroversion) and

smoking initiation for both adolescents and college students. As early as the 1970s, for

example, researchers of a large, longitudinal study found extroversion to be positively

associated with smoking initiation for both sexes (Cherry & Kieman, 1976). In a

cross-sectional study of undergraduate students conducted in the 1980s, findings were

similar-smokers were significantly more likely than nonsmokers to be extroverts

(Spielberger & Jacobs, 1982).

More recently, sociability was examined as one of a variety of potential predictors

of smoking initiation among high school students (Killen et al., 1997). In a 4-year

prospective study, researchers collected data from two consecutive cohorts of 9th graders









to determine whether sociability influenced smoking initiation. They found that

sociability predicted smoking initiation among the girls only in both cohorts.

A longitudinal study of Australian adolescent twins examined the relationship of

extroversion and other personality factors and smoking initiation among 1,400 twins who

were 11 to 18 years old at baseline (White, Hill, & Hopper, 1996). Three years after the

baseline assessment, the researchers categorized each participant as a nonsmoker, an

experimental smoker, or a recent smoker. A nonsmoker had never smoked a whole

cigarette. An experimental smoker had not smoked in the month preceding the survey,

but had smoked at least 10 cigarettes in his or her lifetime. A recent smoker had smoked

during the month preceding the survey and had smoked at least 10 cigarettes in his or her

lifetime. Results indicated that extroversion, but not self-esteem, psychoticism, or

neuroticism, consistently distinguished nonsmokers who remained smoke-free from

nonsmokers who started smoking. Furthermore, researchers found a positive relationship

at the end of the three-year study between extroversion and smoking both among the

experimental smokers and the recent smokers.

Conclusion

All three social variables in this study have been linked to smoking initiation

among adolescents. Of these three explanatory variables, friend's smoking-related

behavior has the strongest empirical support, followed by perceived prevalence of peers'

smoking behavior, and then sociability. These conclusions are based overwhelmingly on

research on adolescent populations because rigorous research on the social risk factors of

smoking initiation among college students is lacking.









Summary

Chapter 2 began by providing a context of cigarette smoking among college

students. Smoking prevalence rates, trends, characteristics of college student smokers,

reasons for smoking, and smoking initiation rates were provided. Following this

overview, two conceptual frameworks were described-one for the outcome variable and

one for all other variables in the study. First, stage models of cigarette smoking were

described. Then, social cognitive theory was presented to explain the relationship

between the study's explanatory variables and smoking initiation. Finally, evidence to

support each of the three explanatory variables was provided. A comprehensive literature

review confirmed that a considerable amount of empirical support exists between each of

these three variables and smoking initiation among high school populations. Conversely,

little is known about these variables and smoking initiation among college students.

Therefore, the field of college health will benefit from research that examines the

influence of each of these social factors on smoking initiation.














CHAPTER 3
METHODS

This study determined smoking the percentage of college students who began

smoking in college, identified at-risk populations, and pinpointed when college students

are most vulnerable to begin smoking. The effects between three social factors and

smoking initiation during freshman and sophomore years of college were also examined

This chapter describes the methodology used in the study and is organized into five

sections: (1) research design, (2) research variables, (3) instrument development, (4)

pilot administration, and (5) final administration.

Research Design

This study used a cross-sectional survey design. Participants in intact classes at

five universities completed a survey on social factors influencing smoking initiation. The

selected classes were comprised of primarily first- and second-year undergraduate

students representing a variety of academic majors. Participants were traditional-age

college students. They began college by the age of 19 years and were between the ags

of 18 and 22 years old at the time of the survey.

Research Variables

This study examined the relationship between three explanatory variables and one

outcome variable. The three explanatory variables included: (1) perceived prevalence of

peers' smoking behavior, (2) smoking-related behavior of friends and roommates, and (3)

sociability. Smoking initiation during college served as the outcome variable. Smoking









initiation was defined as the first time that a participant smoked a whole cigarette on two

or more days during a month.

Instrument Development

A 77-item College Student Smoking Survey was developed specifically for this

study. The survey was designed to assess the relationship between the three explanatory

variables and smoking initiation and to provide descriptive data that could inform

decision-making and program development on college campuses. Procedures outlined by

Alreck and Settle (1995) were followed to develop and select items and to conduct an

expert review as well as student review. These three steps of the development process

are described in the following subsections.

Item Selection and Development

First, a comprehensive review of current surveys on tobacco use among adolescents

and college students was conducted. Items measuring demographics, frequency and

quantity smoked, alcohol consumption, and susceptibility were selected or adapted from

a variety of sources (ACHA, 2002; CDC, 1997; Johnston et al., 2001; Pierce et al., 1996;

SAMSHA, 2002). Next, possible items for the sociability scale were located and selected

from the International Personality Item Pool (2001). Finally, the principal researcher

created additional items to measure the outcome variable as well as the three explanatory

variables. Potential items were categorized into nine sections: (1) Cigarette Smoking

Initiation, (2) Susceptibility, (3) Social Context of First Ever Cigarette, (4) Social Norms,

(5) Sociability, (6) Frequency of Going Out, (7) Smoking-Related Behavior of Friends

and Roommates During Freshman Year, (8) Control Variables, and (9) Freshman Year

and Current Smoking Behavior.









Expert Review

Six experts in cigarette smoking, health education, and college health reviewed

potential items. They were instructed to rate the relevance of each item, select the best

items within each section, recommend additional items, and make suggestions. Based on

the expert review data, two versions of a pilot survey were developed and reviewed by

nine undergraduate students. Appendix A contains the names of the six reviewers.

Student Evaluation

Students at the University of Florida helped in the preliminary stages of survey

development in two ways-they provided qualitative definitions of various levels of

intoxication, and they reviewed two similar versions of the pilot survey. Seventy-six

male and female students from two sections of a Personal and Family Health class were

asked to "define moderately intoxicated and extremely intoxicated in their own words."

Their responses were tallied and definitions were created based on their feedback.

Furthermore, nine students reviewed two similar versions of the pilot instrument. To

keep the survey to a length that could be completed in less than 15 minutes, two versions

were necessary. Both versions contained many identical items including those that

measured demographics, smoking behavior, grade point average, living situation, and

parents' smoking status. However, items on each of the following four scales differed by

version: sociability, perceptions of peers' smoking behavior, friends and roommates'

behavior, and susceptibility to smoke.

Freshman and sophomore reviewers were recruited from a fall 2002 Personal and

Family Health class at the University of Florida, and ultimately, nine females reviewed

the surveys. Five students reviewed version one of the instrument; and four students

reviewed version two. The researcher instructed participants to record how long it took









them to complete the survey, to identify problems and suggest solutions, and to offer

general comments on how to improve the survey. Based on these results, a final version

of the survey was created and then pilot-tested on undergraduate students.

Pilot Study

Setting

The pilot study was conducted using 20 intact classes at the University of Florida in

fall 2002. Participants were enrolled in five sections of Personal and Family Health, one

section of Human Physiology, or 14 lab sections of Human Anatomy. Classes were

comprised of male and female students from a variety of academic majors.

Participants

Females comprised 77.9% of the pilot sample. On average, participants were 17.9

years old when they began their freshman year of college, and 19.3 years old at the time

of the survey. Nearly 90% lived with other individuals-besides their parents-during

the majority of their freshman year. Only 5.2% lived alone during that time. The

percentages of Asian, Black, Hispanic, and White participants were 9.3%, 11.3%, 9.6%,

and 67.3%, respectively. Approximately 2% of the participants indicated that they were

American Indian, Alaskan Native, or Biracial.

Data Collection Procedures

The University of Florida's Institutional Review Board approved the data

collection procedures for the pilot study (see Appendix B). The principal researcher

collected data in each of the 20 classes. Participants received an informed consent letter,

a survey booklet, a scantron form, and a pencil. The researcher read a script explaining

the study to the students and their rights as a potential research participant. Participation

was voluntary and anonymous. To ensure anonymity, signed consent forms were









collected separately from survey responses (see Appendix C). Participants recorded their

responses on a scantron form, and most completed the survey in 10 to 15 minutes.

Data Analysis

Data from the pilot study were scanned and electronically transmitted onto a

computer disk at the University of Florida's Office of Academic Technology. Raw data

from 733 participants were loaded into SPSS v. 10.0. A total of 688 participants met the

following inclusion criteria for the pilot study: (1) they began college at the age of 17, 18

or 19 years, and (2) they were between the ages of 18 and 22 years old at the time of the

survey, and (3) they were enrolled full-time during their freshman year and at the time of

the pilot study.

Instrument Reliability

Item and scale analyses were conducted to assess the survey's four major scales.

All items on the Perceptions of Peers' Smoking Behavior, the Smoking-Related Behavior

of Friends and Roommates, and the Susceptibility scales had acceptable item

discrimination. In contrast, nine of the 13 items on the Sociability scale had acceptable

item discrimination, and therefore, four weaker items were subsequently removed from

this scale.

The reliability of each scale was then calculated using Cronbach's measure of

internal consistency. All scales measuring the three explanatory variables (Perceived

Prevalence of Peers' Smoking Behavior, Smoking-Related Behavior of Friends and

Roommates, and Sociability) had similar reliability coefficients of .84, .84, and .87,

respectively. Furthermore, the Susceptibility scale had a high reliability coefficient of

.90.









Final Survey Administration


Final Survey Format

The final survey was developed using Remark OMR version 5.5 software

developed by Principia Products. The final survey's format was changed after the pilot

study for four reasons. First, the Remark OMR software became available to the

researcher after the pilot study. Second, the software stored all scanned surveys as an

electronic image. Third, participants recorded their responses directly on the survey and

could use ink rather than only pencil. Consequently, scantron forms and pencils did not

need to be distributed for the final study. This shortened the survey administration time

by five minutes when compared to the pilot survey. Fourth, participant error was reduced

because responses were recorded directly on the survey as opposed to on a scantron form.

Final Survey Description

The College Student Smoking Survey is a 77-item survey that was typically

completed in 10 minutes or less (See Appendix D). Forty-three of the items comprised

one of the survey's five scales. Table 3-1 lists the type of variable measured, the number

of items, the number of response choices, and the reliability coefficient for each scale.

More detailed results of the item and scale analyses are located in Appendix E.

Table 3-1. Instrument Specification Table
Scale Variable Type of Item Response Reliability
Variable Number Number Coefficient
Perceived prevalence of Explanatory 9 5 .880
peers' smoking
behavior
Smoking-related Explanatory 12 4 .916
behavior of friends
and roommates
Sociability Explanatory 9 4 .836
Susceptibility to smoking Covariate 10 4 .953
Alcohol consumption Covariate 4 6 .947









The remaining 35 items measured demographics, control variables, other potential

covariates (besides susceptibility and alcohol consumption), smoking initiation behaviors,

and 30-day prevalence of smoking. The final item on the survey solicited qualitative

comments from the participants.

Setting

Undergraduate students from Ball State University, Texas A & M University, the

University of Alabama, the University of Florida, and the University of Georgia

participated in the final study. These five universities were selected for several reasons:

(1) each had at least 15,000 undergraduate students enrolled during the 2002-2003

academic year, (2) each was able to provide a faculty liaison to coordinate data

collection, and (3) each is located in a state that does not rely on tobacco revenue for its

economic stability.

Participants

Male and female undergraduate students in intact classes participated in the final

study. Classes were chosen based on two criteria: (1) they contained primarily first- and

second-year undergraduate students, (2) they were comprised of students from a variety

of academic majors. An elective introductory health class was chosen at four out of the

five universities. A political science class, required of all undergraduate students, was

selected at Texas A & M University.

Between 200 and 600 students at each university completed a survey for an initial

sample size of 1,775. Five inclusion criteria were applied to the initial data set: (1)

participants started college when they were 17, 18 or 19 years old, (2) participants were

18 to 22 years old at the time of the survey, (3) participants began college within 36

months of the survey, (4) participants were enrolled full-time during their freshman year









of college and at the time of the survey, and (5) participants provided sufficient data to

determine their smoking initiation status. A total of 525 surveys failed to meet all five

inclusion criteria yielding a final sample size of 1,250. The amount of excluded surveys

was not surprising because data was collected in intact classes that obviously contained

students who were over 22 years old at the time of the study.

Protection of Confidentiality

Each of the five university's Institutional Review Board approved the final study.

All participants received a description of the study and information about their rights as

participants. Participation was voluntary and anonymous. Written consent was obtained

from participants at Ball State University, the University of Alabama, the University of

Florida, and the University of Georgia. Participants at Texas A & M University were not

required to provide a written consent form.

Data Collection Procedures

Data were collected in selected classes at five universities between January 6 and

February 6, 2003. The researcher collected data at the University of Florida. She

recruited a faculty advisor at each of the other four sites (see Appendix F) and sent them

the survey materials in early January. The same data collection protocol was followed at

all five sites (see Appendix G). The faculty advisors returned the completed surveys to

the researcher by February 20, 2003.

Data Analysis

Surveys were scanned using Remark OMR version 5.5 software. Data were then

loaded into SPSS version 10.0 for analysis. To verify data entry accuracy, 75 surveys

were randomly selected and the original responses were compared to the scanned

responses. The original responses on all 75 surveys matched their scanned counterparts.









Three types of data analyses were conducted on the final data set to establish a

profile for the sample and to test the research hypotheses. First, descriptive statistics

were calculated to determine the baseline smoking initiation rates as well as the 30-day

prevalence rates of current smoking among students. They were also calculated to obtain

a general demographic profile of the sample and a specific profile of college student

smokers. Second, correlations were calculated between each of the five covariates

(susceptibility to smoke cigarettes during the upcoming year, grade point average, living

situation, alcohol consumption, and parents' smoking status) and smoking initiation to

determine which of these five to include in the final regression analysis. According to

Cohen (1977), correlations of .10 or greater represent at least moderate associations, and

therefore, this criterion was used to determine inclusion of the five covariates. Third,

multiple logistic regression was used to examine the relationship between each of the

three explanatory variables and smoking initiation. Analyses for all three research

questions were tested at a .05 significance level.

Controlling for Bias

Potential bias was controlled in three ways. First, the same data collection protocol

was followed at all five sites. Second, individual differences were controlled. More

specifically, data analyses were limited to responses from participants who (1) began

college when they were 17, 18 or 19 years old, (2) were 18-22 years of age at the time of

the survey, (3) began college within 36 months of completing the survey, and (4) were

enrolled full-time during their first year of college at the time of survey. Third, data was

collected to measure five potential covariates. Those with a .10 or greater association

with the outcome variable were included as covariates in the multiple logistic regression

analyses.









Summary

This chapter described the research design, research variables, instrument

development, and pilot and final studies' methodologies. A cross-sectional survey

design was used for this study. The College Student Smoking Survey was developed and

pilot tested according to recommended theory and practice (Alreck & Settle, 1995).

Findings indicated that the survey was reliable.

Data were collected in intact classes during January and February of 2003. A total

of 1,775 students from five universities participated in the final study, however, after five

inclusion criteria were applied, 1,250 participants remained. The majority of participants

who were excluded from the study did not meet the criterion of being a traditional

undergraduate student or were over the age of 22 years.

Descriptive statistics were calculated to identify sample demographics and

characteristics of college smoking behavior, including the percentage of participants who

started smoking during college. Furthermore, smoking prevalence rates were calculated.

Covariates with moderate associations to smoking initiation were identified for inclusion

in the multiple logistic regression analyses to test the three research hypotheses. The

results of the final study are reported in Chapter 4.














CHAPTER 4
RESULTS

This study determined the percentage of college students who began smoking

cigarettes in college, identified at-risk populations, and pinpointed when college students

are most vulnerable to begin smoking. The effects between three social factors and

smoking initiation during freshman and sophomore years were also examined. This

chapter presents the results of the study and is organized into five major sections. The

first three sections provide descriptive analyses of the sample, current smoking behavior,

and smoking initiation behavior. The fourth section describes the associations between

the suspected covariates and smoking initiation, and the fifth section presents the results

for the three research questions.

Sample Demographics and Characteristics

This section describes demographics and characteristics of the sample. A total of

1,772 male and female undergraduate students from five universities participated in the

study. Five criteria were selected to ensure that participation was limited to only full-

time, traditional-aged freshman, sophomore, and junior students. Seniors were excluded

because of the concern for the accuracy of recalling information from their freshman year

of college. After applying the inclusion criteria to the data set, 1,250 surveys remained,

yielding a 70% inclusion rate. Approximately 60% of those excluded were seniors, and

another 10% had failed to indicate how many months they had been in college. The

remaining 30% of those who were excluded were either not traditional-aged students or

failed to provide sufficient data to determine their smoking initiation status.









Participation by University, Gender, and Race

Table 4-1 provides a summary of participants by university, gender, and race. Of

the 1,250 participants, one out of three attended the University of Florida (37.5%).

Nearly 21% were enrolled at Ball State University, followed by 15.8% from Texas

A & M University, 14.1% from the University of Georgia, and 11.6% from the

University of Alabama. Two out of three participants were female (68%), and nearly 8

out of 10 were White (78.7%).

Table 4-1. Distribution of Participants by University, Gender, and Race
f %
University
Ball State University 261 20.9
University of Alabama 145 11.6
University of Florida 467 37.5
University of Georgia 175 14.1
Texas A & M University 197 15.8
Total 1,245
Gender
Male 398 32.0
Female 845 68.0
Total 1,243
Race
African American 117 9.4
American Native 1 0.1
Asian 45 3.6
Latino 67 5.4
Pacific Islander 3 .2
White 978 78.7
Multi-racial 32 2.6
Total 1,243
Note. University affiliation was missing for 5 participants. Gender
and race were missing for 7 participants.


Participation by Age and Months in College

Table 4-2 shows the age of the participants at the time of the study and also the

number of months since they began their freshman year. Approximately 37% of the









participants were 19 years old. The mean age of the sample was 19.25 years. One out of

four were 18 years old, and another 25% were 20 years old. Only 10% of participants

were 21 or 22 years old at the time of the study. Freshmen comprised the largest

percentage of the sample (46.2%), followed by sophomores (32.5%), and then juniors

(21.4%).

Table 4-2. Distribution of Participants by Age and Months in College
f % Cum %
Age
18 319 25.5 25.5
19 464 37.1 62.6
20 324 25.9 88.6
21 123 9.8 98.4
22 20 1.6 100.0
Total 1,250 100.0
Months since beginning
college
<6 447 35.8 35.8
7- 12 130 10.4 46.2
13 18 261 20.9 67.0
19-24 145 11.6 78.6
25-30 160 12.8 91.4
31 36 107 8.6 100.0
Total 1,250 100.0

Current Smoking Behavior

This section describes how many participants smoked, how much, and how often

during the 30 days preceding the study. National surveys typically use a dichotomous

question to measure current smoking behavior: "Have you used (or smoked) cigarettes in

the past 30 days?" In this study, participants were given four response choices to

describe their smoking behavior during the 30 days preceding the survey: (1) never

smoked, not even a puff or two, (2) tried a cigarette, but only a puff or two, (3) took puffs

from cigarettes on occasion, but didn't smoke a whole cigarette at one sitting, or (4)









smoked one or more whole cigarettes. Table 4-3 presents these results and distinguishes

between three groups of participants: never users, pufferss," and those who smoked at

least one or more whole cigarettes during the 30 days before the survey.

Table 4-3. Past 30-Day Smoking Prevalence Rates
f %
Never smoked, not even a puff or two 916 73.5
Smoked, but only took puffs 75 6.3
Smoked one or more whole cigarettes 255 20.5
Total 1,246
Note. 30-day smoking behavior was missing for 4 participants.

Nearly three out of four participants (73.5%) did not smoke in the 30-days before

the survey. In contrast, a total of 26.8% reported some cigarette use during that time

period. Among these cigarette users, 6.3% had only puffed on one or more cigarettes

during that time, whereas, 20.5% had smoked one or more whole cigarettes.

Smoking Behavior by Gender and Race

Table 4-4 shows 30-day smoking prevalence rates by gender and race. Nearly

28% of the males and approximately 25.% of the females had smoked in the 30 days

before the survey. The same percentage of males (20.1%) as females (20.2%) smoked

one or more whole cigarettes, however, slightly more males (7.8%) than females (5.1%)

only puffed during that time. Pacific Islanders had the highest prevalence rate among the

different racial groups (66.7%), however, these results must be interpreted with caution

because only three Pacific Islanders participated in the study. Approximately one of out

four White participants (28.7%), and a similar number of multiracial participants

(28.1%), had smoked during the month before the survey. The prevalence rate for

Latinos was 22.4%, for Asians, 19.9%, and for African Americans, 12%.









Smoking rates for participants who only puffed on cigarettes during the 30 days

before the study also differed among the various racial groups. For example, Latino

participants had the highest rate of puffing on cigarettes during the 30 days before the

study (12.0%), followed by White (6.0%), African American (6.0%), and Asian

participants (2.2%). No Pacific Islander or multi-racial participant reported puffing on

cigarettes during the month before the study.

Table 4-4. Past 30-Day Smoking Prevalence Rates by Gender and Race
Smoked < 1 Smoked > 1 Any cigarette use
cigarette cigarettes
f % f % f %
Gender
Male 31 7.8 80 20.1 111 27.9
Female 43 5.1 171 20.2 214 25.3
Missing 1 4
Total 75 255
Race
African 7 6.0 7 6.0 14 12.0
American
American Native 0 0.0 0 0 0 0
Asian 1 2.2 8 17.7 9 19.9
Latino 8 12.0 7 10.4 15 22.4
Pacific Islander 0 0.0 2 66.7 2 66.7
White 59 6.0 222 22.7 281 28.7
Multi-racial 0 0.0 9 28.1 9 28.1
Total 75 255


Frequency and Quantity Smoked

This section describes the behavior of the participants who smoked one or more

whole cigarettes during the 30 days before the study. This group is referred to as "current

smokers." Table 4-5 shows how often and how much the current smokers smoked during

the month before the study. The majority of current smokers reported smoking on 20 or

fewer days (64%), and 33.6% smoked on five or fewer days during the month prior to the

survey. Only 20.2% had smoked on all 30 days prior to the survey.









Current smokers were also asked how many cigarettes, on average, they smoked on

the days when they smoked. Three out of four reported smoking five or fewer cigarettes

on the days that they smoked-22.4% had smoked one cigarette, and 51.6% had smoked

two to five cigarettes, on the days that they smoked. Only 10% reported smoking

between 11 and 20 cigarettes, and only one out of the 255 current smokers in the sample

reported smoking more than a pack a day when he or she smoked.

Table 4-5. Number of Days and Quantity Smoked
f % Cum %
# of days
1 19 7.5 7.5
2 29 11.5 19.0
3-5 37 14.6 33.6
6-10 34 13.0 46.6
11-20 44 17.4 64.0
21-29 40 15.8 79.8
All 30 51 20.2 100.0
Total 253

# of cigarettes
1 56 22.4 22.4
2-5 129 51.6 74.0
6-10 40 16.0 90.0
11-20 24 9.6 99.6
21 or more 1 0.40 100.0
Total 250
Note. Results are reported for participants who smoked one
or more whole cigarettes during the 30 days prior to the survey
and represent their smoking behavior during that month. The
number of days was missing from 2 participants, and the
number of cigarettes smoked was missing from 5 participants.

Smoking Initiation Behavior

This section describes smoking initiation behavior among all participants as well

as subpopulations within the sample. Findings showing the age of when participants first

smoked a whole cigarette are presented first. Then, the results of four types of smoking

initiation behavior are presented including the first time that participants had (1) smoked

whole cigarette, (2) smoked on 2 days during any one month, (3) smoked on 20 days in









any one month, and (4) smoked on 30 consecutive days. These results are presented by

time in school as well as year in school.


First Ever Smoking Behavior by Age

Table 4-6 shows how old participants were when they first smoked a whole

cigarette. Of the 1,250 participants, 32.7% had smoked their first whole cigarette by the

time they were 17 years old, and another 7.8% had smoked their first whole cigarette

when they were 18 years old. Collectively, only 3.5% of the participants had smoked

their first whole cigarette after the age of 18 years.

Table 4-6. First Ever Smoking Behavior by Age for Sample and Current Smokers
Sample Current Smokers
Smoked first whole f % Cum % f % Cum %
cigarette
17 or younger 408 32.7 32.7 185 72.8 72.8
18 98 7.8 40.5 44 17.3 90.1
19 30 2.4 42.9 17 6.7 96.8
20 12 1.0 43.9 7 2.8 99.6
21 1 .1 44.0 1 0.4 100.0
Never 700 56.0 100.0 -
Total 1,249 100.0 254 100.0
Note. Data was missing for 1 participant and 1 current smoker. Current smokers
are participants who smoked one or more whole cigarettes during the 30 days
before the survey (n = 255).


Table 4-6 also shows how old current smokers were when they first smoked a

whole cigarette. Current smokers were defined as participants who had smoked one or

more whole cigarettes during the 30 days before the survey. Among this group of 255

participants, 72.8% had first smoked a whole cigarette by the time they were 17 years

old. A total of 17.3% had smoked their first whole cigarette at the age of 18 years, and

6.7% had first smoked a whole cigarette when they were 19 years old. Collectively, only

3.2% of these current smokers had first smoked a whole cigarette when they were 20 or









21 years old. Moreover, no current smoker reported first smoking a whole cigarette after

the age of 21 years.

First Ever Smoking Behavior by Time in College

One purpose of this study was to pinpoint when smoking initiation occurred during

college. Table 4-7 shows these results for the sample, current smokers, and "ever"

smokers. Ever smokers were defined as participants who had ever engaged in each of the

four smoking initiation behaviors listed in the table. Among the sample, 43.8% had ever

smoked a whole cigarette in their lifetime and 36.5% had done this prior to college. Only

7.3% had first smoked a whole cigarette during college. A total of 27% of the sample

had first smoked on two or more days in any one month before they came to college,

whereas, only 9.3% had engaged in this behavior after they came to college.

Results are also presented for more frequent smoking. A total of 16.4% of the

participants had ever smoked on 20 days in any one month with 11.2% first smoking this

frequently prior to college, and 5.2% starting some time during college. Finally, 13.5%

of the sample reported ever smoking on 30 consecutive days. Just over 9% had first

smoked this frequently prior to college, whereas, 4.3% had first smoked on 30

consecutive days during college.

In addition to showing first ever smoking behavior for all participants in the

sample, Table 4-7 also presents this information for current smokers. Over 84% of

current smokers had first smoked a whole cigarette prior to college. Approximately 70%

had first smoked on two days in any month; 42.5% had first smoked on 20 days during

any one month; and 36.1% had first smoked on 30 consecutive days prior to college.

Results also show that 3.5% of current smokers had never smoked on 2 days in one













Table 4-7. First Ever Smoking Behavior by Time in College for Sample, Current Smokers, and Ever Smokers
Whole cigarette 2 days/month 20 days/month 30 consecutive days
f % Cum % f % Cum % f % Cum % f % Cum %


Sample
First 6 months
Second 6 months
Second year
After 2nd year
Prior to college
Never
Total

Current Smokers
First 6 months
Second 6 months
Second year
After 2nd year
Prior to college
Never
Total

Ever Smokers
First 6 months
Second 6 months
Second year
After 2nd year
Prior to college
Total


55
10
24
2
456
702
1249


4.4
0.8
1.9
0.2
36.5
56.2
100.0


8.3
1.6
5.5
0.4
84.2


254 100.0


10.0
1.8
4.4
0.4
83.4


4.4
5.2
7.1
7.3
43.8
100.0


8.3
9.9
15.4
15.8
100.0


66
16
22
12
337
797
1250


35
10
14
7
179
9
254


66
16
22
12
337
453


5.3
1.3
1.8
0.9
27.0
63.7
100.0


13.8
3.9
5.5
2.8
70.5
3.5
100.0


14.6
3.5
4.9
2.6
74.4


5.3
6.6
8.4
9.3
36.3
100.0


13.8
17.7
23.2
26.0
96.5
100.0



14.6
18.1
23.0
25.6
100.0


39
7
15
4
140
1044
1249


32
5
13
2
108
94
254


39
7
15
4
160
235


3.1
0.6
1.2
0.3
11.2
83.6
100.0


12.6
2.0
5.1
0.8
42.5
37.0
100.0


19.1
3.4
7.4
2.0
68.1


3.1
3.7
4.9
5.2
16.4
100.0


12.6
14.6
19.7
20.5
63.0
100.0



19.1
22.5
29.9
31.9
100.0


27
8
16
3
114
1079
1247


23
7
13
2
91
116
254


27
8
16
3
114
168


2.2
0.6
1.3
0.2
9.2
86.5
100.0


9.1
2.8
5.2
0.8
36.1
46.0
100.0


16.1
4.7
9.5
1.8
67.9


2.2
2.8
4.1
4.3
13.5
100.0


9.1
11.9
17.1
17.9
54.0
100.0


16.1
20.8
30.3
32.1
100.0


10.0
11.8
16.2
16.6
100.0


Note. Current smokers smoked one or more cigarettes during the 30 days preceding the survey. Ever smokers excluded participants
who reported "never" within each category of smoking behavior.









month; 37.0% had never smoked on 20 out of 30 days; and 46% had never smoked on 30

consecutive days.

Finally, Table 4-7 presents first ever smoking behavior results for "ever" smokers.

These results include only those participants who had ever engaged in that specific

smoking behavior. For example, the percentages for ever smokers under the "whole

cigarette" category were calculated from the 43.8% of the sample who indicated that they

had smoked a whole cigarette. The 56.2% of the sample who reported never smoking a

whole cigarette were excluded from these results. These descriptive results answer the

question, "Among participants who had ever smoked at the various smoking level, how

many engaged in that particular behavior prior to college, and how many engaged in that

behavior during college?"

Results show that as smoking behavior progresses, the percentage of students who

initiate that particular behavior during college increases. For example, 16.6% of

participants who had ever smoked a whole cigarette had first smoked it during college.

This percentage increases to 25.6% for those who had ever smoked on 2 days in any one

month, and to 32.1% for participants who had ever smoked on 30 consecutive days.

Another interesting way to present these smoking behavior data are by year in

school. Table 4-8 presents first ever smoking behavior by year in school for all

participants. Freshmen were defined participants who had been in college for one to 12

months. Sophomores had been enrolled for 13-24 months, and juniors had been in

college for 25-36 months. For reasons mentioned at the beginning of this chapter, seniors

were excluded from this study.














Table 4-8. First Ever Smoking Behavior by Year in School for all Participants
Whole cigarette 2 days/month 20 days/month 30 consecutive days
f % Cum % f % Cum % f % Cum % f % Cum %
Freshmen
First 6 months 22 3.8 3.8 28 4.9 4.9 18 3.1 3.1 10 1.8 1.8
Second 6 months 1 0.2 4.0 1 0.2 5.1 0 0.0 3.1 0 0.0 1.8
Prior to college 209 36.2 40.2 162 28.1 33.2 71 12.3 15.1 59 10.2 12.0
Never 344 59.8 100.0 385 66.8 100.0 487 84.6 100.0 506 88.0 100.0
Total 577 100.0 576 100.0 576 100.0 575 100.0

Sophomores
First 6 months 21 5.2 5.2 22 5.4 5.4 11 2.7 2.1 9 2.2 2.2
Second 6 months 7 1.7 6.9 10 2.5 7.9 4 1.0 3.7 5 1.2 3.4
Second year 11 2.7 9.6 15 3.7 11.6 11 2.7 6.4 9 2.2 5.6
Prior to college 142 35.0 44.6 104 25.6 37.2 36 8.9 15.3 29 7.2 12.8
Never 225 55.4 100.0 255 62.8 100.0 343 84.7 100.0 353 87.2 100.0
Total 406 100.0 406 100.0 405 100.0 405 100.0

Juniors
First 6 months 12 4.5 4.5 16 5.9 5.9 10 3.8 3.8 8 3.0 3.0
Second 6 months 2 0.7 5.2 5 1.9 7.8 3 1.1 4.9 3 1.1 4.1
Second year 13 4.8 10.0 7 2.7 10.5 4 1.5 6.4 7 2.6 6.7
After 2nd year 2 0.7 10.7 11 4.1 14.6 3 1.1 7.5 2 0.8 7.5
Prior to college 105 39.3 50.0 71 26.6 41.2 32 12.0 19.5 26 9.7 17.2
Never 133 50.0 100.0 157 58.8 100.0 214 80.5 100.0 220 82.8 100.0
Total 267 100.0 267 100.0 266 100.0 266 100.0


13-24 months. Juniors = 25-36 months. Seniors were not part of the study.


Note. Freshmen = 1-12 months in college. Sophomores











Noteworthy of these data by year in school are the percentages of students who had

never smoked at the four various levels shown in Table 4-8. The largest percentage of

participants who had never smoked within each category of smoking initiation

consistently is freshmen, followed by sophomores, and then juniors. For example, 66.8%

of freshmen had never smoked on two or more days in any month, however, this

percentage for sophomores was 62.8%, and for juniors, 58.8%.

The last two tables in this section on smoking initiation behavior focus specifically

on college-initiated smoking behavior. In other words, the results listed in Tables 4-9 and

4-10 represent only those participants who reported beginning to smoke during college.

Two groups are shown in Table 4-9 each of the four types of smoking initiation

behavior: (1) participants who had ever initiated the behavior during college, and (2)

current smokers. Table 4-10 summarizes the four types of smoking initiation behavior

by year in school.

Results in Table 4-9 indicate that the majority of college-initiated smoking

behavior among the entire sample, and among current smokers, occurred during the first

six months of college. For example, among those who had ever first smoked on 20 days

during any one month in college, 60% did so during their first six months of college.

Among current smokers, 61.5% smoked on 20 days in any one month for the first time

during their first six months of college. Furthermore, relatively little smoking initiation

behavior occurred after the sophomore year (3.9%).















Table 4-9. College-Initiated Smoking Behavior by Time in College
Whole Cigarette 2 Days/Month
f % Cum % f % Cum % f
Ever Initiators
First 6 mos. 55 60.4 60.4 66 56.9 56.9 39
Second 6 mos. 10 11.0 71.4 16 13.8 70.7 7
Second year 24 26.4 97.8 22 19.0 89.7 15
After 2n year 2 2.2 100.0 12 10.3 100.0 4
Total 91 116 65


20 Days/Month 30 Consecutive Days
% Cum % f % Cum %


60.0 60.0
10.8 70.8
23.1 93.9
6.1 100.0


27 50.0 50.0
8 14.8 64.8
16 29.6 94.4
3 5.6 100.0
54


Current Smokers
First 6 mos.
Second 6 mos.
Second year
After 2na year
Total


21 52.5 52.5 35 53.0 53.0 32
4 10.0 62.5 10 15.2 68.2 5
14 35.0 97.5 14 21.2 89.4 13
1 2.5 100.0 7 10.6 100.0 2


61.5 61.5 23
9.6 71.1 7
25.0 96.1 13
3.9 100.0 2


51.1 51.1
15.6 66.7
28.9 95.6
4.4 100.0









Another way to present these college-initiated data is by year in school. These data

help to pinpoint when the majority of students in each class initiated each type of

smoking behavior. Table 4-10 shows that over 95% of the freshmen who had ever

initiated these four smoking behaviors in college had engaged in these behaviors for the

first time during their first six months of college. Similarly, the largest percentage of

college-initiated smoking behavior among sophomores and juniors also occurred during

the first six months of college. Moreover, the data representing the juniors shows that the

majority of college smoking initiation among this group had occurred during their first

two years of college. For example, 85% of the juniors who first smoked on 20 days

during any one month, and 90% of the juniors who had ever smoked on 30 consecutive

days, engaged in these behaviors by the end of their sophomore year in college.

Correlations Between the Covariates and Smoking Initiation

This section will describe the relationship between five covariates and the

dependent variable. To briefly review, the behavior of interest for the study was not the

first time that participants tried a cigarette, or even when they first smoked a whole

cigarette; rather, the behavior of interest was when participants began to smoke

repeatedly. Therefore, the dependent variable of the study-smoking initiation-was

defined as the first time a participant smoked on 2 or more days in any one month during

their freshman or sophomore year of college. A total of 104 students, or 8.3% of the

sample, met this definition of smoking initiation. Data from this subpopulation of

students was used to assess the association between five suspected covariates

(susceptibility to smoke, grade point average, living situation, alcohol consumption, and

parents' smoking status) and smoking initiation. Data from this group was also used to

analyze each of the three research questions.












Table 4-10. College-Initiated Smoking Behavior by Year in School
Whole cigarette 2 days/month 20 days/month 30 consecutive days
f % Cum % f % Cum % f % Cum % f % Cum %


Freshmen
First 6 months
Second 6 months
Total

Sophomores
First 6 months
Second 6 months
Second year
Total

Juniors
First 6 months
Second 6 months
Second year
After 2nd year
Total


95.7
4.3
100.0


53.8
18.0
28.2
100.0


41.4
6.9
44.8
6.9
100.0


95.7
100.0


53.8
71.8
100.0


96.6
3.4


46.8
21.3
31.9
100.0


41.0
12.8
18.0
28.2


41.4
48.3
93.1
100.0


96.6
100.0


46.8
68.1
100.0


41.0
53.8
71.8
100.0


100.0
0.0
100.0


42.3
15.4
42.3
100.0


50.0
15.0
20.0
15.0


100.0
100.0


42.3
57.7
100.0


100.0
0.0
100.0


39.1
21.8
39.1
100.0


40.0
15.0
35.0
10.0


50.0
65.0
85.0
100.0


100.0
100.0


39.1
60.9
100.0


40.0
55.0
90.0
100.0


Note. Freshmen = 1-12 months in college. Sophomores


13-24 months. Juniors


25-36 months. Seniors were not part of the study.


College initiators are participants who reported engaging in smoking initiation behavior during college.









Table 4-11 shows the correlations between each of five suspected covariates and

smoking initiation. Only two covariates were moderately associated with smoking

initiation as defined by a predetermined inclusion criterion of 0.10. Susceptibility to

smoke during freshman year of college (r =.40), and alcohol consumption (r =.20) each

showed a moderate association to smoking initiation, and subsequently, both were

included in the multiple logistic regression analyses used to answer the study's three

research questions. In contrast to these two variables, the other three variables (grade

point average freshman year, living with others during freshman year, and parents

smoking status) failed to meet the inclusion criterion of 0.10. Consequently, they were

excluded from further analyses.

Table 4-11. Correlations between Five Suspected Covariates and Smoking Initiation
Variable r
Susceptibility to smoke freshman year .40
Alcohol consumption freshman year .20
Grade point average freshman year -.08
Living with others (besides parents) freshman year .04
Parents' smoking status -.02


Research Question Results

This section describes the results of the multiple logistic regression analyses

conducted to answer each of three research questions.

Research Question One

Does perceived prevalence of peers' smoking behavior during freshman year of

college increase the risk of smoking initiation during freshman or sophomore year of

college?

A multiple logistic regression analysis was conducted to examine the relationship

between perceived prevalence of peers' smoking behavior and smoking initiation during









freshman or sophomore year of college. The approximate chi-square statistic was used to

assess model fit. A p-value of greater than 0.05 indicates an adequate model fit (Hosmer

& Lemeshow, 1980). Results of the chi-square analysis confirmed an adequate fit

(?2 = 15.145, p = .056). Table 4-12 shows the results of the regression analysis. After

controlling for the effects of two covariates (susceptibility to smoke during freshman year

and alcohol consumption during freshman year), perceived prevalence of peers' smoking

behavior was not significantly related to smoking initiation during freshman or

sophomore year of college (p = .587).

Table 4-12. Multiple Logistic Regression Analysis of Perceived Peer Smoking Behavior
and Smoking Initiation
Variable b SE Wald p-value
Perceived Peer Smoking .095 .176 .295 .587
Susceptibility 1.576 .204 59.735 < .001*
Alcohol Consumption .507 .105 23.192 <.001*
* = Significant at the .05 level


Research Question Two

Does the smoking-related behavior of friends and roommates during freshman year

of college increase the risk of smoking initiation during freshman or sophomore year of

college?

A multiple logistic regression analysis was conducted to test for the significance of

smoking-related behavior of friends and roommates and smoking initiation during

freshman or sophomore year of college. Smoking-related behavior of friends and

roommates, susceptibility to smoke during freshman year, and alcohol consumption

during freshman year were included in the regression model. A chi-square analysis

confirmed an adequate model fit (?2 = 14.031, p = .051). Results indicated that the









smoking-related behavior of friends and roommates was a statistically significant factor

of smoking initiation during freshman and sophomore years of college (p = .003).

A logistic regression coefficient can be interpreted as the increase in the odds for

each unit increase in the explanatory variable. For example, the coefficient (b) for

friends' and roommates' smoking-related behavior was 0.832. This value suggests that

the odds of initiating smoking during freshman or sophomore year of college increases by

eo.32= 2.298, or more than doubles, for each unit increase in the composite score of

smoking-related behavior of friends and roommates. Table 4-13 summarizes the results

of this regression analysis.

Table 4-13. Multiple Logistic Regression Analysis of Friends' and Roommates'
Smoking-Related Behavior and Smoking Initiation
Variable b SE Wald p-value
Friends' and Roommate's .832 .281 8.740 .003*
Behavior
Susceptibility 1.380 .214 41.565 < .001*
Alcohol Consumption .421 .110 14.600 < .001*
* = Significant at the .05 level

Research Question Three

Does sociability during freshman year of college increase the risk of smoking

initiation during freshman or sophomore year of college?

A multiple logistic regression analysis was conducted to examine the relationship

between the sociability and smoking initiation during freshman or sophomore year of

college. The approximate chi-square statistic was used to assess the fit of the model.

Results indicated an adequate model fit (.05 (?2 = 11.928, p = .154). In addition to

sociability, susceptibility and alcohol consumption were also included in the model as

covariates. Table 4-14 shows that sociability failed to be a statistically significant factor

of smoking initiation (p = .411).









Table 4-14. Multiple Logistic Regression Analysis of Sociability and Smoking Initiation
Variable b SE Wald p-value
Sociability .801 .975 .675 .411
Susceptibility 1.613 .202 63.666 <.001*
Alcohol Consumption .449 .131 11.764 .001*
* = Significant at the .05 level

Regression Analysis of All Three Explanatory Variables

Finally, a multiple logistic regression analysis was conducted to determine the

relationship between all three explanatory variables and smoking initiation. To assess the

fit of the model, the approximate chi-square statistic was used. The model fit the data as

indicated by ap-value of greater than .05 (?2 = 10.671, p = .221). Susceptibility and

alcohol consumption during freshman year were included as covariates in the model.

The smoking behavior of friends and roommates proved to be the only statistically

significant explanatory variable (p = .003). A logistic regression coefficient can be

interpreted as the increase in the odds for each unit increase in the explanatory variable.

The coefficient (b) for friends' and roommates' smoking-related behavior in this model

was 0.842. This value suggests that the odds of initiating smoking during freshman or

sophomore year of college increases by e0.842= 2.321, or more than doubles, for each unit

increase in the composite score of smoking-related behavior of friends and roommates.

Table 4-15 summarizes the results of this regression analysis.

Table 4-15. Multiple Logistic Regression Analysis of Three Explanatory Variables of
Smoking Initiation
Variable b SE Wald p-value
Perceived Prevalence of Peer Smoking .100 .177 .320 .572
Friends' and Roommates' Behavior .842 .283 8.855 .003*
Sociability .808 .981 .678 .410
Susceptibility 1.363 .216 39.780 < .001*
Alcohol Consumption .350 .135 6.659 <.001*
* = Significant at the .05 level









Summary

This chapter provided relevant descriptive results as well as findings for each of the

three research questions. A profile of the sample and description of smoking behavior

was presented, including 30-day smoking prevalence and the percentage of college

students who began smoking in college. Correlations between each of five suspected

covariates and smoking initiation were calculated using a predetermined inclusion

criterion of 0.10. Only two of these five variables-susceptibility to smoke during

freshman year and alcohol consumption during freshman year-met this inclusion

criterion. Consequently, these variables were included as covariates in all regression

analyses. A separate regression analysis was conducted to answer each of the three

research questions. A final regression analysis was conducted that included all three

explanatory variables and the two covariates. A discussion of the results will be provided

in Chapter 5.














CHAPTER 5
DISCUSSION

This chapter has four purposes: (1) to discuss the study's results including current

smoking behavior, smoking initiation behavior, and findings of the research questions,

(2) to describe the study's strengths and weaknesses, (3) to present conclusions, and,

(4) to offer recommendations for future research and practice.

Current Smoking Behavior

Consistent with national surveys of college students, current smoking behavior was

measured by asking participants whether they had smoked at any time during the 30 days

before the survey. The 30-day smoking prevalence rate of 26.8% found in this study was

consistent with the rates found in other national surveys (ACHA, 2002, Johnston et al.,

2001, SAMHSA, 2002). Also consistent were findings that indicated similar smoking

rates for males and females, and higher smoking rates for Whites, as opposed to Latinos

or African American students.

Unlike national surveys, however, this study distinguished between the number of

participants who just puffed on cigarettes during the 30 days before the study, and those

who had smoked at least one whole cigarette during that time. Currently, national

prevalence rates include individuals who have only puffed on cigarettes in the past 30

days. These individuals are typically labeled "current smokers." Results from this study

indicated that 6.3% of those who smoked on the 30 days before the study were pufferss,"

whereas, 20.5% had smoked one or more cigarettes in the month prior to the survey.









These data provide a more precise description of 30-day smoking behavior and are useful

to practitioners who deliver smoking interventions.

This study also determined the frequency of smoking and the quantity of cigarettes

smoked, on average, on the days when smoking occurred. Participants who smoked one

or more whole cigarettes during the previous 30 days were included in these analyses.

Findings indicated that the majority of smokers smoked on relatively few days, and

smoked relatively few cigarettes. For example, one out of five students smoked on one

or two days during the month before the study, and close to half had smoked on ten or

fewer days. In contrast, only one out of five smokers smoked daily. This daily rate is

fairly consistent with recent Monitoring the Future reports showing that 19% of college

smokers smoked daily in 1999, and 17.8% smoked daily in 2000.

This study also determined the quantity of cigarettes smoked, on average, on the

days that participants smoked. One out of four smokers smoked only one cigarette on the

days that they smoked, and three out of four smoked five or fewer cigarettes. Moreover,

only one out often students smoked 11-20 cigarettes, and only one participant out of the

255 smokers in the study reported smoking more than a pack on the days that he or she

smoked.

Collectively, the findings on the frequency of smoking and the quantity smoked

suggest that many college students smoke moderately and occasionally. Therefore,

smoking cessation efforts on college campuses should not only target heavy, frequent

smokers, but should also target occasional and moderate smokers. What motivates this

latter population to quit may be inherently different from what motivates the more

frequent, heavy smokers to quit. For example, using a financial appeal may not be an









effective approach with these moderate smokers, because many may only spend around

$10.00 a month on cigarettes. Consequently, college health practitioners should find

other, more suitable, ways to appeal to these moderate smokers if they intend to motivate

them to quit.

Smoking Initiation Results

Currently, 25-30% of college students in that nation smoke. This is more than twice

the Healthy Campus 2010 goal of 12%. Reducing college smoking rates by the year

2010 can be accomplished in one of two ways: (1) either nonsmoking college students

must remain smoke-free, or (2) college students who already smoke must quit. This

study focused on helping nonsmokers remain smoke-free by investigating smoking

initiation among college students. Prior to this study, data were limited that addressed

how many college students start smoking in college, when college students are most

vulnerable to start smoking, and why college students start smoking. This information is

essential not only for the development of effective smoking prevention programs, but

also for the wise allocation of limited health promotion resources.

How Many Students Start Smoking in College?

Findings from this study are consistent with previous studies showing that the

majority of smoking initiation among college students occurs by the age of 18 years.

Among current smokers in this study, 90.1% had smoked their first whole cigarette when

they were 18 years old or younger, and 9.9% had smoked their first cigarette when they

were 19 years old or older. Wechsler and colleagues (1998) reported similar rates among

the current smokers in their sample. They reported that 89% had smoked a whole

cigarette for the first time at the age of 18 or younger, and 11% smoked their first

cigarette at the age of 19 years or older.









Collecting data by age, however, fails to address how many college students started

smoking in college because many college students begin college when they are 18 years

old. Findings from this study indicated that 7.3% of the sample had smoked their first

whole cigarette during college. Moreover, 9.3% had first smoked on two or more days in

any one month; 5.2% had first smoked on 20 or more days in any one month; and 4.3%

had first smoked on 30 consecutive days during college. These data are consistent with a

study done at Yale University (Debernardo et al., 1999) showing a 5.3% smoking

initiation rate among college students. Collectively, these data clearly indicate that fewer

than 10% of college students begin smoking during college.

Does this mean that we should abandon smoking prevention efforts on our college

campuses? Not necessarily-to assess this question more fairly, the data must be

evaluated in a different light. As a whole, smoking initiation rates ranging from 4.3% to

9.3% appear to appear rather insignificant, but when participants who never smoked are

excluded from the analyses the picture changes. For example, 25.6% of students who

had ever smoked on two or more days in one month did so for the first time during

college. Moreover, one-third of students who had ever smoked 20 days/month, and who

had smoked on 30 consecutive days, had engaged in this behavior for the first time while

in college.

A similar picture emerges when current smokers, rather than ever smokers, are the

population of interest. Although 16% of the current smokers in the sample had first

smoked a whole cigarette during college, 26% had first smoked on two days in any

month during college. Moreover, one-third of the current smokers who had ever smoked

on 20 or more days in any month, and one-third of current smokers who had ever smoked









on 30 consecutive days, first reached these smoking frequencies during college. These

findings are consistent with a previous report by Wechsler and colleagues (1999). They

found that 28% of current smokers had progressed to regular smoking at the age of 19 or

older-most likely when they were in college.

Conversely, these smoking initiation percentages among current smokers are

inconsistent with those reported by Debernardo and colleagues (1999). In their study of

Yale University students, half of the current smokers reported starting to smoke in

college. This discrepancy between current smokers' initiation rates in this study and

those of Yale students may be attributable to differences in the studies' methodologies.

The survey done at Yale University used an email format, whereas this study surveyed

students in intact classrooms at five universities. More smokers who had begun smoking

in college may have responded to their survey than smokers who had begun smoking

prior to college. Another possibility may be that Yale students may not be representative

of students from other campuses, and that they have an unusually high percentage of

students who start smoking after coming to college.

When Do Most College Student Begin Smoking?

Although the percentage of current smokers in this study who started smoking in

college was inconsistent with the percentage in Debernardo's study, findings from both

studies indicated that the majority of smoking initiation occurred during the first two

years of college. Approximately 80% of the students at Yale University who reported

starting to smoke in college began smoking during their freshman or sophomore year.

Results of this study confirmed these findings. For example, 89.4% of the students in this

study who reported first smoking two or more cigarettes during college started this

behavior during their first two years of college.









Other descriptive results from this study further pinpoint when college student tend

to begin smoking. These data clearly indicate that the first six months of college is an

especially vulnerable time for potential smokers. Among students who had ever smoked

on two or more days in any month during college, 56.9% had started smoking during

their first six months. Furthermore, 60% had first smoked on 20 days during any one

month, and 50% first smoked 30 consecutive days during this time.

Result of smoking initiation behavior among freshmen, sophomores, and juniors,

corroborate that freshman year of college is a critical time for potential smokers. For

example, 41% of juniors who first smoked on two days during any one month during

college reached this frequency of smoking during their first six months of college, and

54% had reached this frequency by the end of their first year of college. Moreover, a

total of 72% had first started smoking this often by the end of their second year. The

implications of these findings by year in school further support a recommendation to first

target prevention efforts to freshmen, and then to sophomores.

Although students who begin smoking in college represent approximately 7-10% of

all college students, separating smoking initiation behavior by time in college is useful

for several reasons. First, these data pinpoint when the majority of smoking initiation and

progression occurs. Second, practitioners can use these results to time their prevention

efforts and to wisely use limited resources. Third, these data may help practitioners tailor

prevention messages to at-risk audiences, primarily freshmen and sophomores.

Correlations Between Covariates and Smoking Initiation

So far, this chapter has addressed current smoking behavior and smoking initiation

results. The percentage of college students in the study who currently smoke, how often,

and how much were discussed, and implications were described. Smoking initiation









results revealed that a relatively small percentage of college students begin smoking in

college, but much of this behavior occurs during the first two years of college. The first

six months of college appears to be a particularly vulnerable time for students.

This chapter will now focus on possible reasons why some college students begin

to smoke in college. This section will address the results of the associations between five

suspected covariates and the outcome variable (grade point average, parents' smoking

status, susceptibility to smoke, alcohol consumption, and living situation). These five

variables were not of primary interest in the study; rather, they were measured to be used

as control variables if necessary.

Four of the five variables had empirical support that linked them to smoking

initiation behavior primarily among a high school population. Three of these variables

(grade point average, parents' smoking status, and susceptibility to smoke) have been

identified as predicators of smoking initiation among adolescents. Binge drinking during

college has been associated with college smoking status. The fifth variable, living

situation, did not have empirical support for smoking initiation, however, data were

collected showing whether a participant lived alone or with others because of the social

nature of the three explanatory variables.

Of these five potential covariates, only susceptibility to begin smoking during

freshman year (r =0.40), and alcohol consumption during freshman year (r =0.20), were

moderately correlated with smoking initiation. The magnitude of the correlations

between the other three suspected covariates (parents' smoking status, grade point

average, and living situation) and smoking initiation failed to reach the inclusion criterion

of .10 or greater. Two explanations may be offered as to why parents' smoking status









was not influential among this sample. First, findings from the adolescent literature

indicate that parents' smoking status is more influential in smoking initiation during early

adolescence versus late adolescence. Therefore, if parents were to influence college

students' smoking initiation behavior, this most likely would have occurred prior to

college. Second, over 95% of the participants did not live with their parents during their

freshman year of college, and hence, their parents probably had minimal influence over

this behavior.

Like with parental influence, grade point average was included as a potential

covariate because of evidence showing a strong association among high school student

smoking initiation behavior. One possible reason why GPA was not associated with

smoking initiation in this study may be due to a lack of variability among college GPAs

when compared to high school GPAs. In this study, 87% of participants reported a first

semester GPA of 2.5 or higher; and 65% reported a first semester GPA of 3.0 or higher.

The lack of spread in the GPA scores among this sample could have undoubtedly affected

this result. Likewise, a lack of variability of participants' living situation probably

influenced the results of the correlation between this variable and the outcome variable

because nine out of ten participants lived with other students during their freshman year

of college.

Research Question Results

The three explanatory variables examined in this study were chosen based on

social cognitive theory and empirical support showing an association with cigarette

smoking initiation among middle and high school populations. Multiple logistic

regression analysis was used to answer each of the three research questions. Only one of









the three explanatory variables proved to be significant at a 0.05 level. The results for

each question are discussed below.

Research Question One

Does the perceived prevalence of peers' smoking behavior during freshman year of

college increase the risk of smoking initiation during the first two years of college?

Perceived prevalence of peers' smoking behavior was categorized as a personal

determinant within the social cognitive theory framework. Furthermore, a fair amount of

evidence was presented linking this variable to smoking initiation among adolescent

populations. Although perceived prevalence of peers' smoking behavior has both

theoretical and empirical support, the results from this study showed that it was not a

significant factor of smoking initiation among college students.

The empirical support for this variable comes predominantly from children and

adolescent populations. No study thus far has linked this variable influences smoking

initiation behavior among a college population. Although much evidence exists showing

that college students overestimate the percentage of college students who smoke, none

has attempted to use inferential statistics to link this misperception with smoking

initiation behavior. Thus one plausible explanation for a lack of significance for this

variable among a college population is that this finding is valid.

In contrast, another possible explanation may be that this result is not valid because

this variable could have been inaccurately measured. Of all items on the 77-item survey,

students expressed the most confusion about items measuring perceived prevalence of

peers' smoking behavior. Several students' written comments during instrument

development expressed difficulty in "guessing" the number of peers' who smoked during

freshman year or questioned the accuracy of their estimations. So although the reliability









coefficient of this scale was adequate (0.88), future investigation of how to best measure

this variable may be warranted.

Research Question Two

Does the smoking-related behavior of friends and roommates during freshman year

of college increase the risk of smoking initiation during the first two years of college?

Findings from a multiple logistic regression analysis indicated that the smoking-

related behavior of friends and roommates was a significant factor of smoking initiation

during the first two years of college. This relationship remained even when controlling

for susceptibility to smoke and for alcohol consumption during freshman year. Due to

the strong theoretical and research support for the influence of this variable on smoking

initiation among high school students, this finding was not surprising.

Smoking-related behavior of friends and roommates had the strongest theoretical

support among the three explanatory variables. It was classified not only as an

environmental determinant of smoking initiation within social cognitive theory, but it

also incorporated two key concepts within the theory-observational learning and direct

reinforcement. Seeing others smoke is a form of observational learning, or modeling,

and according to social cognitive theory, increases the likelihood of similar behavior. The

smoking status for four types of relationships (friends, roommates, sorority

sisters/fraternity brothers, and romantic partners) was part of how this variable was

measured, and the score increased as the number of friends who smoked increased.

In addition to observational learning, this variable also incorporated the concept of

direct reinforcement by measuring friends' offers and encouragement to begin smoking.

Like observational learning, as direct reinforcement increases for a particular behavior,

the likelihood of the behavior occurring or continuing also increases. Eight out of the









twelve items on this scale measured either friends' offers to smoke or friends'

encouragement to smoke. Hence, a participant who had received many offers and

encouragement to smoke would have a higher composite score for this variable than a

participant who had received relatively few offers to smoke.

In addition to strong theoretical support, friends' smoking-related behavior had the

strongest empirical support among the three explanatory variables. Longitudinal studies

of high school students provide strong evidence linking friends' behavior and smoking

initiation. Furthermore, in a small, cross-sectional study of college students, participants

who were susceptible to smoke in the upcoming year had indicated that friends' smoking

behavior might motivate them to start smoking. They selected this as a potential reason

to begin smoking more frequently than image, stress level, depression, or expression of

independence.

This finding confirms that smoking initiation is, at least in part, an interpersonal

phenomenon, and therefore, prevention efforts should address interpersonal

communication issues embedded within the behavior. Rather than focusing on the health

consequences or dangers or costs of smoking, interventions should concentrate on the

development of resistance or refusal skills. Health communication theory and planning

models would be especially useful in developing these types of interventions. For

example, a mass communication theme like "Don't let your 'friends' get you hooked on

cigarettes" could be tested on a college population for appeal and effectiveness.

Research Question Three

Does sociability during freshman year of college increase the risk of smoking

initiation during the first two years of college?









A limited number of studies provide evidence that sociability influences smoking

initiation among young people. Two longitudinal studies found extroversion to be

positively associated with smoking initiation among adolescents of both genders;

whereas, another found sociability to be predictive of smoking initiation only among

adolescent girls. A study of male and female college students indicated that extroversion

was positively associated with smoking initiation.

Although the association between sociability and smoking initiation has modest

empirical support among adolescents, the results from this study indicated that it is not a

significant factor of smoking initiation among college students. Two explanations may

be plausible for this finding. First, the finding may be valid. Of the three explanatory

variables in the study, sociability has the weakest amount of evidence linking it to

smoking initiation. Furthermore, these findings are mixed about its influence for each

gender. One study showed a significant association for both genders, whereas another

found such an association only among girls. Second, this finding may be invalid due to a

lack of variability of sociability scores for this population. The participants, as a whole,

were a fairly social which is not surprising considering the nature of undergraduate

student life and culture. Further investigation of this variable among college students is

needed to corroborate this particular finding.

Strengths and Weaknesses

The purpose of this study was to investigate cigarette smoking initiation behavior

among a college population. Prior to this investigation, data was limited and questions

remained about the size of this problem on college campuses. Furthermore, little was

known about the reasons why college students might start to smoke during their college

years. Like any study, this investigation had both strengths and weaknesses.









One strength of this study was expanding the research knowledge in this area by

providing information on how many students start smoking in college and when they are

most vulnerable to begin. Data for smoking initiation were collected by age as well as

point of time in college. Findings by age were consistent with those found in a large,

national study. Results also confirmed a previous report indicating that smoking

initiation is most likely to occur during the first two years of college. This study suggests

that the first year of college is an especially vulnerable time to start smoking. These

results have practical implications for anti-smoking prevention efforts. Health

practitioners can apply them to at-risk groups and can strategically time the delivery of

anti-smoking interventions. Given that health promotion resources are limited, this

information is valuable.

Another contribution of this study was investigating several reasons why college

students may begin to smoke. Results suggest that sociability, perceptions of peers'

smoking behavior, GPA, living situation, and parents' smoking status are not significant

factors in starting to smoke in college. In contrast, friends' behavior, susceptibility to

smoke during freshman year, and alcohol consumption during freshman year were

significantly associated with starting to smoke during freshman or sophomore year of

college. These findings serve as a starting point for more research on this issue and for

application to current campus smoking prevention efforts.

Although this study had several strengths, its findings must be interpreted in the

context of several limitations. First, the study was based on self-reported data and

required participants to recall information from their past. The extent to which recall bias

influenced these findings is unknown. Second, the data are cross-sectional, and therefore,









certainty about the direction of causation between the significant explanatory variable

and the outcome variable cannot be confirmed. What cannot be determined from these

data are whether persons who initiated smoking in college did so because their friends'

behavior influenced them, or whether these "smoking initiators" sought friends and

roommates who tended to smoke, and subsequently, offered and encouraged them also

smoke. Third, the sample was a convenience sample, rather than a random sample. This

limits the study's findings to its sample and not a broader college population.

Conclusions

This study examined whether three social factors affected cigarette smoking

initiation among college undergraduate students: (1) perceived prevalence of peers'

smoking behavior, (2) smoking-related behavior of friends and roommates, and (3)

sociability. Moreover, this study determined the percentage of college students who

began smoking in college, identified at-risk populations, and pinpointed when college

students were most vulnerable to begin smoking. Results from this study indicated that

26.8% of participants smoked during the 30 days preceding the survey. Unfortunately,

this falls far short of the Healthy Campus 2010 goal of 12%, and consequently, confirms

the need for anti-smoking efforts on college campuses.

A total of 9.3% of all participants began smoking cigarettes in college, with many

starting to smoke, and progressing in frequency, during their first two years of college.

These data indicate that the first six months are a particularly vulnerable time for

potential smokers. The specificity of these findings can help college health practitioners

implement their smoking prevention efforts at the most appropriate time during the

academic year.









Only two of the five suspected covariates were moderately correlated with smoking

initiation during the first two years of college. These covariates included susceptibility to

smoke during freshman year and alcohol consumption during freshman year. In contrast,

grade point average during freshman year, parents' smoking status prior to freshman

year, and students' living situation during freshman year failed to be moderately

correlated with smoking initiation during the first two years of college. Although

susceptibility and alcohol consumption were not the variables of interest for this study,

their relationship to smoking initiation warrants further research.

Of the three explanatory variables of this study, smoking-related behavior of

friends and roommates was the only significant factor of cigarette smoking initiation

during college. This relationship remained even after controlling for the other two

explanatory variables and the two covariates. In contrast, perceived prevalence of peers'

smoking behavior, as well as sociability, were not significantly associated with smoking

initiation during the first two years of college.

Recommendations

Although this study had some limitations, its findings can guide future practice

and research. The following recommendations are offered to help practitioners

implement anti-smoking initiatives for college students and for researchers who are

interested in investigating cigarette smoking initiation among a college population.

Recommendations for College Health Practice

Based on this study's findings, the following recommendations are provided for

college health practice.

1. Develop anti-smoking interventions specifically (a) tailored to susceptible
freshmen, (b) designed to address friends' behaviors and influence on the smoking
initiation process, (c) designed to help susceptible freshmen make a firm









commitment not to smoke, and (d) implemented within six months from when the
majority of freshmen begin college, typically during fall semester.

2. Address the misconception that once individuals graduate from high school, they
are no longer vulnerable to start smoking.

3. Use The College Student Smoking Survey, in part or in whole, to identify the
number and characteristics of college students who are likely to begin smoking.

4. Apply the findings from this study to inform decisions about the relative
importance of this health issue on college campuses and the amount of resources to
allocate to smoking prevention.

5. Create health interventions and messages that address the relationship between
cigarette smoking initiation and alcohol consumption.

Recommendations for Future Research

In addition to recommendations for college health practices, the following

recommendations are provided for research on smoking initiation among undergraduate

college students.

1. Replicate this study using a longitudinal research design and a random sample.

2. Conduct both quantitative and qualitative studies about the nature of friends' offers
and encouragement to smoke.

3. Conduct research on the effectiveness of an anti-smoking interventions targeted to,
and implement for, traditional-aged freshmen during their first six months of
college.

4. Investigate the nature of alcohol consumption and smoking initiation during the
first semester of college.

5. Conduct quantitative and qualitative research on susceptibility to start smoking
during freshman year of college.

This chapter focused on a discussion of a multi-campus study designed to

determine how many traditional-aged, undergraduate students start smoking cigarettes in

college, when they are most vulnerable to begin, and whether three social factors put

them at greater risk to start smoking in college. These findings expand the limited body

of literature in the area of smoking initiation among college students and should be






86


corroborated by future research. In the mean time, however, college health practitioners

can use these findings to guide the development and delivery of their smoking prevention

efforts. Such efforts could help at-risk students remain smoke-free during their college

years, and ultimately, could help to achieve the Healthy Campus 2010 goal of reducing

the percentage of college smokers to 12% or less by the year 2010.




Full Text

PAGE 1

SOCIAL FACTORS OF CIGARETTE SMOKING INITIATION AMONG UNDERGRADUATE COLLEGE STUDENTS By JANE F. EMMER E A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS F OR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2003

PAGE 2

Copyright 2003 by Jane F. Emmere

PAGE 3

In Dedication to Jill Margaret Fameree and JoAnn McFarlin

PAGE 4

iv ACKNOWLEDGMENTS My first name, Jane, means God is gracious. During my entire life, but particularly during these past four years, I have been graced with the love and support of so many outstanding people. Over the past four years, I have had wonderf ul support from professors and graduate students on campus. Dr. Jill Varnes is the best chairperson on the planet. I am grateful for her competence, caring, wisdom, and flexibility. I also thank Dr. R. Morgan Pigg for his calm and gentle spirit. I app reciate Dr. Bill Chen and Dr. Sadie Sanders for their expertise in the area of smoking and your encouraging words. I would also like to thank Dr. Randy Penfield for his patience and expertise with my survey development and data analyses. I learned so ma ny new things in his classes and during the dissertation process. I would also like to thank two professors who did not sit on my committee but had a significance influence on me. They were especially kind, enthusiastic, and supportive: Dr. Linda Behar Horenstein and Dr. Kim Walsh Childers. I am also indebted to my office mate, Terri Mitchell, and other fellow graduate students who have become dear friends: Teresa Lyles, Laura Hill, and Jamie Fisher. They are great friends. I am also graced with loving and positive individuals in my personal life. Without my mother JoAnns help, I would not be where I am today. I thank her for taking care of me during the most difficult times and celebrating with me during the most joyous ones. My sister, Jill, is my b est friend. I appreciate her prayers, optimism, and unfailing

PAGE 5

v encouragement, not only during the past four years but also for as long as I can remember! My brother, Jody; his wife, Julie; and my niblings (a made up word for niece and nephews) Jenna, Jac kson, and Jordan were a great distraction. I thank all of them for the camping trips, baseball and basketball games, rock climbing, and kayaking. I love hanging out with them. I thank my father, Gale, and my grandmother, Minerva, for their prayers and wise advice. My non biological family also gave me great support and help. I thank John Barbour, Stacey Hoffman, and Joy Asbell for their on going friendship, for bringing me meals on wheels, and for letting me borrow their children for kid therapy. Th eir children are a special blessing and helped me keep life in perspective. I would forget any academic worries while searching for the moon during the daylight, walking dogs and bathing cats, throwing stones in the pond, playing on the patio, and answe ring questions like Aunt Jane, can I come and live with you for six and a half days? It is my hope that one day Kaitee, Luke, John William, and Joseph will read this and realize how much joy they all gave me during a very challenging time in my life. I am also indebted to my family at Faith Presbyterian Church for their love for God, on going prayers and petitions, sound counsel and teaching, and ministering to me at a crucial time in my life. Finally, and most significantly, I thank the Lord, for bl essing me with His grace and with these remarkable people.

PAGE 6

vi TABLE OF CONTENTS page ACKNOWLEDGMENTS ................................ ................................ ................................ .. iv LIST OF TABLES ................................ ................................ ................................ ............... x LIST OF FIGURES ................................ ................................ ................................ ........... xii ABSTRACT ................................ ................................ ................................ ...................... xiii CHAPTER 1 INTRODUCTION ................................ ................................ ................................ ........ 1 Research Problem ................................ ................................ ................................ ......... 2 Purpose of the Study ................................ ................................ ................................ ..... 2 Rationale for the Study ................................ ................................ ................................ 2 The Negative Health Consequences of Smoking ................................ .................. 3 Smoking Reduction as a Current Health Priority ................................ .................. 4 Smoking Cessation as a Difficult Goal ................................ ................................ 4 Limited Research on College Students ................................ ................................ .. 5 Smoking Initiation Among College Students ................................ ............................... 5 Research Questions ................................ ................................ ................................ ....... 8 Delimitations ................................ ................................ ................................ ................. 9 Limitations ................................ ................................ ................................ .................... 9 Assumptions ................................ ................................ ................................ ................. 9 Definition of Terms ................................ ................................ ................................ .... 10 Summary ................................ ................................ ................................ ..................... 11 2 REVIEW OF THE LITERATURE ................................ ................................ ............ 13 Cigarette Smoking Among College Students ................................ ............................. 13 Smoking Prevalence Rates Among College Students ................................ ......... 13 Frequency and Quant ity Smoked ................................ ................................ ........ 14 Smoking Rates by Gender ................................ ................................ ................... 15 Smoking Rates by Race ................................ ................................ ....................... 15 Smok ing Trends Among College Students ................................ ......................... 16 Smoking Initiation Among College Students ................................ ............................. 17 Reasons Why College Students Smoke ................................ ................................ ...... 18 A Conceptual Framework for Smoking Initiation ................................ ...................... 20 Incorporating Susceptibility into the Stages ................................ ........................ 21

PAGE 7

vii Risk Factors of Smoking by Stage ................................ ................................ ...... 25 Summary ................................ ................................ ................................ ..................... 25 Social Cognitive Theory ................................ ................................ ............................. 27 Personal Determinants ................................ ................................ ......................... 27 Environmental Determinants ................................ ................................ ............... 28 Behavioral Determinants ................................ ................................ ..................... 29 Observational Learning ................................ ................................ ....................... 30 Reinforcement ................................ ................................ ................................ ..... 30 Summary ................................ ................................ ................................ .............. 31 The Relationship between the Three Explanatory Variables and Smoking Initiation ................................ ................................ ................................ ................. 31 Perceived Prevalence of Peers Smoking Behavior ................................ ............ 31 Smoking Related Behavior of Friends ................................ ................................ 34 Sociability ................................ ................................ ................................ ............ 37 Conclusion ................................ ................................ ................................ .................. 38 Summary ................................ ................................ ................................ ..................... 39 3 METHODS ................................ ................................ ................................ ................. 40 Research Design ................................ ................................ ................................ ......... 40 Research Va riables ................................ ................................ ................................ ..... 40 Instrument Development ................................ ................................ ............................ 41 Item Selection and Development ................................ ................................ ......... 41 Expert Review ................................ ................................ ................................ ..... 42 Student Evaluation ................................ ................................ ............................... 42 Pilot Study ................................ ................................ ................................ .................. 43 Setting ................................ ................................ ................................ .................. 43 Participants ................................ ................................ ................................ .......... 43 Data Collection Procedures ................................ ................................ ................. 43 Data Analysis ................................ ................................ ................................ ....... 44 Instrument Reliability ................................ ................................ .......................... 44 Final Survey Administration ................................ ................................ ....................... 45 Final Survey Format ................................ ................................ ............................ 45 Final Survey Description ................................ ................................ ..................... 45 Setting ................................ ................................ ................................ .................. 46 Participants ................................ ................................ ................................ .......... 46 Protection of Confidentiality ................................ ................................ ............... 47 Data Collection Procedures ................................ ................................ ................. 47 Data Analysis ................................ ................................ ................................ ....... 47 Controlling for Bias ................................ ................................ ............................. 48 Summary ................................ ................................ ................................ ..................... 49 4 RESULTS ................................ ................................ ................................ ................... 50 Sample Demographics and Characteristics ................................ ................................ 50 Participation by University, Gender, and Race ................................ ................... 51 Parti cipation by Age and Months in College ................................ ...................... 51

PAGE 8

viii Current Smoking Behavior ................................ ................................ ......................... 52 Smoking Behavior by Gender and Race ................................ ............................. 53 Frequency and Quantity Smoked ................................ ................................ ........ 54 Smoking Initiation Behavior ................................ ................................ ....................... 55 First Ever Smoking Behavior by Age ................................ ................................ 56 First Ever Smoking Behavior by Time in College ................................ .............. 57 Correlations Between the Covariates and Smoking Initiation ................................ .... 63 Research Question Results ................................ ................................ ......................... 65 Research Question One ................................ ................................ ....................... 65 Research Question Two ................................ ................................ ....................... 66 Research Question Three ................................ ................................ ..................... 67 Regression Analysis of All Three Explanatory Variables ................................ ... 68 Summary ................................ ................................ ................................ ..................... 69 5 DISCUSSION ................................ ................................ ................................ ............. 70 Current Smoking Behavior ................................ ................................ ......................... 70 Sm oking Initiation Results ................................ ................................ ......................... 72 How Many Students Start Smoking in College? ................................ ................. 72 When Do Most College Student Begin Smoking? ................................ .............. 74 Correlations Between Covariates and Smoking Initiation ................................ .......... 75 Research Question Results ................................ ................................ ......................... 77 Research Question One ................................ ................................ ....................... 78 Research Question Two ................................ ................................ ....................... 79 Research Question Three ................................ ................................ ..................... 80 Strengths and Weaknesses ................................ ................................ .......................... 81 Conclusions ................................ ................................ ................................ ................. 83 Recommendations ................................ ................................ ................................ ....... 84 Recommendations for College Health Practice ................................ ................... 84 Recommendations for Future Research ................................ ............................... 85 APPENDIX A LIST OF EXPERT PANEL MEMBER S ................................ ................................ ... 87 B INSTITUTIONAL REVIEW BOARD APPLICATIONS AND APPROVAL LETTERS ................................ ................................ ................................ ................... 88 C PARTICIPANT CONSENT LETTERS ................................ ................................ ... 119 D THE COLLEGE STUDent SMOKING SURVEY ................................ .................. 127 E FINAL SURVEY SCALE ANALYSES ................................ ................................ .. 137 F LIST OF FACULTY ADVI SORS ................................ ................................ ........... 142 G DATA COLLECTION PROTOCOL ................................ ................................ ....... 143

PAGE 9

ix LIST OF REFERENCES ................................ ................................ ................................ 145 BIOGRAP HICAL SKETCH ................................ ................................ ........................... 151

PAGE 10

x LIST OF TABLES Table page 2 1 Cigarette Smoking Prevalence Rates by Year and Source ................................ ....... 14 2 2 Cigarette Smoking Prevalence R ates by Race ................................ .......................... 16 2 3 College Students Motivations for Smoking ................................ ............................ 19 2 4 Stage Models of Smoking Acquisition, 1987 1998 ................................ ................. 22 2 5 Stage Models of Smoking Acquisition, 2000 2002 ................................ ................. 23 2 6 Actual Prevalence and Perceived Prevalence of Smoking Among College Students ................................ ................................ ................................ .................... 33 3 1 Instrument Specification Table ................................ ................................ ................. 45 4 1 Distribution of Participants by University, Gender, and Race ................................ 51 4 2 Distribution of Participants by Age and Months in College ................................ .... 52 4 3 Past 30 Day Smoking Prevalence Rates ................................ ................................ .. 53 4 4 Past 30 Day Smoking Prevalence Rates by Gender and Race ................................ 54 4 5 Number of Days and Quantity Smoked ................................ ................................ ... 55 4 6 First Ever S moking Behavior by Age for Sample and Current Smokers ................ 56 4 7 First Ever Smoking Behavior by Time in College for Sample, Current Smokers, and Ever Smokers ................................ ................................ ................................ ..... 58 4 8 First Ever Smoking Behavior by Year in School for all Participants ...................... 60 4 9 College Initiated Smoking Behavior by Time in College ................................ ........ 62 4 10 College Initiated Smoking Behavior by Year in School ................................ .......... 64 4 11 Correlations between Five Suspected Covariates and Smoking Initiation ............... 65

PAGE 11

xi 4 12 Multiple Logistic Regression Analysis of Perceived Peer Smoking Behavior and Smoking Initiation ................................ ................................ ............................. 66 4 13 Multiple Logistic Regression Analysis of Friends and Roommates Smoking Related Behavior and Smoking Initiation ................................ ................................ 67 4 14 Multiple Logistic Regression Analysis of Sociability and Smoking Initiation ........ 68 4 15 Multiple Logistic Regression Analysis of Three Explanatory Variables of Smoking Initiation ................................ ................................ ................................ .... 68 E 1 Reliability Analysis for Perceived Prevalence of Peers Smoking Behavi or Scale ................................ ................................ ................................ ....................... 137 E 2 Item Total Statistics for Perceived Prevalence of Peers Smoking Behavior Scale ................................ ................................ ................................ ....................... 137 E 3 Reliability Analysis for Smoking Related Behavior of Friends and Roommates Scale ................................ ................................ ................................ ... 138 E 4 Item Total Statistics for Smoking Related Behavior of Friends and Roommates Scale ................................ ................................ ................................ ... 138 E 5 Reliability Analysis for Sociability Scale ................................ ............................... 139 E 6 Item Total Statistics for Sociability Scale ................................ .............................. 139 E 7 Rel iability Analysis for Susceptibility to Smoke Scale ................................ .......... 140 E 8 Item Total Statistics for Susceptibility to Smoke Scale ................................ ......... 140 E 9 Reliabil ity Analysis for Alcohol Consumption Scale ................................ ............ 141 E 10 Item Total Statistics for Alcohol Consumption Scale ................................ ............ 141

PAGE 12

xii LIST OF FIGURES Figure page 2 1 Psychosocial risk factors of cigarette smoking. ................................ ....................... 26 2 2 A schematic of the explanatory va riables and covariates within a Social Cognitive Theory framework. ................................ ................................ .................. 28

PAGE 13

xiii Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy SOCIAL FACTORS OF CIGARETTE SMOKING INITIATION AMONG UNDERGRADUATE COLLEGE STUDENTS By Jane F. Emmere August 2003 Chair: Jill Varnes Major Department: Health Science Education This study examined whether three social factors affected cigarette smoking initiation during the freshman and sophomore years of college: (1) perceived prevalence of peers smoking behavior, (2) smoking related behavior of friends and roommates, and (3) sociability. Moreover, this study determined the percentage of students who initiated smoking during college, identified at risk populations, and pinpointed when college students are most vulnerable to begin smoking. A total of 1,772 undergraduate students in intact classrooms from five universities completed a researcher constructed self report survey. Questionnaires from 1,250 full time, tra ditional aged freshmen, sophomores, and juniors were analyzed. A total of 26.8% of the sample had smoked cigarettes during the 30 days before the survey, and 9.3% of the sample had smoked for the first time on two or more days during any one month during college. Nearly 57% of these smoking initiators had started smoking during their first six months of college; 70.7%, by the end

PAGE 14

xiv of their freshman year; and, 90%, by the end of their sophomore year. Moreover, one out of four participants who reported sm oking at least one cigarette during the 30 days before the study had started smoking in college. Results of multiple logistic regression analyses showed that the smoking related behavior of friends and roommates was the only statistically significant expl anatory variable ( p = .003). This relationship remained after controlling for the effects of perceived prevalence of peers smoking behavior, sociability, susceptibility to smoke during freshman year, and alcohol consumption during freshman year. The coe fficient ( b ) for friends and roommates smoking related behavior was 0.842. This value suggests that the odds of initiating smoking during freshman or sophomore year of college increases by e 0.842 = 2.321 or more than doubles for each unit increase in the composite score of the smoking related behavior of friends and roommates. Findings will assist college health professionals (1) to identify students at risk to begin smoking, (2) to develop and tailor anti smoking initiatives, (3) to implement intervent ions at the most strategic times during the academic year, and, (4) to help at risk students remain smoke free during their college years.

PAGE 15

1 CHAPTER 1 INTRODUCTION Cigarette smoking has many negative consequences. Between 1995 and 1999, smoking caused over 440,000 adult and over 1,000 infant deaths annually (Centers for Disease Control and Prevention [CDC], 2002). Currently, smoking causes on e out of every five deaths, making it the leading preventable cause of death in the United States (CDC, 1999). Furthermore, annual economic losses total over $157 billion the equivalent of $3,400 for every smoker in the United States (CDC, 2002). If cur rent smoking patterns continue, an estimated 25 million people alive today will die of smoking related illnesses (CDC, 1999, p. 41). The Surgeon General first warned Americans about the health hazards of cigarette smoking in 1964. At that time, one out o f every two adults smoked. Since then, over 10 million people have died from smoking related illnesses (CDC, 1997), and 56.3 million Americans continue to jeopardize their health despite repeated warnings (Substance Abuse and Mental Health Services Admini stration [SAMHSA], 2002). Persons of all ages choose to ignore public health warnings related to smoking, but adolescents and young adults have the highest smoking rates. Approximately 11% of middle school students (CDC, 2001), and 34% of students in gr ades 9 12 have smoked at least one cigarette in the past 30 days (United States Department of Health and Human Services [USDHHS], 2000). Moreover, in 2001, approximately 33% of full time, traditional aged college students have smoked in the past 30 days ( SAMHSA, 2002).

PAGE 16

2 Smoking rates among college students are comparable with those of high school students, and therefore, pubic health efforts for smoking prevention and cessation are warranted. Research Problem This study determined the percentage of colle ge students who began smoking in college and examined the influence of three social factors on cigarette smoking initiation during the freshman and sophomore years of college. These factors included (1) perceived prevalence of peer smoking behavior, (2) s moking related behavior of friends and roommates, and (3) sociability. Purpose of the Study This study was conducted to accomplish three purposes. (1) This study determined the percentage of college students who began smoking in college, identified at ris k populations, and pinpointed when college students are most vulnerable to begin smoking. This information is useful for targeting at risk populations and for implementing these anti smoking initiatives at the most appropriate times. (2) The study examin ed the effect of three social factors and smoking initiation during freshman and sophomore years of college. Findings will help college health professionals develop tailored interventions to help vulnerable students remain smoke free during college. (3) T he study helped to fill a void in the professional literature regarding smoking initiation among college students. Currently, a voluminous amount of research on smoking initiation has been conducted on middle school and high school students, whereas only a limited amount of research exists on smoking initiation among college students. Rationale for the Study Research on cigarette smoking initiation among college students is important for four reasons: (1) Smoking negatively contributes to serious health consequences, and a

PAGE 17

3 significant number of college students currently smoke, (2) reducing cigarette smoking is both a national and a college health priority, (3) nicotine is a highly addictive substance, and subsequently, smoking cessation is extremely diff icult, and (4) research on smoking initiation among college students is limited. The Negative Health Consequences of Smoking A significant number of college students are putting themselves and others at risk for serious health consequences because of smok ing, and the magnitude of these consequences is enormous. Smoking is the number one preventable illness behavior in the nation, resulting in more deaths each year than AIDS, alcohol, cocaine, heroine, homicide, suicide, motor vehicle crashes and fires com bined (USDHHS, 2000). Smoking has been conclusively linked to lung, kidney, bladder, laryngeal, oral, esophageal, and pancreatic cancer, cardiovascular disease, chronic obstructive pulmonary disease, and stroke in both males and females (Samet, 2001). M oreover female smokers have an increased risk for infertility, delayed conception, ectopic pregnancy, and spontaneous abortion when compared with nonsmokers (USDHHS, 2001a). Smoking not only negatively affects the health of smokers, but also harms nonsmo kers. Each year, 35,000 nonsmoking adults die of heart disease, and another 3,000 die of lung cancer, from passive smoke (CDC, n.d.) Furthermore, environmental tobacco smoke has been linked to the exacerbation of asthma, reduced lung functioning, and res piratory symptoms (Samet, 2001). Due to the enormous health hazards to smokers and nonsmokers alike, helping nonsmoking college students refrain smoke free, and assisting those who currently smoke with cessation efforts, is critical.

PAGE 18

4 Smoking Reduction as a Current Health Priority Because smoking is the number one preventable cause of disease, health professionals want to reduce adult smoking in the United States from 22% to 16% by the year 2010 (USDHHS, 2000). Although the prevalence rate for smoking amo ng college students (26 33%) exceeds the national rate (22%), campus health professionals have set the same goal as those who set the national agenda a reduction of smoking to 12% or less among college students by the year 2010 (Task Force on National Heal th Objectives in Higher Education, 2002). To accomplish this ambitious goal, national and college public health efforts must prevent the initial onset and progression of smoking as well as encourage and help smokers to quit. Smoking Cessation as a Diffic ult Goal Reducing smoking to 12% among college students by the year 2010 can be achieved in two ways either nonsmoking college students must remain smoke free, or those who currently smoked must quit. Quitting smoking, however, proves to be very difficult and is typically characterized by frequent relapses (CDC, 1998). For instance, seven out of ten current smokers in the United States want to quit smoking, but because of the highly addictive nature of nicotine, most smokers who try to quit repeatedly fai l ( USDHHS, 2000). Annually, 45% of smokers who try to quit smoking succeed for at least 24 hours every year, but fewer than 3% quit permanently (USDHHS, 2000). A large percentage of college smokers also want to quit smoking. Nearly two thirds have ever tr ied to quit smoking (CDC, 1997), and 50% to 80% have tried to quit smoking during the year prior to being surveyed (Debernardo et al., 1999; Everett et al., 1999). One study reported that one in five college students tries to quit five or more times annua lly (Wechsler, Rigotti, Gledhill Hoyt, & Lee, 1998). Although the majority

PAGE 19

5 of college smokers want to quit, only one in four succeed (Everett et al., 1999). Because the success rate of smoking cessation is so bleak, the optimal public health strategy is to prevent cigarette smoking completely (CDC, 1998, p. 387). Limited Research on College Students To make progress toward the Healthy Campus 2010 goal of reducing smoking on college campuses, the American College Health Association (ACHA) has recommende d that prevention and education initiatives address the risks of tobacco use and support non use (ACHA, 2000). Since the 1998 multi billion dollar Master Tobacco Settlement Agreement, funding for tobacco prevention and cessation programs has increased on college campuses nationwide. To develop effective anti smoking initiatives, however, more than money is required. College health practitioners also must have relevant information and tools. Currently, research on smoking initiation among college studen ts is limited and cannot adequately inform college health practitioners. This study will help college health professionals move toward their 2010 goal of reducing smoking among college students by providing useful information. Findings will assist colleg e health professionals in four ways: (1) to identify students at risk to begin smoking, (2) to develop or refine tailored anti smoking initiatives, (3) to implement interventions at the most appropriate times during the academic year, and ultimately, (4) to help at risk students remain smoke free during their college years. Furthermore, college health practitioners can use the College Student Smoking Survey to determine the nature and scope of cigarette smoking on their particular campus. Smoking Initiat ion Among College Students Primary prevention involves actions and interventions designed to identify risks and reduce susceptibility or exposure to health threats prior to disease onset (American

PAGE 20

6 Association for Health Education [AAHE], 2001, p. 101). In the case of smoking initiation among college students, the goal is to identify nonsmoking students who are at risk of becoming smokers in college, and then to create and deliver tailored interventions to help them remain smoke free. Although prevention of smoking among college students is warranted, the preponderance of primary prevention research and interventions have focused on middle and high school populations because four out of five smokers begin to smoke before the age of 18 (USDHHS, 2000). Appr oximately 20% of smokers begin smoking in their young adult years, representing around 11 million individuals who can be categorized as late onset smokers. Rates of smoking initiation among college students cannot be determined from existing data. Smokin g initiation among college students is not well studied. Furthermore, all national data collected from college students about smoking initiation have been collected by age rather than by educational setting. Results from the 1995 National College Health Risk Behavior Survey (NCHRBS) indicated that 22.6% of college students who had ever smoked began smoking when they were either 17 or 18 years old. A total of 11% began when they were age 19 years old or older (Everett et al.,1999; Wechsler et al., 1998 ). Results from a statewide survey administered in California showed that among students who had ever smoked, 14.3% smoked a whole cigarette for the first time at the age of 19 or older (Patrick, Covin, Fulop, Calfas, & Lovato, 1997). From these two studies, we can presume that at least 11% to 14.3% of college students who had ever smoked began the behavior during college. To accurately assess how many college students start smoking during college, however, data must be collected and analyzed by educational setting rather than by age.

PAGE 21

7 Only one study of college students has assessed smoking initiation by educational setting. Researchers at Yale University determined whether college smokers began smoking prior to or during college (Debernardo et al., 1999). Th ey also assessed how many nonsmoking students were at risk to become smokers in the future. Researchers found that 5.7% of the participants began to smoke cigarettes in college. Among the 11.3% of the sample who considered themselves current smokers, hal f indicated that they started smoking in college. Furthermore, 80% of participants who began smoking in college started this behavior during their freshman or sophomore year. The researchers also found that 19.4% of nonsmoking college students lacked a f irm commitment to refrain from smoking in the upcoming year. This finding was similar to national findings from a high school sample 22.5% of students who had never smoked were considered susceptible to trying cigarettes in the upcoming year (CDC, 2000). The collective results from the study of Yale University students suggested that a significant number are vulnerable to initiate smoking, particularly during their freshman or sophomore year of college. Although intriguing, these findings need validation for several reasons. First, college students from only one university participated in the study. Second, only 13% of the sample smoked, which was significantly lower than the corresponding national rate of 29% (CDC, 1997). Third, the definition for smok ing was not explicit. Participants were asked, Do you smoke cigarettes now? Current smoking is typically measured using a specific time frame, usually the 30 days prior to the survey. For example, many national surveys use the following question to me asure current smoking: In the past 30 days, have you smoked cigarettes? The inconsistency between how these two ways measured smoking prevalence may explain the relatively low

PAGE 22

8 prevalence rate in the Yale study. Fourth, and finally, the small sample siz e of college smokers limits the conclusions of this study. Only 58 college smokers and only 29 smoking initiators participated in this study. Because of these limitations, findings from this study are inconclusive and offer limited direction to health ed ucators for designing smoking prevention programs. As mentioned previously, most individuals who smoke begin smoking during adolescence (USDHHS, 2000; Everett et al., 1999). Subsequently, the preponderance of research on cigarette smoking initiation inves tigates middle and high school populations. However, this voluminous body of literature may shed light on why college students begin to smoke. Rigorous studies of middle and high school students unequivocally show that a variety of social factors increa se the risk of starting to smoke. These social risk factors include having friends, parents or older siblings who smoke; receiving offers and encouragement to smoke; being socially competent; and having an inflated perception of the norm for smoking (Carv ajal, Wiatrek, Evans, Knee, & Nash, 2000; Conrad, Flay, & Hill, 1992; Flay, Hu, & Richardson, 1998; USDHHS, 1994; West & Michell, 1999; West, Sweeting, & Ecob, 1999). Research Questions The following questions were investigated in this study. 1. Does the perc eived prevalence of peers smoking behavior during freshman year of college increase the risk of smoking initiation during the first two years of college? 2. Does the smoking related behavior of friends and roommates during freshman year of college increase the risk of smoking initiation during the first two years of college? 3. Does sociability during freshman year of college increase the risk of smoking initiation during the first two years of college?

PAGE 23

9 Delimitations The following delimitations were noted in t his study. 1. Participants were selected from undergraduate students at one of five universities: (1) Ball State University, (2) Texas A & M University, (3) University of Alabama, (4) University of Florida, or (5) University of Georgia. 2. Participants were tra ditional undergraduate students. 3. Participants reported beginning college within 36 months of the survey. 4. Participants reported having full time status during their first year of college and at the time of the survey. Limitations The following limitations w ere identified for this study. 1. Participants selected from the five universities may not represent a nationwide undergraduate student population. 2. Participation in the study was voluntary. 3. The diversity of the sample was limited to students who chose to part icipate in the study. 4. The generalizability of the results was limited to the sample of participants in this study. Assumptions The following assumptions were made for this study. 1. The five selected universities were representative of other universities in the United States. 2. Individuals who voluntarily agreed to participate were representative of all undergraduate students enrolled at their university. 3. The data collection instrument developed for this study was adequate for the purpose of this study. 4. Part icipants were able to accurately recall their perceptions and behaviors during their freshman year of college.

PAGE 24

10 Definition of Terms The following is a list of relevant terms for this study. Extroversion. A preference for attending to the outer world of ev ents by seeking active stimulation and involvement in the environment (Morris, 1979). First year of college. The first 12 months of attendance at a two or four year college or university, after graduating from high school. Synonymous to freshman year o f college. Freshman year of college. The first 12 months that a student attends a two or four year college or university, after graduating from high school. Health promotion. Any planned combination of educational, political, environmental, regulatory, or organizational mechanisms that support actions and conditions of living conducive to the health of individuals, groups, and communities (AAHE, 2001). Perceived prevalence of peers smoking behavior. An estimate of the percentage of the participants pe ers who engaged in a variety of smoking related behaviors during their freshman year of college. Primary prevention. Actions or interventions designed to identify risk and reduce susceptibility or exposure to health threats prior to disease onset (AAHE, 20 01). Second year of college. The second 12 months of attendance at a four year college or university, after graduating from high school. Synonymous to sophomore year of college. Smoking. Actively inhaling smoke from regular or clove cigarettes. Smoking initiation. The first time that a participant smoked on two or more days during any one month.

PAGE 25

11 Smoking related behavior of friends. Smoking cigarettes, offering cigarettes to others, or encouraging others to smoke. Sociability. The tendency to prefer the presence of others to being alone and seeking and being especially gratified by the presence, attention, sharing of activities, and stimulation of others (Buss & Plomin, 1986). Sophomore year of college The second 12 months in attendance at a four y ear college or university, after graduating from high school. Same as Second Year of College. Susceptibility. The lack of a firm commitment not to smoke in the future among nonsmokers (Pierce, Choi, Gilpin, Farkas, & Merritt, 1996). Traditional, under graduate student. A student who began college between the ages of 17 and 19 years old, and who were 18 to 22 years old at the time of the survey. Summary Each year, cigarette smoking costs thousands of premature deaths and millions of dollars in lost pro ductivity and health care costs in the United States. Consequently, national and college health goals include reducing smoking prevalence rates to 12% or less by the year 2010. To achieve this goal, aggressive efforts to prevent individuals from starting to smoke, and to help smokers stop smoking, must occur nationwide including on college campuses. Because most cigarette smoking begins during adolescence, research has overwhelmingly focused on middle and high school students rather than on college stude nts. Nonetheless, over 46 million people in the United States currently smoke, and one fifth or 9 million began smoking as young adults. Recent data suggest that a substantial number of young adults will begin to smoke during college, particularly durin g their freshman or sophomore year (DeBernardo et al., 1999;

PAGE 26

12 Everett et al., 1999; Wechsler et al., 1998). Therefore, research that examines potential risk factors of smoking initiation among this population is important. This chapter provided an overvie w of the study and included the research problem, purpose, rationale, as well as a brief overview of smoking initiation among college students. Additionally, research questions, delimitations, limitations, assumptions, and definitions of key terms for the study were presented. Chapter 2 provides background and context of cigarette smoking among college students, conceptual frameworks for the studys variables, and a review of relevant literature on smoking initiation among college students.

PAGE 27

13 CHAPTER 2 REVIEW OF THE LITERA TURE This study determined the percentage of college students who began smoking in college, identified at risk populations, and pinpointed when college students are most vulnerable to begin smoking. The effects between three social factors and smoking initiation during freshman and sophomore years of college were also examined. The first section of this chapter provides an overview of smoking and college students, including prevalence rates and trends, smoking initiation amon g college students, and reasons why college students smoke. The second section provides a conceptual framework for smoking initiation. The third section provides a conceptual framework to explain the relationship between the three explanatory variables a nd smoking initiation. The fourth section reviews the empirical support between each of the three explanatory variables and smoking initiation. Cigarette Smoking Among College Students The purpose of this section is to provide a background and context of cigarette smoking among college students. First, prevalence rates and trends will be described, and then, smoking initiation among college students will follow. This section will conclude with reasons why college students smoke. Smoking Prevalence Rates Among College Students National surveys administered between 1993 and 2001 indicate 30 day smoking prevalence rates for college students ranging from 22.3% 32.9%. (ACHA, 2002; CDC,

PAGE 28

14 1997; Johnston, OMalley, & Bachman, 2001; SAMHSA, 2002; Wechsler et al., 1998). Table 2 1 shows various prevalence rates by year and source. Table 2 1. Cigarette Smoking Prevalence Rates by Year and Source Year of Survey 30 Day Prevalence Annual Prevalence Life time EVER Prevalence Source 1993 22.3% 31.6% Wechsler et al., 1998 1995 29.0% 74.8% CDC, 1997 1997 28.5% 39.5% Wechsler et al., 1998 1999 40.4% 48.5% 69.0% SAMSHA, 2000 2000 28.2% 32.0% 41.3 Johnston et al., 2001 ACHA, 2002 2001 25.0% 32.9% 69% ACHA, 2002 SAMSHA, 2002 2002 26.5% ACHA, 2002 Frequency and Quantity Smoked The 30 day prevalence rates reported above include any amount of cigarette use during the month prior to the survey. Prevalence rates are also calculated based on how frequently students smoke, and how many days they smoke per month. For example, in 1995, 16.5% students smoked on 20 or more days during the month prior to being surveyed (CDC, 1997). In the fall of 2000 and the spring of 2002, these rates decreased from 15% to 11.9% (ACHA, 2002). In 2000, daily smoking rates among col lege students ranged between 10.6% and 18% (ACHA, 2002; Johnston et al., 2001). In the spring of 2002, 8.0% of college students smoked daily (ACHA, 2002). In addition to tracking the number of days per month smoked, Wechsler and colleagues (1998) assessed the number of cigarettes that college students smoked on the days that they smoked. Among the 28.5% of participants who smoked during the month prior to the survey, 43.6% smoked fewer than one cigarette on the days that they smoked. Moreover, 24.2% smok ed between one and nine cigarettes; 20.4% smoked 10 19

PAGE 29

15 cigarettes; and 11.8% smoked 20 or more cigarettes, on average, on the days that they smoked. Smoking Rates by Gender Findings from five national studies between 1993 and 2002 show similar smoking rate s for male and female college students (ACHA, 2002; CDC, 1997; Emmons, Wechsler, Dowdall, & Abraham, 1998; Johnston et al., 2001; Wechsler et al., 1998). Thirty day prevalence rates measuring any amount of cigarette use, 30 day daily prevalence rates, and annual prevalence rates were similar between males and females (ACHA, 2002; Johnston et al., 2001). For example, National College Health Assessment data collected in the spring of 2002 show that 27.6% of males and 25.8% of females reported smoking at lea st once in the 30 days prior to the survey (ACHA, 2002). A total of 12.1% of males, and 11.7% of females smoked on 20 or more days, and 8.5% of males and 7.7% of females smoked on all 30 days prior to the survey. Smoking Rates by Race Although rates of sm oking do not significantly differ by gender, they do significantly differ by race. According to the 1995 National College Health Risk Behavior Survey, (CDC, 1997) white students were significantly more likely than black students or Hispanic students to re port smoking: (1) any time in their lifetime, (2) any time during the month prior to the survey, (3) on 11 or more days during the month prior to the survey, (4) on 20 or more days during the month prior to the survey, and (5) 30 days during any month duri ng their lifetime. Additionally, Hispanic students were significantly more likely to report lifetime cigarette use and current use as were black students. Rates for other racial groups were not available from this study.

PAGE 30

16 Although more white students smoke than do blacks or Hispanics, researchers at Harvard University found significant increases in smoking among all racial groups during between 1993 and 1997 (Wechsler et al., 1998). Table 2 2 shows these data. Although black students had the lowest 3 0 day prevalence rate (13.7%), they had the highest increase in smoking between 1993 and 1997 (47%). The rates for Hispanic students also increased (12.0%), but this group had the lowest rate of increase compared to other racial groups. Table 2 2. Cigaret te Smoking Prevalence Rates by Race Race 30 day prevalence in 1993 30 day prevalence in 1997 % Increase African American 9.6 13.7 42.7 Asian/Pacific Islander 18.3 22.4 22.5 Hispanic 22.7 28.8 12.0 White 23.2 30.4 31.2 Smoking Trends Among College S tudents During the 1990s, overall smoking rates among college students steadily increased. Wechsler and his colleagues (1998) used data from the Harvard School of Public Health Alcohol Study to compare smoking rates between 1993 and 1997. In 1993, 22.3% of students smoked during the month prior to the survey; however, 28.5% smoked in 1997. This represented a 27.8% increase in smoking over this 4 year time period. During that time, rates also significantly increased for all student subgroups including ge nder, race/ethnicity, and year in school, as well as among students attending both private and public universities. Findings from the Monitoring the Future Study verified the Harvard researchers results (Johnston et al., 2001). Thirty day prevalence rat es in 1991 and 1999 were 23% and 31%, respectively, representing a 30% increase. Most recent data from this on going

PAGE 31

17 study, however, suggest that this upward trend may be changing. Findings from the year 2000 indicate that the 30 day smoking prevalence rate among college students slightly decreased for the first time since 1991. This 2.4% decline (from 31% to 28.2%) was not a statistically significant change. Results from the 2001 National Household Survey on Drug Abuse, however, show a higher 30 day s moking prevalence rate than that found in the 2000 Monitoring the Future Study (SAMHSA, 2002). Nearly 33% of full time college students, and 44.6% of part time students, aged 18 22 years had smoked some time during the month prior to the survey. Data from the Monitoring the Future Study and from the National College Health Assessment suggest that daily smoking may be leveling off after a steady increase throughout the 1990s (ACHA, 2002; Johnston et al., 2000). Findings from the former study show that the number of college students who smoked daily increased between 1991 and 1999 by 40% from 14% to 19%, respectively. Between 1999 and 2000, however, the rate dropped from 19% to 17.8%. Similarly, results from three annual administrations of the National Col lege Health Assessment also showed a downward trend in daily smoking among college students. These data indicated that 10.6%, 8.6%, and 8.0% of college students smoked daily in the 2000, 2001, and 2002, respectively (ACHA, 2002). Smoking Initiation Among College Students Research has been conducted that consistently assesses how many college students currently smoke, how much, and how often, however, research is lacking that assesses how many college students begin to smoke during college. Only two nation al studies have collected smoking initiation data, but unfortunately, these data were collected by age rather than educational setting. Nonetheless, these age specific data can shed some

PAGE 32

18 light on smoking initiation behavior among college students. For ex ample, results from the 1995 National College Health Risk Behavior Survey indicate that among students who had ever smoked, nearly nine out of ten smoked a whole cigarette for the first time at the age of 18 or younger (Everett et al., 1999). More specifi cally, 18.5% smoked a whole cigarette for the first time when they were 12 years old or younger; 24.5% when they were 13 or 14 years old; 23.7% at the age of 15 or 16 years, and; 22.6% when they were either 17 or 18 years old. Approximately 11% smoked the ir first cigarette at age 19 years or older. Furthermore, a national study conducted in 1997 found that 11% of student smokers smoked their first whole cigarette at the age of 19 or older (Wechsler et al., 1998). From these findings, we can assume that at least 11% of college smokers may begin smoking during college. However, to precisely assess smoking initiation during college, students must be asked whether they began smoking before or during college. A random study conducted among Yale University students assessed how many current smokers began smoking in college (Debernardo et al., 1999). Although current smokers comprised only 13.0% of the sample, half of them indicated that they started smoking in college. Furthermore, 80% of these smokers wh o began smoking in college started smoking during their freshman or sophomore year. These results suggest that a significant number of college students are vulnerable to start smoking in college, particularly during their freshman or sophomore year. Reas ons Why College Students Smoke Two studies of college students have investigated reasons why college students smoke. One national study used multivariate logistic regression to assess predictors of smoking among current college smokers (Emmons et al., 199 8). Their sample included students from 140 randomly selected 4 year colleges in the United States. The found that

PAGE 33

19 seven variables increased the likelihood of being a smoker in college: binge drinking prior to college; binge drinking during college; usi ng marijuana during college; having multiple sex partners in college; participation in leisure activities during college; dissatisfaction with college education; and choosing not to participate in athletics during college. Moreover, females who engaged in three high risk behaviors binge drinking, using marijuana, and having multiple sex partners during college were more likely to smoke than their male counterparts. However, researchers of this study did not assess how many smokers in their sample began to smoke in college, nor did they investigate whether these seven variables predicted smoking initiation during college. Nevertheless, their findings provide insight into possible correlates of smoking behavior among college students. The second study too k place at Yale University. This small scale study examined reasons why college students smoked and also why those who were contemplating smoking would begin to smoke (Debernardo et al., 1999). Descriptive results are shown in Table 2 3. According to th ese findings, large differences exist between smokers and Table 2 3. College Students Motivations for Smoking Motivation Smokers (%) Nonsmokers (%) Stress 49.3 6.5 Image smokers project 39.1 2.5 Expression of independence 36.2 1.8 Friends smoke 11.6 1 1.1 Depression 31.9 2.7 nonsmokers motivations for smoking. For example, nearly one out of two smokers reported that stress motivated them to smoke. Conversely, less than one in ten susceptible nonsmokers thought that stress would motivate them to beg in to smoke. A similar percentage of smokers and nonsmokers reported friends smoking behavior as

PAGE 34

20 motivator for smoking. Approximately 11.6% of the smokers claimed that friends who smoked influenced their current smoking behavior. Similarly, 11.1% of the non smokers implicated friends who smoke as potentially influencing their future smoking behavior. A Conceptual Framework for Smoking Initiation Cigarette smoking has been conceptualized as a gradual progression through a series of stages. In general, re searchers have divided the process into four to six stages beginning with not smoking and ending with established smoking and addiction (Flay, 1993, Mayhew, Flay, & Mott, 2000; Pallonen, Prochaska, Velicer, Prokhorov, & Smith, 1998; Stern, Prochaska, Velic er, & Elder, 1987). Two similar, yet distinct, types of these stage models have emerged (See Table 2 4). The first type, based on the Transtheoretical Model of smoking cessation, classifies smoking into four or five stages: precontemplation, contemplati on, decision making (or preparation), action, and maintenance (Stern et al., 1987; Pallonen et al., 1998). The second type of stage model also divides the process of smoking into five stages, but uses different labels for each stage: preparatory, trial, experimentation, regular use, and nicotine dependence/addiction (Flay, 1993) [See Table 2 4]. Recently, scholars have merged these two types of stage models into yet another five stage framework ranging from non smoking to established/daily use (Mayhew e t al., 2000). Their conceptualization is depicted in Table 2 5. The researchers collapsed the precontemplation, contemplation, and preparation stages of the Transtheoretical Model into one, non smoking stage with two sub stages based on smoking intention The four stages that follow the two nonsmoking stages use Flays (1993) labels and descriptions: tried, experimentation, regular use, and established/daily use.

PAGE 35

21 Incorporating Susceptibility into the Stages The concept of susceptibility is a recent addit ion to the stages of smoking acquisition (Prokhorov et al., 2002). Susceptibility as a measure for smoking initiation was first introduced in 1993 in a report of tobacco use in California. Since then, it has been found to be a strong predictor of smoking initiation among children and adolescents (Choi, Pierce, Gilpin, Farkas, & Berry, 1997; Distefan, Gilpin, Choi, & Pierce, 1998; Jackson, 1998; Pierce et al., 1996; Pierce et al., 1993; Unger, Johnson, Stoddard, Nezami, & Chou, 1997). Consequently, suscep tibility has been incorporated into the Youth Tobacco Survey (USDHHS, 2001). Pierce and colleagues (1993) define susceptibility as the absence of a determined decision not to smoke in the future (p. 43), and measure this construct with three questions: ( 1) Do you think you will smoke a cigarette soon? (2) Do you think you will smoke a cigarette in the next year? and (3) Do you think you would smoke if your closest friend offered you a cigarette? Those who respond "definitely not to all of three items are classified as nonsusceptible, whereas, all others are considered susceptible to future smoking behavior (Pierce et al., 1996). Prokhorov and colleagues (2002) have recently suggested that integrating susceptibility into the Transtheoretical Stage Model would improve its validity. More specifically, they hypothesized that an adolescents response to the question about a closest friends offer to smoke would further determine who was susceptible among those who also had no intention of smoking in the fut ure. Their integrated model proved to have better predictive and concurrent validity than did either of the two individual

PAGE 36

22 Table 2 4. Stage Models of Smoking Acquisition, 1987 1998 Source Stage 1 Stage 2 Stage 3 Stage 4 Stage 5 Stern et al., 1987 Pre contemplation Has not yet considered smoking and reports no desire to do so in the future. Perceives few positive consequences of smoking. Contemplation Thinking about starting to smoke. Perceives relatively more positive consequences than in Stage 1. Deci sion Making Pros and cons of smoking are closely balanced. Action Experiments with cigarettes, but not totally committed to smoking in the future. Receives minimal pleasure from smoking. Maintenance Smokes on a regular basis. Committed to smoking now and in the future. Has no desire to stop. Receives pleasure from smoking. Pallonen et al., 1998 Precontemplation Has not tried smoking and has no intentions of smoking in the next 6 months. Contemplation Never smokers who are thinking about trying to smoke i n the next 6 months. Preparation Never smokers who are thinking about trying to smoke in the next 30 days. Recent Action Smoked cigarettes regularly, but for less than 6 months. Flay, 1993 Preparatory Involves the formation of knowledge, beliefs and ex pectations about smoking and the functions that it can serve. Trying The first two or three times a person smokes. Peers are usually involved. Experimentation Repeated, but irregular smoking; often situation specific (like at a party). Regular Use Weekly smoking across situations and personal interactions. Nicotine Dependence and Addiction. Physiological need for nicotine. Established tolerance and withdrawal symptoms with quit attempts.

PAGE 37

23 Table 2 5. Stage Models of Smoking Acquisition, 2000 2002 Source Stage 1 Stage 2 Stage 3 Stage 4 Stage 5 Stage 6 Mayhew et al., 2000 Non smoking Preparation Does not intend to smoke. Non smoking Contemplation Intends to smoke. Belief and attitude formation; susceptible to peer pressure. Tried Has smoked, but no mo re than one or two cigarettes. Has not smoked in the past year. Experimentation Smokes occasionally on an experimental basis. Does not intend to become a permanent smoker. Regular Use Smokes at least monthly, but not as frequently as daily. Established/ D aily Use Smokes daily or almost daily. May smoke heavily on occasion. Nicotine dependence. Prokhorov, et al., 2002 Precontemplation Nonsusceptible Never smokers who have no intention of smoking in the next 6 months and state that they would definitely not smoke if a friend offered them a cigarette. Precontemplation Susceptible Never smokers who have no intention of smoking in the next 6 months, but fail to state that they would definitely not smoke if a friend offered them a cigarette. Contemplation Never smokers who are thinking about trying to smoke in the next 6 months. Preparation Never smokers who are thinking about smoking in the next 30 days.

PAGE 38

24 models. Furthermore, their integrated model predicted susceptibility among adolescents who report ed no intentions to smoke in the future. Based on these findings, they divided the precontemplation stage of smoking acquisition into two sub stages based on their definition of susceptibility. They categorized individuals with strong convictions to refu se a friends offer to smoke as Nonsusceptible, and those with weak convictions as susceptible (See Table 2 5). Since the introduction of susceptibility as a relevant construct in smoking behavior, researchers have validated ways to measure it among adoles cents (Pierce et al., 1996; Unger et al., 1997). Furthermore, the relationship between susceptibility and smoking acquisition has been examined primarily among adolescent populations (Choi et al., 1997, Jackson, 1997, 1998; Pierce et al., 1996; Prokhorov et al., 2002; Unger et al., 1997; Unger, Rohrbach, Howard Pitney, Ritt Olsen, & Mouttapa, 2001). Little investigation of this construct, however, has occurred among college populations. Only one study has examined how many nonsmoking college students may be susceptible to begin smoking. DeBernardo and colleagues (1999) measured susceptibility by asking students whether they would try cigarette smoking in the next 12 months (p. 63). They found that nearly one out of five nonsmoking participants were su sceptible to start smoking in the upcoming year. More specifically, their findings indicated that 11.5% of non smokers were considering smoking, and an additional 7.9% were unsure whether they would continue to abstain from smoking, during the next year. This study, however, did not examine the relationship between susceptibility and smoking initiation among college students.

PAGE 39

25 Risk Factors of Smoking by Stage In addition to susceptibility to smoke in the future, a variety of other psychosocial risk factors have emerged for smoking initiation and progression through the stages of smoking. Conrad, Flay, and Hill (1992) reviewed 27 prospective studies on smoking initiation published since 1980 and identified risk factors for each stage of smoking. Their find ings indicated that some type of social influence like parent, adult, sibling, and/or peers smoking behavior and attitudes plays a role at every stage of smoking acquisition. Their findings are summarized in Figure 2 1. Summary This section provided a co nceptual framework for smoking initiation and progression to addiction. The process of becoming a cigarette smoker was depicted on a continuum represented by four to six stages beginning with receptivity to smoking and ending with nicotine dependence and addiction. Within the last 10 years, the concept of susceptibility has been incorporated into two stage models of smoking acquisition. Prospective studies of smoking initiation among adolescents have identified a variety of psychosocial risk factors for each stage of the smoking process. Now that a conceptual framework for the outcome variable of the study has been described, a conceptual framework for the three explanatory variables, and the five covariates, will be provided.

PAGE 40

26 1. PREPARATO RY STAGE 2. TRYING STAGE No longer smokes 3.EXPERIMENTAL STAGE 4. REGULAR USE 5. ADDICTION/DEPENDENT SMOKER Figure 2 1. Psychosocial risk factors of cig arette smoking. Adapted from Flay (1993), Flay et al. (1998), Prokhorov et al. (2002), USDHHS (1991), and USDHHS (1994, p. 92), Psychosocial risk factors include advertising, adult/sibling role models who smoke cigarettes, and susceptibility to smoke. Adolescent forms attitu des and beliefs about smoking Risk factors include peer influences to smoke -friends smoking, approval, and offers to smoke, the perception that smoking is normative, the availability of cigarettes, alcohol use, and marijuana use. Adolescent smokes first few cigarettes. Risk factors include social situations and peers that support smoking, parental smoking, self efficacy in ability to refu se offers to smoke, and the availability of cigarettes. Risk factors include peers who smoke, the pe rception that smoking has personal utility, and few restrictions on smoking in school, home, and community settings Adolescent smokes at least weekly across a variety of situations and personal Adolescent has developed the physiological need for nicotine. Never smokes Quits smoking No longer smokes Adolescent smokes repeatedly, but irregularly.

PAGE 41

27 Social Cognitive Theory Psychosocial risk factors have been implicated at every stage of the smoking acquisition process, and therefore, a psychosocial theory is necessary to help explain the progression of smoking from cognitive susceptibility to nicotine dependence. One of the most promising models applied to smoking initiation is social cognitive theory (Wang, Fitzhugh, Eddy, Fu, & Turner, 1997). Social cognitive theory explains behavior as a process involving the dynamic interactions between personal, behavioral, and environmental factors (Bandura, 2002). This on going process is known as reciprocal determinism. Figure 2 2 diagrams the studys three explanatory variables and five covariates as a schematic showing reciprocal determinism between each pair of the three types of determinants in the model. Each of these three determinants as well as two concepts that are releva nt to smoking initiation, observational learning and reinforcement, are depicted and will be briefly described. Personal Determinants Personal determinants within social cognitive theory include thoughts, perceptions, and emotions, as well as biological fa ctors and genetic predisposition. More specific examples of personal determinants include beliefs, expectations, self perceptions, gender, ethnicity, and temperament. In this study, two explanatory variables, sociability and perceptions of peers smoking behaviors fit within this category. Furthermore, two of the five covariates in the study, susceptibility to smoke and grade point average, can be considered personal determinants of smoking initiation.

PAGE 42

28 Figure 2 2. A schematic of the explanatory variables and covariates within a Social Cognitive Theory framework. Adapted from: Bandura (2002) Environmental Determinants In contrast to personal determinants that are internal to an individual, environmental determinants are external factors that influence an individuals thoughts and behavior. With few exceptions, smoking initiation occurs within social environments. According to Bandura (2002), self development, adaptation, and change Smoking related behavior of friends Parents smoking status Living situation Cigarette Smoking Alcohol Consumption Perceived prevalence of peers smoking behavior Sociability Susceptibility Grade point average Observational Learning Direct Reinforcement Vicarious Reinforcement Personal Dete rminants Behavioral Determinants Environmental Determinants Cigarette smoking Alcohol consumption Smoking related behavior of friends Parents smoking status Living situation

PAGE 43

29 are embedded in social systems (p. 121). T raditional aged college students exist in a very dynamic and diverse social environment. Undergraduate students typically interact with roommates, housemates, dorm mates, classmates, co workers, instructors, friends, and/or romantic partners on a regular basis. Moreover, college students also tend to have many group affiliations, like being members of academic clubs, intramural sports teams, sororities or fraternities, or community groups. Because of the dynamic and diverse nature of relationships withi n college students lives, the social environment plays a crucial role in this study. In particular, the relationship of smoking related behavior and interactions between friends, roommates, romantic partners, and sorority sisters and fraternity brothers (if applicable) and smoking initiation was examined. In addition, parents smoking status and whether participants lived alone or with other students was assessed. Behavioral Determinants Behavioral determinants interact with both personal and environment al determinants. Social cognitive theory posits that a reciprocal relationship exists between a persons actions and his or her environment. In the case of cigarette smoking, for example, individuals who smoke may choose friends or roommates who also smo ke. Conversely, friends or roommates who smoke may influence a nonsmoker to start smoking. Another behavioral determinant that is relevant to smoking behavior is alcohol consumption. As mentioned earlier in this chapter, binge drinking both before high school and during college was predictive of smoking behavior in college. Therefore, this study measured alcohol consumption during freshman year among its participants.

PAGE 44

30 Observational Learning Two other constructs within social cognitive theory are also v ery relevant social influences of smoking initiation behavior. The first, observational learning, is defined as behavioral acquisition that occurs by watching the actions and outcomes of others behavior (Baranowski, Perry, & Parcel, 1997, p. 157). A k ey tenant of social cognitive theory is that human beings learn not only by direct experiences, but also vicariously through observing others (Bandura, 2002). Moreover, vicarious learning can occur both inside and outside of ones immediate environment. An example of vicarious learning within ones environment would be observing a group of college students smoking outside ones residence hall or sorority house, and creating interpretations about that situation. An example of observational learning occurr ing outside ones environment would be seeing an anti smoking commercial on TV or seeing a character in a movie smoking a cigarette, and subsequently forming thoughts, feelings, and interpretations about what was seen. Reinforcement In addition to observat ional learning, social cognitive theory suggests that two types of reinforcement influence thoughts and behavior. The first, known as vicarious reinforcement, is related to observational learning. In the previous example of the group of students smoking outside of a residence hall, vicarious reinforcement could occur if the observer admired, liked, or was attracted to one or more of the individuals who were smoking. In this instance, smoking may be positively reinforced in the observer. The second type of reinforcement direct reinforcement is inherent in the explanatory variable Smoking Related Behavior of Friends as measured in this study. Respondents indicated how many of their friends, roommates, romantic partners, and sorority sisters or

PAGE 45

31 fraternit y brothers smoked, offered them cigarettes, or encouraged them to smoke during their freshman year of college. The composite score of these three behaviors for these four types of relationships provided a measure of the level of positive reinforcement rec eived for initiating smoking. Summary This section provided a conceptual framework for the studys three explanatory variables and for five suspected covariates. Social cognitive theory posits that substance use is learned through observation and experien ces within the social environment. Moreover, behavioral, personal, and environmental determinants interact collectively and dynamically to influence an individuals thoughts and actions. Observational learning, vicarious reinforcement, and direct reinfor cement were also described as they relate to the studys variables. A diagram was created to integrate within a social cognitive theory framework all three of the explanatory variables, all five of the covariates, and pertinent concepts for this study. T he next, and final section of this chapter, will review evidence of the relationship between each of the three explanatory variables and cigarette smoking initiation. The Relationship between the Three Explanatory Variables and Smoking Initiation Perceive d Prevalence of Peers Smoking Behavior Social cognitive theory suggests that personal perceptions, whether real or imagined, influence behavior. Considerable evidence indicates that perceived prevalence of peers smoking behavior influences the perceiver s smoking behavior. In particular, children and adolescents are at an increased risk of becoming smokers if they overestimate the number of their peers who smoke (Bauman, Botvin, Botvin, & Baker,

PAGE 46

32 1992; Carvajal et al., 2000; Flay et al., 1998; Jackson, 1 997; Unger & Rohrbach, 2002; USDHHS, 1994). Flay and his associates (1998), for example, conducted a longitudinal study of the psychosocial predictors of the different stages of smoking among high school students. Perceived smoking behavior of peers and adults was among one of the many variables they measured among 7 th graders. Cigarette smoking stages were assessed 6 years later, when participants were in the 12 th grade. Students who never smoked, not even a puff or two, were classified as never smoke rs. Those who had smoked only one whole cigarette or less in their entire life were labeled triers Those who had smoked more than one whole cigarette in their lifetime, but had not smoked in the week preceding the assessment, were classified as experi menters Finally, students who smoked during the week before they had completed the survey were categorized as regular smokers The relationship between perceptions of smoking and all three active smoking stages, as well as, transitions between trying an d experimentation, and experimentation and regular use were assessed. The composite score of a 3 item scale served as the measure for perceived prevalence of smoking. Two items assessed perceived prevalence of peers smoking behavior, and one measured pe rceived prevalence of adult smoking behavior. The scales internal consistency was .71 as measured by Cronbachs reliability coefficient. Results indicated that perceived prevalence of peers smoking significantly predicted experimentation and regular us e. Prevalence estimates also predicted one of the two transitions measured between smoking stages. The transition between trying cigarettes to experimental use was significantly predicted by the participants perceptions

PAGE 47

33 of prevalence, however, this rel ationship was not found for the transition between experimental smoking and regular use. The study just described served as an example of the quality of many other longitudinal studies of perceived prevalence and smoking among children and adolescents (B auman et al., 1992; Carvajal et al., 2000; Flay et al., 1998; Jackson, 1997; Unger & Rohrbach, 2002). Collectively, findings provide strong evidence that perceived prevalence of peers smoking behavior influences smoking behavior. However, no published st udy currently exists that has examined the relationship between perceived peers smoking prevalence and smoking initiation among college students. Only descriptive data are available on prevalence estimates of peer smoking among college students. Descript ive data show that college students greatly overestimate the number of peers who have never smoked, as well as those who currently smoke (as measured as any cigarette use during the 30 days before data collection). Furthermore, the magnitude of these mis perceptions has been fairly consistent over the past three years (see Table 2 6). Table 2 6. Actual Prevalence and Perceived Prevalence of Smoking Among College Students Spring 2000 Spring 2001 Spring 2002 Actual Perceived Actual Perceived Actual Per ceived Never smoked 70.0% 9.0% 71.0% 8.0% 68.2% 6.1% Smoked one or more days in past month 30.0% 91.0% 25.0% 93.0% 26.6% 93.9% Source: American College Health Association, 2002. Evidence also suggests that college students overestimate daily smoking be havior among their peers (Page, 1998). In a study of 775 college students from one university,

PAGE 48

34 female students, on average, estimated that 42.3% of male, and 42.2% of female peers, smoked daily. Men also overestimated these daily prevalence rates but by a smaller margin. They perceived that 36.9% of men, and 33.9% of women on their campus smoked daily. Actual rates of daily smoking proved to be much lower 13.4% for men and 17.8% for women. Smoking Related Behavior of Friends Evidence from studies of ele mentary school students (Jackson, 1997), middle school students (Carvajal et al., 2000; Conrad et al., 1992), and high school students (Duncan, Tildesley, Duncan, & Hops, 1995; Flay et al., 1998; Killen et al., 1997; Wang et al., 1997) clearly indicate tha t friends play a major role in smoking initiation and escalation of use. For example, the first time an individual smokes most commonly occurs with a friend who already smokes, and reinforcement from friends has been implicated in further experimentation and progression of smoking (USDHHS, 1994). Furthermore, over 90% of high school students who smoke report that one or more of their closest friends smoke, compared to 33% of nonsmokers reporting having one or more close friends who smoke (CDC, 2001). Pr ospective studies of high school students provide strong evidence that friends smoking status and reinforcement are more significant than many other psychosocial risk factors for smoking initiation and progression (Duncan et al., 1995; Flay et al., 1998; Killen et al., 1997; Wang et al., 1997). For example, in a longitudinal study described above (Flay et al., 1998), friends smoking behavior significantly predicted trying, experimenting, and regular smoking. Moreover, this relationship became stronger a s the stage of smoking progressed and as the number of friends who smoked increased. Friends smoking behavior also predicted movement from trying cigarettes to

PAGE 49

35 experimentation for both genders, but not progression from experimentation to regular use. K illen and associates (1997) also examined the relationship between friends smoking behavior and smoking initiation among high school students. For several years, they annually assessed these behaviors among participants in 2 consecutive cohorts of 9 th gr aders. Among girls and boys with no history of smoking at baseline, those with friends who smoked at baseline were significantly more likely to try smoking over the studys duration. The researchers concluded that friends smoking was the most important predictor of smoking (p. 1013) for both genders when compared to other factors including temperament, depression, and drive for thinness. In another study of secondary school students, smoking initiation and progression were assessed over a 3 year time s pan (Wang et al., 1997). The relationship between parents, siblings and best friends smoking status and smoking initiation and progression was examined. Best friends smoking status proved to be the only consistent and significant predictor of smoking initiation and progression for males and females. A 4 year prospective study of five age cohorts who began the study between the ages of 11 and 15 years olds also found evidence that peer reinforcement was predictive of cigarette use. Results indicated that encouragement to smoke by a best friend and lack of family cohesion predicted initial cigarette use. Encouragement but not family cohesion also predicted progression to more advanced stages of smoking (Duncan et al., 1995). In contrast to the numero us prospective studies that examine friends smoking behavior and smoking initiation among high school students, only a limited amount of

PAGE 50

36 cross sectional data link friends smoking behavior to smoking initiation during college. In a study of African Ameri can students, current residence, parents and friends who smoked during the participants childhood, and current friends who smoked were significantly correlated with smoking at least one cigarette but less than one hundred total in ones lifetime (Hestick, Perrino, Rhodes, & Sydnor, 2001). When these variables were subsequently entered into a prediction model, however, only parents and friends who smoked during the participants childhood proved to be statistically significant in predicting trial smoking b ehavior in college. Current friends smoking status was not a significant predictor of trial smoking in this study, however, it did predict lifetime smoking behavior defined as smoking 100 or more cigarettes in ones lifetime. Furthermore, the risk of l ifetime smoking was reduced when neither friends nor parents smoked and the student viewed spirituality as important. An email assessment of undergraduate students at Yale University also provided a limited amount of cross sectional data on friends beh avior and smoking status (Debernardo et al., 1999). Nonsmoking students who were susceptible to start smoking within a year after the survey were asked about motivations for smoking. Approximately 11% indicated that they would be motivated to begin smoki ng because of friends who smoked. The most frequently selected motivator for future smoking initiation behavior was friends behavior (11%), followed by stress (6.5%), depression (2.7%), image that smokers project (2.5%), and expression of independence ( 1.8%). Obviously, these descriptive data cannot assess whether having friends who smoke actually increased the risk of smoking among their susceptible friends.

PAGE 51

37 Sociability Sociability is defined as the tendency to prefer the presence of others to being alone and seeking and being especially gratified by the presence, attention, and sharing of activities (Buss & Plomin, 1986). Psychologists generally agree that sociability is one component of extroversion (Morris, 1979; Sipps & Alexander, 1987). Extrov ersion is defined as the preference for attending to the outer world of events by seeking active stimulation and involvement in the environment (Morris, 1979). Extroverts inherently value, enjoy, and engage in a wide range of social and affiliative activi ties and [are] socially confident (Morris, 1979, p. 89). Not surprisingly, similar items that measure both constructs are included in the International Personality Item Pool (2001). Compared to the amount of research on the relationship between friends behavior and smoking initiation, extroversion and its related construct of sociability has been studied less extensively. Yet overall, a limited number of longitudinal and cross sectional studies have found an association between sociability (or extrover sion) and smoking initiation for both adolescents and college students. As early as the 1970s, for example, researchers of a large, longitudinal study found extroversion to be positively associated with smoking initiation for both sexes (Cherry & Kiernan, 1976). In a cross sectional study of undergraduate students conducted in the 1980s, findings were similar smokers were significantly more likely than nonsmokers to be extroverts (Spielberger & Jacobs, 1982). More recently, sociability was examined as one of a variety of potential predictors of smoking initiation among high school students (Killen et al., 1997). In a 4 year prospective study, researchers collected data from two consecutive cohorts of 9 th graders

PAGE 52

38 to determine whether sociability influen ced smoking initiation. They found that sociability predicted smoking initiation among the girls only in both cohorts. A longitudinal study of Australian adolescent twins examined the relationship of extroversion and other personality factors and smoking i nitiation among 1,400 twins who were 11 to 18 years old at baseline (White, Hill, & Hopper, 1996). Three years after the baseline assessment, the researchers categorized each participant as a nonsmoker, an experimental smoker, or a recent smoker. A nonsm oker had never smoked a whole cigarette. An experimental smoker had not smoked in the month preceding the survey, but had smoked at least 10 cigarettes in his or her lifetime. A recent smoker had smoked during the month preceding the survey and had smoked at least 10 cigarettes in his or her lifetime. Results indicated that extroversion, but not self esteem, psychoticism, or neuroticism, consistently distinguished nonsmokers who remained smoke free from nonsmokers who started smoking. Furthermore, researc hers found a positive relationship at the end of the three year study between extroversion and smoking both among the experimental smokers and the recent smokers. Conclusion All three social variables in this study have been linked to smoking initiation am ong adolescents. Of these three explanatory variables, friends smoking related behavior has the strongest empirical support, followed by perceived prevalence of peers smoking behavior, and then sociability. These conclusions are based overwhelmingly on research on adolescent populations because rigorous research on the social risk factors of smoking initiation among college students is lacking.

PAGE 53

39 Summary Chapter 2 began by providing a context of cigarette smoking among college students. Smoking prevale nce rates, trends, characteristics of college student smokers, reasons for smoking, and smoking initiation rates were provided. Following this overview, two conceptual frameworks were described one for the outcome variable and one for all other variables in the study. First, stage models of cigarette smoking were described. Then, social cognitive theory was presented to explain the relationship between the studys explanatory variables and smoking initiation. Finally, evidence to support each of the thr ee explanatory variables was provided. A comprehensive literature review confirmed that a considerable amount of empirical support exists between each of these three variables and smoking initiation among high school populations. Conversely, little is kn own about these variables and smoking initiation among college students. Therefore, the field of college health will benefit from research that examines the influence of each of these social factors on smoking initiation.

PAGE 54

40 CHAPTER 3 METHODS This study determined smoking the percentage of college students who began smoking in college, identified at risk populations, and pinpointed when college students are most vulnerable to begin smoking. The effects between three social fa ctors and smoking initiation during freshman and sophomore years of college were also examined This chapter describes the methodology used in the study and is organized into five sections: (1) research design, (2) research variables, (3) instrument devel opment, (4) pilot administration, and (5) final administration. Research Design This study used a cross sectional survey design. Participants in intact classes at five universities completed a survey on social factors influencing smoking initiation. T he selected classes were comprised of primarily first and second year undergraduate students representing a variety of academic majors. Participants were traditional age college students. They began college by the age of 19 years and were between the ag es of 18 and 22 years old at the time of the survey. Research Variables This study examined the relationship between three explanatory variables and one outcome variable. The three explanatory variables included: (1) perceived prevalence of peers smoki ng behavior, (2) smoking related behavior of friends and roommates, and (3) sociability. Smoking initiation during college served as the outcome variable. Smoking

PAGE 55

41 initiation was defined as the first time that a participant smoked a whole cigarette on tw o or more days during a month. Instrument Development A 77 item College Student Smoking Survey was developed specifically for this study. The survey was designed to assess the relationship between the three explanatory variables and smoking initiation and to provide descriptive data that could inform decision making and program development on college campuses. Procedures outlined by Alreck and Settle (1995) were followed to develop and select items and to conduct an expert review as well as student review These three steps of the development process are described in the following subsections. Item Selection and Development First, a comprehensive review of current surveys on tobacco use among adolescents and college students was conducted. Items measuri ng demographics, frequency and quantity smoked, alcohol consumption, and susceptibility were selected or adapted from a variety of sources (ACHA, 2002; CDC, 1997; Johnston et al., 2001; Pierce et al., 1996; SAMSHA, 2002). Next, possible items for the soci ability scale were located and selected from the International Personality Item Pool (2001). Finally, the principal researcher created additional items to measure the outcome variable as well as the three explanatory variables. Potential items were cate gorized into nine sections: (1) Cigarette Smoking Initiation, (2) Susceptibility, (3) Social Context of First Ever Cigarette, (4) Social Norms, (5) Sociability, (6) Frequency of Going Out, (7) Smoking Related Behavior of Friends and Roommates During Fresh man Year, (8) Control Variables, and (9) Freshman Year and Current Smoking Behavior.

PAGE 56

42 Expert Review Six experts in cigarette smoking, health education, and college health reviewed potential items. They were instructed to rate the relevance of each item, sel ect the best items within each section, recommend additional items, and make suggestions. Based on the expert review data, two versions of a pilot survey were developed and reviewed by nine undergraduate students. Appendix A contains the names of the six r eviewers. Student Evaluation Students at the University of Florida helped in the preliminary stages of survey development in two ways they provided qualitative definitions of various levels of intoxication, and they reviewed two similar versions of the pilot survey. Seventy six male and female students from two sections of a Personal and Family Health class were asked to define moderately intoxicated and extremely intoxicated in their own words. Their responses were tallied and definitions were creat ed based on their feedback. Furthermore, nine students reviewed two similar versions of the pilot instrument. To keep the survey to a length that could be completed in less than 15 minutes, two versions were necessary. Both versions contained many identi cal items including those that measured demographics, smoking behavior, grade point average, living situation, and parents smoking status. However, items on each of the following four scales differed by version: sociability, perceptions of peers smoking behavior, friends and roommates behavior, and susceptibility to smoke. Freshman and sophomore reviewers were recruited from a fall 2002 Personal and Family Health class at the University of Florida, and ultimately, nine females reviewed the surveys. Fi ve students reviewed version one of the instrument; and four students reviewed version two. The researcher instructed participants to record how long it took

PAGE 57

43 them to complete the survey, to identify problems and suggest solutions, and to offer general com ments on how to improve the survey. Based on these results, a final version of the survey was created and then pilot tested on undergraduate students. Pilot Study Setting The pilot study was conducted using 20 intact classes at the University of Florida in fall 2002. Participants were enrolled in five sections of Personal and Family Health, one section of Human Physiology, or 14 lab sections of Human Anatomy. Classes were comprised of male and female students from a variety of academic majors. Partici pants Females comprised 77.9% of the pilot sample. On average, participants were 17.9 years old when they began their freshman year of college, and 19.3 years old at the time of the survey. Nearly 90% lived with other individuals besides their parents du ring the majority of their freshman year. Only 5.2% lived alone during that time. The percentages of Asian, Black, Hispanic, and White participants were 9.3%, 11.3%, 9.6%, and 67.3%, respectively. Approximately 2% of the participants indicated that they were American Indian, Alaskan Native, or Biracial. Data Collection Procedures The University of Floridas Institutional Review Board approved the data collection procedures for the pilot study (see Appendix B). The principal researcher collected data in each of the 20 classes. Participants received an informed consent letter, a survey booklet, a scantron form, and a pencil. The researcher read a script explaining the study to the students and their rights as a potential research participant. Participat ion was voluntary and anonymous. To ensure anonymity, signed consent forms were

PAGE 58

44 collected separately from survey responses (see Appendix C). Participants recorded their responses on a scantron form, and most completed the survey in 10 to 15 minutes. Data Analysis Data from the pilot study were scanned and electronically transmitted onto a computer disk at the University of Floridas Office of Academic Technology. Raw data from 733 participants were loaded into SPSS v. 10.0. A total of 688 participants met the following inclusion criteria for the pilot study: (1) they began college at the age of 17, 18 or 19 years, and (2) they were between the ages of 18 and 22 years old at the time of the survey, and (3) they were enrolled full time during their fresh man year and at the time of the pilot study. Instrument Reliability Item and scale analyses were conducted to assess the surveys four major scales. All items on the Perceptions of Peers Smoking Behavior, the Smoking Related Behavior of Friends and Roo mmates, and the Susceptibility scales had acceptable item discriminations. In contrast, nine of the 13 items on the Sociability scale had acceptable item discriminations, and therefore, four weaker items were subsequently removed from this scale. The re liability of each scale was then calculated using Cronbachs measure of internal consistency. All scales measuring the three explanatory variables (Perceived Prevalence of Peers Smoking Behavior, Smoking Related Behavior of Friends and Roommates, and Soc iability) had similar reliability coefficients of .84, .84, and .87, respectively. Furthermore, the Susceptibility scale had a high reliability coefficient of .90.

PAGE 59

45 Final Survey Administration Final Survey Format The final survey was developed using Rem ark OMR version 5.5 software developed by Principia Products. The final surveys format was changed after the pilot study for four reasons. First, the Remark OMR software became available to the researcher after the pilot study. Second, the software sto red all scanned surveys as an electronic image. Third, participants recorded their responses directly on the survey and could use ink rather than only pencil. Consequently, scantron forms and pencils did not need to be distributed for the final study. T his shortened the survey administration time by five minutes when compared to the pilot survey. Fourth, participant error was reduced because responses were recorded directly on the survey as opposed to on a scantron form. Final Survey Description The Col lege Student Smoking Survey is a 77 item survey that was typically completed in 10 minutes or less (See Appendix D). Forty three of the items comprised one of the surveys five scales. Table 3 1 lists the type of variable measured, the number of items, t he number of response choices, and the reliability coefficient for each scale. More detailed results of the item and scale analyses are located in Appendix E. Table 3 1. Instrument Specification Table Scale Variable Type of Variable Item Number Response Number Reliability Coefficient Perceived prevalence of peers smoking behavior Explanatory 9 5 .880 Smoking related behavior of friends and roommates Explanatory 12 4 .916 Sociability Explanatory 9 4 .836 Susceptibility to smoking Covariate 10 4 .953 Alcohol consumption Covariate 4 6 .947

PAGE 60

46 The remaining 35 items measured demographics, control variables, other potential covariates (besides susceptibility and alcohol consumption), smoking initiation behaviors, and 30 day prevalence of smoking. The fin al item on the survey solicited qualitative comments from the participants. Setting Undergraduate students from Ball State University, Texas A & M University, the University of Alabama, the University of Florida, and the University of Georgia participated in the final study. These five universities were selected for several reasons: (1) each had at least 15,000 undergraduate students enrolled during the 2002 2003 academic year, (2) each was able to provide a faculty liaison to coordinate data collection, and (3) each is located in a state that does not rely on tobacco revenue for its economic stability. Participants Male and female undergraduate students in intact classes participated in the final study. Classes were chosen based on two criteria: (1) t hey contained primarily first and second year undergraduate students, (2) they were comprised of students from a variety of academic majors. An elective introductory health class was chosen at four out of the five universities. A political science clas s, required of all undergraduate students, was selected at Texas A & M University. Between 200 and 600 students at each university completed a survey for an initial sample size of 1,775. Five inclusion criteria were applied to the initial data set: (1) participants started college when they were 17, 18 or 19 years old, (2) participants were 18 to 22 years old at the time of the survey, (3) participants began college within 36 months of the survey, (4) participants were enrolled full time during their fre shman year

PAGE 61

47 of college and at the time of the survey, and (5) participants provided sufficient data to determine their smoking initiation status. A total of 525 surveys failed to meet all five inclusion criteria yielding a final sample size of 1,250. The amount of excluded surveys was not surprising because data was collected in intact classes that obviously contained students who were over 22 years old at the time of the study. Protection of Confidentiality Each of the five universitys Institutional Rev iew Board approved the final study. All participants received a description of the study and information about their rights as participants. Participation was voluntary and anonymous. Written consent was obtained from participants at Ball State Universi ty, the University of Alabama, the University of Florida, and the University of Georgia. Participants at Texas A & M University were not required to provide a written consent form. Data Collection Procedures Data were collected in selected classes at fiv e universities between January 6 and February 6, 2003. The researcher collected data at the University of Florida. She recruited a faculty advisor at each of the other four sites (see Appendix F) and sent them the survey materials in early January. Th e same data collection protocol was followed at all five sites (see Appendix G). The faculty advisors returned the completed surveys to the researcher by February 20, 2003. Data Analysis Surveys were scanned using Remark OMR version 5.5 software. Data w ere then loaded into SPSS version 10.0 for analysis. To verify data entry accuracy, 75 surveys were randomly selected and the original responses were compared to the scanned responses. The original responses on all 75 surveys matched their scanned counte rparts.

PAGE 62

48 Three types of data analyses were conducted on the final data set to establish a profile for the sample and to test the research hypotheses. First, descriptive statistics were calculated to determine the baseline smoking initiation rates as well as the 30 day prevalence rates of current smoking among students. They were also calculated to obtain a general demographic profile of the sample and a specific profile of college student smokers. Second, correlations were calculated between each of the five covariates (susceptibility to smoke cigarettes during the upcoming year, grade point average, living situation, alcohol consumption, and parents smoking status) and smoking initiation to determine which of these five to include in the final regressi on analysis. According to Cohen (1977), correlations of .10 or greater represent at least moderate associations, and therefore, this criterion was used to determine inclusion of the five covariates. Third, multiple logistic regression was used to examine the relationship between each of the three explanatory variables and smoking initiation. Analyses for all three research questions were tested at a .05 significance level. Controlling for Bias Potential bias was controlled in three ways. First, the same data collection protocol was followed at all five sites. Second, individual differences were controlled. More specifically, data analyses were limited to responses from participants who (1) began college when they were 17, 18 or 19 years old, (2) were 1 8 22 years of age at the time of the survey, (3) began college within 36 months of completing the survey, and (4) were enrolled full time during their first year of college at the time of survey. Third, data was collected to measure five potential covari ates. Those with a .10 or greater association with the outcome variable were included as covariates in the multiple logistic regression analyses.

PAGE 63

49 Summary This chapter described the research design, research variables, instrument development, and pilot a nd final studies methodologies. A cross sectional survey design was used for this study. The College Student Smoking Survey was developed and pilot tested according to recommended theory and practice (Alreck & Settle, 1995). Findings indicated that th e survey was reliable. Data were collected in intact classes during January and February of 2003. A total of 1,775 students from five universities participated in the final study, however, after five inclusion criteria were applied, 1,250 participants rem ained. The majority of participants who were excluded from the study did not meet the criterion of being a traditional undergraduate student or were over the age of 22 years. Descriptive statistics were calculated to identify sample demographics and cha racteristics of college smoking behavior, including the percentage of participants who started smoking during college. Furthermore, smoking prevalence rates were calculated. Covariates with moderate associations to smoking initiation were identified for inclusion in the multiple logistic regression analyses to test the three research hypotheses. The results of the final study are reported in Chapter 4.

PAGE 64

50 CHAPTER 4 RESULTS This study determined the percentage of college students who began smoking cigarettes in college, identified at risk populations, and pinpointed when college students are most vulnerable to begin smoking. The effects between three social factors and smoking initiation during freshman and sophomore years were also examined. This chapter presents the results of the study and is organized into five major sections. The first three sections provide descriptive analyses of the sample, current smoking behavior, and smoking initiation behavior. The fourth section describes the associations between the suspected covariates and smoking initiation, and the fifth section presents the results for the three research questions. Sample Demographics a nd Characteristics This section describes demographics and characteristics of the sample. A total of 1,772 male and female undergraduate students from five universities participated in the study. Five criteria were selected to ensure that participation was limited to only full time, traditional aged freshman, sophomore, and junior students. Seniors were excluded because of the concern for the accuracy of recalling information from their freshman year of college. After applying the inclusion criteria t o the data set, 1,250 surveys remained, yielding a 70% inclusion rate. Approximately 60% of those excluded were seniors, and another 10% had failed to indicate how many months they had been in college. The remaining 30% of those who were excluded were either not traditional aged students or failed to provide sufficient data to determine their smoking initiation status.

PAGE 65

51 Participation by University, Gender, and Race Table 4 1 provides a summary of participants by university, gender, and race. Of the 1,2 50 participants, one out of three attended the University of Florida (37.5%). Nearly 21% were enrolled at Ball State University, followed by 15.8% from Texas A & M University, 14.1% from the University of Georgia, and 11.6% from the University of Alabama Two out of three participants were female (68%), and nearly 8 out of 10 were White (78.7%). Table 4 1. Distribution of Participants by University, Gender, and Race f % University Ball State University 261 20.9 University of Alabama 145 11.6 Uni versity of Florida 467 37.5 University of Georgia 175 14.1 Texas A & M University 197 15.8 Total 1,245 Gender Male 398 32.0 Female 845 68.0 Total 1,243 Race African American 117 9.4 American Native 1 0.1 Asian 45 3.6 Latino 67 5.4 Pacifi c Islander 3 .2 White 978 78.7 Multi racial 32 2.6 Total 1,243 Note. University affiliation was missing for 5 participants. Gender and race were missing for 7 participants. Participation by Age and Months in College Table 4 2 shows the age of the participants at the time of the study and also the number of months since they began their freshman year. Approximately 37% of the

PAGE 66

52 participants were 19 years old. The mean age of the sample was 19.25 years. One out of four were 18 years old, and another 25% were 20 years old. Only 10% of participants were 21 or 22 years old at the time of the study. Freshmen comprised the largest percentage of the sample (46.2%), followed by sophomores (32.5%), and then juniors (21.4%). Table 4 2. Distribution of Pa rticipants by Age and Months in College f % Cum % Age 18 319 25.5 25.5 19 464 37.1 62.6 20 324 25.9 88.6 21 123 9.8 98.4 22 20 1.6 100.0 Total 1,250 100.0 Months since beginning college 6 447 35.8 35.8 7 12 130 10.4 46.2 1 3 18 261 20.9 67.0 19 24 145 11.6 78.6 25 30 160 12.8 91.4 31 36 107 8.6 100.0 Total 1,250 100.0 Current Smoking Behavior This section describes how many participants smoked, how much, and how often during the 30 days preceding the study. National surveys typically use a dichotomous question to measure current smoking behavior: Have you used (or smoked) cigarettes in the past 30 days? In this study, participants were given four response choices to describe their smoking behavior during the 30 days preceding the survey: (1) never smoked, not even a puff or two, (2) tried a cigarette, but only a puff or two, (3) took puffs from cigarettes on occasion, but didnt smoke a whole cigarette at one sitting, or (4)

PAGE 67

53 smoked one or more whole cigar ettes. Table 4 3 presents these results and distinguishes between three groups of participants: never users, puffers, and those who smoked at least one or more whole cigarettes during the 30 days before the survey. Table 4 3. Past 30 Day Smoking Pre valence Rates f % Never smoked, not even a puff or two 916 73.5 Smoked, but only took puffs 75 6.3 Smoked one or more whole cigarettes 255 20.5 Total 1,246 Note. 30 day smoking behavior was missing for 4 participants. Nearly three out of four part icipants (73.5%) did not smoke in the 30 days before the survey. In contrast, a total of 26.8% reported some cigarette use during that time period. Among these cigarette users, 6.3% had only puffed on one or more cigarettes during that time, whereas, 20 .5% had smoked one or more whole cigarettes. Smoking Behavior by Gender and Race Table 4 4 shows 30 day smoking prevalence rates by gender and race. Nearly 28% of the males and approximately 25.% of the females had smoked in the 30 days before the survey The same percentage of males (20.1%) as females (20.2%) smoked one or more whole cigarettes, however, slightly more males (7.8%) than females (5.1%) only puffed during that time. Pacific Islanders had the highest prevalence rate among the different rac ial groups (66.7%), however, these results must be interpreted with caution because only three Pacific Islanders participated in the study. Approximately one of out four White participants (28.7%), and a similar number of multiracial participants (28.1%) had smoked during the month before the survey. The prevalence rate for Latinos was 22.4%, for Asians, 19.9%, and for African Americans, 12%.

PAGE 68

54 Smoking rates for participants who only puffed on cigarettes during the 30 days before the study also differe d among the various racial groups. For example, Latino participants had the highest rate of puffing on cigarettes during the 30 days before the study (12.0%), followed by White (6.0%), African American (6.0%), and Asian participants (2.2%). No Pacific Isl ander or multi racial participant reported puffing on cigarettes during the month before the study. Table 4 4. Past 30 Day Smoking Prevalence Rates by Gender and Race Smoked < 1 cigarette Smoked > 1 cigarettes Any cigarette use f % f % f % Gender Male 31 7.8 80 20.1 111 27.9 Female 43 5.1 171 20.2 214 25.3 Missing 1 4 Total 75 255 Race African American 7 6.0 7 6.0 14 12.0 American Native 0 0.0 0 0 0 0 Asian 1 2.2 8 17.7 9 19.9 Latino 8 12.0 7 10.4 15 22.4 Pacific Islander 0 0.0 2 66.7 2 66.7 White 59 6.0 222 22.7 281 28.7 Multi racial 0 0.0 9 28.1 9 28.1 Total 75 255 Frequency and Quantity Smoked This section describes the behavior of the participants who smoked one or more whole cigarettes during the 30 days befo re the study. This group is referred to as current smokers. Table 4 5 shows how often and how much the current smokers smoked during the month before the study. The majority of current smokers reported smoking on 20 or fewer days (64%), and 33.6% smok ed on five or fewer days during the month prior to the survey. Only 20.2% had smoked on all 30 days prior to the survey.

PAGE 69

55 Current smokers were also asked how many cigarettes, on average, they smoked on the days when they smoked. Three out of four repo rted smoking five or fewer cigarettes on the days that they smoked 22.4% had smoked one cigarette, and 51.6% had smoked two to five cigarettes, on the days that they smoked. Only 10% reported smoking between 11 and 20 cigarettes, and only one out of the 2 55 current smokers in the sample reported smoking more than a pack a day when he or she smoked. Table 4 5. Number of Days and Quantity Smoked f % Cum % # of days 1 19 7.5 7.5 2 29 11.5 19.0 3 5 37 14.6 33.6 6 10 34 13.0 46.6 11 20 44 17.4 64.0 21 29 40 15.8 79.8 All 30 51 20.2 100.0 Total 253 # of cigarettes 1 56 22.4 22.4 2 5 129 51.6 74.0 6 10 40 16.0 90.0 11 20 24 9.6 99.6 21 or more 1 0.40 100.0 Total 250 Note. Results are reported for participants who smoked one or more whole cigarettes during the 30 days prior to the survey and represent their smoking behavior during that month. The number of days was missing from 2 participants, and the number of cigarettes smoked was missing from 5 participants Smoking Initia tion Behavior This section describes smoking initiation behavior among all participants as well as subpopulations within the sample. Findings showing the age of when participants first smoked a whole cigarette are presented first. Then, the results of f our types of smoking initiation behavior are presented including the first time that participants had (1) smoked whole cigarette, (2) smoked on 2 days during any one month, (3) smoked on 20 days in

PAGE 70

56 any one month, and (4) smoked on 30 consecutive days. The se results are presented by time in school as well as year in school. First Ever Smoking Behavior by Age Table 4 6 shows how old participants were when they first smoked a whole cigarette. Of the 1,250 participants, 32.7% had smoked their first whole ci garette by the time they were 17 years old, and another 7.8% had smoked their first whole cigarette when they were 18 years old. Collectively, only 3.5% of the participants had smoked their first whole cigarette after the age of 18 years. Table 4 6. Fi rst Ever Smoking Behavior by Age for Sample and Current Smokers Sample Current Smokers Smoked first whole cigarette f % Cum % f % Cum % 17 or younger 408 32.7 32.7 185 72.8 72.8 18 98 7.8 40.5 44 17.3 90.1 19 30 2.4 42.9 17 6.7 96.8 20 12 1.0 43.9 7 2.8 99.6 21 1 .1 44.0 1 0.4 100.0 Never 700 56.0 100.0 Total 1,249 100.0 254 100.0 Note. Data was missing for 1 participant and 1 current smoker. Current smokers are participants who smoked one or more whole cigarettes during the 30 days b efore the survey (n = 255). Table 4 6 also shows how old current smokers were when they first smoked a whole cigarette. Current smokers were defined as participants who had smoked one or more whole cigarettes during the 30 days before the survey. Among this group of 255 participants, 72.8% had first smoked a whole cigarette by the time they were 17 years old. A total of 17.3% had smoked their first whole cigarette at the age of 18 years, and 6.7% had first smoked a whole cigarette when they were 19 year s old. Collectively, only 3.2% of these current smokers had first smoked a whole cigarette when they were 20 or

PAGE 71

57 21 years old. Moreover, no current smoker reported first smoking a whole cigarette after the age of 21 years. First Ever Smoking Behavior by Time in College One purpose of this study was to pinpoint when smoking initiation occurred during college. Table 4 7 shows these results for the sample, current smokers, and ever smokers. Ever smokers were defined as participants who had ever engaged in each of the four smoking initiation behaviors listed in the table. Among the sample, 43.8% had ever smoked a whole cigarette in their lifetime and 36.5% had done this prior to college. Only 7.3% had first smoked a whole cigarette during college. A to tal of 27% of the sample had first smoked on two or more days in any one month before they came to college, whereas, only 9.3% had engaged in this behavior after they came to college. Results are also presented for more frequent smoking. A total of 16.4 % of the participants had ever smoked on 20 days in any one month with 11.2% first smoking this frequently prior to college, and 5.2% starting some time during college. Finally, 13.5% of the sample reported ever smoking on 30 consecutive days. Just over 9% had first smoked this frequently prior to college, whereas, 4.3% had first smoked on 30 consecutive days during college. In addition to showing first ever smoking behavior for all participants in the sample, Table 4 7 also presents this information for current smokers. Over 84% of current smokers had first smoked a whole cigarette prior to college. Approximately 70% had first smoked on two days in any month; 42.5% had first smoked on 20 days during any one month; and 36.1% had first smoked on 30 conse cutive days prior to college. Results also show that 3.5% of current smokers had never smoked on 2 days in one

PAGE 72

58 Table 4 7. First Ever Smoking Behavior by Time in College for Sample, Current Smokers, and Ever Smokers Whole cigarette 2 days/month 20 days/m onth 30 consecutive days f % Cum % f % Cum % f % Cum % f % Cum % Sample First 6 months 55 4.4 4.4 66 5.3 5.3 39 3.1 3.1 27 2.2 2.2 Second 6 months 10 0.8 5.2 16 1.3 6.6 7 0.6 3.7 8 0.6 2.8 Second year 24 1.9 7.1 22 1.8 8.4 15 1.2 4.9 16 1 .3 4.1 After 2 nd year 2 0.2 7.3 12 0.9 9.3 4 0.3 5.2 3 0.2 4.3 Prior to college 456 36.5 43.8 337 27.0 36.3 140 11.2 16.4 114 9.2 13.5 Never 702 56.2 100.0 797 63.7 100.0 1044 83.6 100.0 1079 86.5 100.0 Total 1249 100.0 1250 100.0 1249 100.0 1247 10 0.0 Current Smokers First 6 months 21 8.3 8.3 35 13.8 13.8 32 12.6 12.6 23 9.1 9.1 Second 6 months 4 1.6 9.9 10 3.9 17.7 5 2.0 14.6 7 2.8 11.9 Second year 14 5.5 15.4 14 5.5 23.2 13 5.1 19.7 13 5.2 17.1 After 2 nd year 1 0.4 15.8 7 2.8 26.0 2 0.8 20.5 2 0.8 17.9 Prior to college 214 84.2 100.0 179 70.5 96.5 108 42.5 63.0 91 36.1 54.0 Never 9 3.5 100.0 94 37.0 100.0 116 46.0 100.0 Total 254 100.0 254 100.0 254 100.0 254 100.0 Ever Smokers First 6 months 55 10.0 10.0 66 14.6 14.6 39 19.1 19.1 27 16.1 16.1 Second 6 months 10 1.8 11.8 16 3.5 18.1 7 3.4 22.5 8 4.7 20.8 Second year 24 4.4 16.2 22 4.9 23.0 15 7.4 29.9 16 9.5 30.3 After 2 nd year 2 0.4 16.6 12 2.6 25.6 4 2.0 31.9 3 1.8 32.1 Prior to college 456 83.4 100.0 337 74.4 100.0 160 68.1 100.0 114 67.9 100.0 Total 547 453 235 168 Note. Current smokers smoked one or more cigarettes during the 30 days preceding the survey. Ever smokers excluded participants who reported never within each category of smoking behavior.

PAGE 73

59 month; 37.0% had never smoked on 20 out of 30 days; and 46% had never smoked on 30 consecutive days. Finally, Table 4 7 presents first ever smoking behavior results for ever smokers. These results include o nly those participants who had ever engaged in that specific smoking behavior. For example, the percentages for ever smokers under the whole cigarette category were calculated from the 43.8% of the sample who indicated that they had smoked a whole cigar ette. The 56.2% of the sample who reported never smoking a whole cigarette were excluded from these results. These descriptive results answer the question, Among participants who had ever smoked at the various smoking level, how many engaged in that par ticular behavior prior to college, and how many engaged in that behavior during college? Results show that as smoking behavior progresses, the percentage of students who initiate that particular behavior during college increases. For example, 16.6% of participants who had ever smoked a whole cigarette had first smoked it during college. This percentage increases to 25.6% for those who had ever smoked on 2 days in any one month, and to 32.1% for participants who had ever smoked on 30 consecutive days. Another interesting way to present these smoking behavior data are by year in school. Table 4 8 presents first ever smoking behavior by year in school for all participants. Freshmen were defined participants who had been in college for one to 12 months. Sophomores had been enrolled for 13 24 months, and juniors had been in college for 25 36 months. For reasons mentioned at the beginning of this chapter, seniors were excluded from this study.

PAGE 74

60 Table 4 8. First Ever Smoking Behavior by Year in School for all Participants Whole cigarette 2 days/month 20 days/month 30 consecutive days f % Cum % f % Cum % f % Cum % f % Cum % Freshmen First 6 months 22 3.8 3.8 28 4.9 4.9 18 3.1 3.1 10 1.8 1.8 Second 6 months 1 0.2 4.0 1 0.2 5.1 0 0.0 3.1 0 0.0 1.8 Prior to college 209 36.2 40.2 162 28.1 33.2 71 12.3 15.1 59 10.2 12.0 Never 344 59.8 100.0 385 66.8 100.0 487 84.6 100.0 506 88.0 100.0 Total 577 100.0 576 100.0 576 100.0 575 100.0 Sophomores First 6 months 21 5.2 5.2 22 5.4 5.4 11 2.7 2.1 9 2.2 2.2 Second 6 months 7 1.7 6.9 10 2.5 7.9 4 1.0 3.7 5 1.2 3.4 Second year 11 2.7 9.6 15 3.7 11.6 11 2.7 6.4 9 2.2 5.6 Prior to college 142 35.0 44.6 104 25.6 37.2 36 8.9 15.3 29 7.2 12.8 Never 225 55.4 100.0 255 62.8 100.0 343 84.7 100.0 353 87.2 100.0 Total 406 100.0 406 100.0 405 100.0 405 100.0 Juniors First 6 months 12 4.5 4.5 16 5.9 5.9 10 3.8 3.8 8 3.0 3.0 Second 6 months 2 0.7 5.2 5 1.9 7.8 3 1.1 4.9 3 1.1 4.1 Second year 13 4 .8 10.0 7 2.7 10.5 4 1.5 6.4 7 2.6 6.7 After 2 nd year 2 0.7 10.7 11 4.1 14.6 3 1.1 7.5 2 0.8 7.5 Prior to college 105 39.3 50.0 71 26.6 41.2 32 12.0 19.5 26 9.7 17.2 Never 133 50.0 100.0 157 58.8 100.0 214 80.5 100.0 220 82.8 100.0 Total 267 100.0 26 7 100.0 266 100.0 266 100.0 Note. Freshmen = 1 12 months in college. Sophomores = 13 24 months. Juniors = 25 36 months. Seniors were not part of the study.

PAGE 75

61 Noteworthy of these data by year in school are the percentages of students who had never smok ed at the four various levels shown in Table 4 8. The largest percentage of participants who had never smoked within each category of smoking initiation consistently is freshmen, followed by sophomores, and then juniors. For example, 66.8% of freshmen ha d never smoked on two or more days in any month, however, this percentage for sophomores was 62.8%, and for juniors, 58.8%. The last two tables in this section on smoking initiation behavior focus specifically on college initiated smoking behavior. In o ther words, the results listed in Tables 4 9 and 4 10 represent only those participants who reported beginning to smoke during college. Two groups are shown in Table 4 9 each of the four types of smoking initiation behavior: (1) participants who had eve r initiated the behavior during college, and (2) current smokers. Table 4 10 summarizes the four types of smoking initiation behavior by year in school. Results in Table 4 9 indicate that the majority of college initiated smoking behavior among the enti re sample, and among current smokers, occurred during the first six months of college. For example, among those who had ever first smoked on 20 days during any one month in college, 60% did so during their first six months of college. Among current smok ers, 61.5% smoked on 20 days in any one month for the first time during their first six months of college. Furthermore, relatively little smoking initiation behavior occurred after the sophomore year (3.9%).

PAGE 76

62 Table 4 9. College Initiated Smoking Behavio r by Time in College Whole Cigarette 2 Days/Month 20 Days/Month 30 Consecutive Days f % Cum % f % Cum % f % Cum % f % Cum % Ever Initiators First 6 mos. 55 60.4 60.4 66 56.9 56.9 39 60.0 60.0 27 50.0 50.0 Second 6 mos. 10 11.0 71.4 16 13 .8 70.7 7 10.8 70.8 8 14.8 64.8 Second year 24 26.4 97.8 22 19.0 89.7 15 23.1 93.9 16 29.6 94.4 After 2 nd year 2 2.2 100.0 12 10.3 100.0 4 6.1 100.0 3 5.6 100.0 Total 91 116 65 54 Current Smokers First 6 mos. 21 52.5 52.5 35 53.0 53.0 32 61.5 61.5 23 51.1 51.1 Second 6 mos. 4 10.0 62.5 10 15.2 68.2 5 9.6 71.1 7 15.6 66.7 Second year 14 35.0 97.5 14 21.2 89.4 13 25.0 96.1 13 28.9 95.6 After 2 nd year 1 2.5 100.0 7 10.6 100.0 2 3.9 100.0 2 4.4 100.0 Total 40 66 52 45

PAGE 77

63 Another way to present these college initiated data is by year in school. These data help to pinpoint when the majority of students in each class initiated each type of smoking behavior. Table 4 10 shows that over 95% of the freshmen who had e ver initiated these four smoking behaviors in college had engaged in these behaviors for the first time during their first six months of college. Similarly, the largest percentage of college initiated smoking behavior among sophomores and juniors also oc curred during the first six months of college. Moreover, the data representing the juniors shows that the majority of college smoking initiation among this group had occurred during their first two years of college. For example, 85% of the juniors who fi rst smoked on 20 days during any one month, and 90% of the juniors who had ever smoked on 30 consecutive days, engaged in these behaviors by the end of their sophomore year in college. Correlations Between the Covariates and Smoking Initiation This secti on will describe the relationship between five covariates and the dependent variable. To briefly review, the behavior of interest for the study was not the first time that participants tried a cigarette, or even when they first smoked a whole cigarette; r ather, the behavior of interest was when participants began to smoke repeatedly. Therefore, the dependent variable of the study smoking initiation was defined as the first time a participant smoked on 2 or more days in any one month during their freshman or sophomore year of college. A total of 104 students, or 8.3% of the sample, met this definition of smoking initiation. Data from this subpopulation of students was used to assess the association between five suspected covariates (susceptibility to smok e, grade point average, living situation, alcohol consumption, and parents smoking status) and smoking initiation. Data from this group was also used to analyze each of the three research questions.

PAGE 78

64 Table 4 10. College Initiated Smoking Behavior by Ye ar in School Whole cigarette 2 days/month 20 days/month 30 consecutive days f % Cum % f % Cum % f % Cum % f % Cum % Freshmen First 6 months 22 95.7 95.7 28 96.6 96.6 18 100.0 100.0 10 100.0 100.0 Second 6 months 1 4.3 100.0 1 3.4 100.0 0 0.0 100.0 0 0.0 100.0 Total 23 100.0 29 18 100.0 10 100.0 Sophomores First 6 months 21 53.8 53.8 22 46.8 46.8 11 42.3 42.3 9 39.1 39.1 Second 6 months 7 18.0 71.8 10 21.3 68.1 4 15.4 57.7 5 21.8 60.9 Second year 11 28.2 100.0 15 31.9 100.0 11 42.3 100.0 9 39.1 100.0 Total 39 100.0 47 100.0 26 100.0 23 100.0 Juniors First 6 months 12 41.4 41.4 16 41.0 41.0 10 50.0 50.0 8 40.0 40.0 Second 6 months 2 6.9 48.3 5 12.8 53.8 3 15.0 65.0 3 15.0 55.0 Second year 13 44.8 93.1 7 18.0 71.8 4 20.0 85.0 7 35.0 90.0 After 2 nd year 2 6.9 100.0 11 28.2 100.0 3 15.0 100.0 2 10.0 100.0 Total 29 100.0 39 20 Note. Freshmen = 1 12 months in college. Sophomores = 13 24 months. Juniors = 25 36 month s. Seniors were not part of the study. College initiators are participants who reported engaging in smoking initiation behavior during college.

PAGE 79

65 Table 4 11 shows the correlations between each of five suspected covariates and smoking initiation. Only two co variates were moderately associated with smoking initiation as defined by a predetermined inclusion criterion of 0.10. Susceptibility to smoke during freshman year of college ( r =.40), and alcohol consumption ( r =.20) each showed a moderate association to smoking initiation, and subsequently, both were included in the multiple logistic regression analyses used to answer the studys three research questions. In contrast to these two variables, the other three variables (grade point average freshman year, l iving with others during freshman year, and parents smoking status) failed to meet the inclusion criterion of 0.10. Consequently, they were excluded from further analyses. Table 4 11. Correlations between Five Suspected Covariates and Smoking Initiation Variable r Susceptibility to smoke freshman year .40 Alcohol consumption freshman year .20 Grade point average freshman year .08 Living with others (besides parents) freshman year .04 Parents smoking status .02 Research Question Results This sec tion describes the results of the multiple logistic regression analyses conducted to answer each of three research questions. Research Question One Does perceived prevalence of peers smoking behavior during freshman year of college increase the risk of s moking initiation during freshman or sophomore year of college? A multiple logistic regression analysis was conducted to examine the relationship between perceived prevalence of peers smoking behavior and smoking initiation during

PAGE 80

66 freshman or sophomore y ear of college. The approximate chi square statistic was used to assess model fit. A p value of greater than 0.05 indicates an adequate model fit (Hosmer & Lemeshow, 1980). Results of the chi square analysis confirmed an adequate fit (? 2 = 15.145, p = .056). Table 4 12 shows the results of the regression analysis. After controlling for the effects of two covariates (susceptibility to smoke during freshman year and alcohol consumption during freshman year), perceived prevalence of pee rs smoking behavior was not significantly related to smoking initiation during freshman or sophomore year of college ( p = .587). Table 4 12. Multiple Logistic Regression Analysis of Perceived Peer Smoking Behavior and Smoking Initiation Variable b SE Wald p value Perceived Peer Smoking .095 .176 .295 .587 Susceptibility 1.576 .204 59.735 < .001* Alcohol Consumption .507 .105 23.192 < .001* = Significant at the .05 level Research Question Two Does the smoking related behavior of friends and roommates during freshman year of college increase the risk of smoking initiation during freshman or sophomore year of college? A multiple logistic regression analysis was conducted to test for the significance of smoking related behavior of friends and roommates a nd smoking initiation during freshman or sophomore year of college. Smoking related behavior of friends and roommates, susceptibility to smoke during freshman year, and alcohol consumption during freshman year were included in the regression model. A chi square analysis confirmed an adequate model fit (? 2 = 14.031, p = .051). Results indicated that the

PAGE 81

67 smoking related behavior of friends and roommates was a statistically significant factor of smoking initiation during freshman and sophomore years of col lege ( p = .003). A logistic regression coefficient can be interpreted as the increase in the odds for each unit increase in the explanatory variable. For example, the coefficient ( b ) for friends and roommates smoking related behavior was 0.832. This value suggests that the odds of initiating smoking during freshman or sophomore year of college increases by e 0.832 = 2.298, or more than doubles, for each unit increase in the composite score of smoking related behavior of friends and roommates. Table 4 1 3 summarizes the results of this regression analysis. Table 4 13. Multiple Logistic Regression Analysis of Friends and Roommates Smoking Related Behavior and Smoking Initiation Variable b SE Wald p value Friends and Roommates Behavior .832 .281 8.740 .003* Susceptibility 1.380 .214 41.565 < .001* Alcohol Consumption .421 .110 14.600 < .001* = Significant at the .05 level Research Question Three Does sociability during freshman year of college increase the risk of smoking initiation during freshma n or sophomore year of college? A multiple logistic regression analysis was conducted to examine the relationship between the sociability and smoking initiation during freshman or sophomore year of college. The approximate chi square statistic was used to assess the fit of the model. Results indicated an ade quate model fit (.05 (? 2 = 11.928, p = .154). In addition to sociability, susceptibility and alcohol consumption were also included in the model as covariates. Table 4 14 shows that sociability failed to be a statistically significant factor of smoking i nitiation ( p = .411).

PAGE 82

68 Table 4 14. Multiple Logistic Regression Analysis of Sociability and Smoking Initiation Variable b SE Wald p value Sociability .801 .975 .675 .411 Susceptibility 1.613 .202 63.666 < .001* Alcohol Consumption .449 .131 11.764 .00 1* = Significant at the .05 level Regression Analysis of All Three Explanatory Variables Finally, a multiple logistic regression analysis was conducted to determine the relationship between all three explanatory variables and smoking initiation. To as sess the fit of the model, the approximate chi square statistic was used. The model fit the data as indicated by a p value of greater than .05 (? 2 = 10.671, p = .221). Susceptibility and alcohol consumption during freshman year were included as covariate s in the model. The smoking behavior of friends and roommates proved to be the only statistically significant explanatory variable ( p = .003). A logistic regression coefficient can be interpreted as the increase in the odds for each unit increase in the explanatory variable. The coefficient ( b ) for friends and roommates smoking related behavior in this model was 0.842. This value suggests that the odds of initiating smoking during freshman or sophomore year of college increases by e 0.842 = 2.321, or m ore than doubles, for each unit increase in the composite score of smoking related behavior of friends and roommates. Table 4 15 summarizes the results of this regression analysis. Table 4 15. Multiple Logistic Regression Analysis of Three Explanatory Var iables of Smoking Initiation Variable b SE Wald p value Perceived Prevalence of Peer Smoking .100 .177 .320 .572 Friends and Roommates Behavior .842 .283 8.855 .003* Sociability .808 .981 .678 .410 Susceptibility 1.363 .216 39.780 < .001* Alcohol Co nsumption .350 .135 6.659 < .001* = Significant at the .05 level

PAGE 83

69 Summary This chapter provided relevant descriptive results as well as findings for each of the three research questions. A profile of the sample and description of smoking behavior was pr esented, including 30 day smoking prevalence and t he percentage of college students who began smoking in college Correlations between each of five suspected covariates and smoking initiation were calculated using a predetermined inclusion criterion of 0 .10. Only two of these five variables susceptibility to smoke during freshman year and alcohol consumption during freshman year met this inclusion criterion. Consequently, these variables were included as covariates in all regression analyses. A separat e regression analysis was conducted to answer each of the three research questions. A final regression analysis was conducted that included all three explanatory variables and the two covariates. A discussion of the results will be provided in Chapter 5.

PAGE 84

70 CHAPTER 5 DISCUSSION This chapter has four purposes: (1) to discuss the studys results including current smoking behavior, smoking initiation behavior, and findings of the research questions, (2) to describe the studys strengths and weaknesses, (3) to pr esent conclusions, and, (4) to offer recommendations for future research and practice. Current Smoking Behavior Consistent with national surveys of college students, current smoking behavior was measured by asking participants whether they had smoked at a ny time during the 30 days before the survey. The 30 day smoking prevalence rate of 26.8% found in this study was consistent with the rates found in other national surveys (ACHA, 2002, Johnston et al., 2001, SAMHSA, 2002). Also consistent were findings t hat indicated similar smoking rates for males and females, and higher smoking rates for Whites, as opposed to Latinos or African American students. Unlike national surveys, however, this study distinguished between the number of participants who just puf fed on cigarettes during the 30 days before the study, and those who had smoked at least one whole cigarette during that time. Currently, national prevalence rates include individuals who have only puffed on cigarettes in the past 30 days. These individu als are typically labeled current smokers. Results from this study indicated that 6.3% of those who smoked on the 30 days before the study were puffers, whereas, 20.5% had smoked one or more cigarettes in the month prior to the survey.

PAGE 85

71 These data pro vide a more precise description of 30 day smoking behavior and are useful to practitioners who deliver smoking interventions. This study also determined the frequency of smoking and the quantity of cigarettes smoked, on average, on the days when smoking occurred. Participants who smoked one or more whole cigarettes during the previous 30 days were included in these analyses. Findings indicated that the majority of smokers smoked on relatively few days, and smoked relatively few cigarettes. For example, one out of five students smoked on one or two days during the month before the study, and close to half had smoked on ten or fewer days. In contrast, only one out of five smokers smoked daily. This daily rate is fairly consistent with recent Monitoring the Future reports showing that 19% of college smokers smoked daily in 1999, and 17.8% smoked daily in 2000. This study also determined the quantity of cigarettes smoked, on average, on the days that participants smoked. One out of four smokers smoked o nly one cigarette on the days that they smoked, and three out of four smoked five or fewer cigarettes. Moreover, only one out of ten students smoked 11 20 cigarettes, and only one participant out of the 255 smokers in the study reported smoking more than a pack on the days that he or she smoked. Collectively, the findings on the frequency of smoking and the quantity smoked suggest that many college students smoke moderately and occasionally. Therefore, smoking cessation efforts on college campuses shoul d not only target heavy, frequent smokers, but should also target occasional and moderate smokers. What motivates this latter population to quit may be inherently different from what motivates the more frequent, heavy smokers to quit. For example, using a financial appeal may not be an

PAGE 86

72 effective approach with these moderate smokers, because many may only spend around $10.00 a month on cigarettes. Consequently, college health practitioners should find other, more suitable, ways to appeal to these moderate smokers if they intend to motivate them to quit. Smoking Initiation Results Currently, 25 30% of college students in that nation smoke. This is more than twice the Healthy Campus 2010 goal of 12%. Reducing college smoking rates by the year 2010 can be a ccomplished in one of two ways: (1) either nonsmoking college students must remain smoke free, or (2) college students who already smoke must quit. This study focused on helping nonsmokers remain smoke free by investigating smoking initiation among colle ge students. Prior to this study, data were limited that addressed how many college students start smoking in college, when college students are most vulnerable to start smoking, and why college students start smoking. This information is essential not o nly for the development of effective smoking prevention programs, but also for the wise allocation of limited health promotion resources. How Many Students Start Smoking in College? Findings from this study are consistent with previous studies showing tha t the majority of smoking initiation among college students occurs by the age of 18 years. Among current smokers in this study, 90.1% had smoked their first whole cigarette when they were 18 years old or younger, and 9.9% had smoked their first cigarette when they were 19 years old or older. Wechsler and colleagues (1998) reported similar rates among the current smokers in their sample. They reported that 89% had smoked a whole cigarette for the first time at the age of 18 or younger, and 11% smoked the ir first cigarette at the age of 19 years or older.

PAGE 87

73 Collecting data by age, however, fails to address how many college students started smoking in college because many college students begin college when they are 18 years old. Findings from this study indicated that 7.3% of the sample had smoked their first whole cigarette during college. Moreover, 9.3% had first smoked on two or more days in any one month; 5.2% had first smoked on 20 or more days in any one month; and 4.3% had first smoked on 30 conse cutive days during college. These data are consistent with a study done at Yale University (Debernardo et al., 1999) showing a 5.3% smoking initiation rate among college students. Collectively, these data clearly indicate that fewer than 10% of college students begin smoking during college. Does this mean that we should abandon smoking prevention efforts on our college campuses? Not necessarily to assess this question more fairly, the data must be evaluated in a different light. As a whole, smoking in itiation rates ranging from 4.3% to 9.3% appear to appear rather insignificant, but when participants who never smoked are excluded from the analyses the picture changes. For example, 25.6% of students who had ever smoked on two or more days in one month did so for the first time during college. Moreover, one third of students who had ever smoked 20 days/month, and who had smoked on 30 consecutive days, had engaged in this behavior for the first time while in college. A similar picture emerges when curre nt smokers, rather than ever smokers, are the population of interest. Although 16% of the current smokers in the sample had first smoked a whole cigarette during college, 26% had first smoked on two days in any month during college. Moreover, one third o f the current smokers who had ever smoked on 20 or more days in any month, and one third of current smokers who had ever smoked

PAGE 88

74 on 30 consecutive days, first reached these smoking frequencies during college. These findings are consistent with a previous r eport by Wechsler and colleagues (1999). They found that 28% of current smokers had progressed to regular smoking at the age of 19 or older most likely when they were in college. Conversely, these smoking initiation percentages among current smokers ar e inconsistent with those reported by Debernardo and colleagues (1999). In their study of Yale University students, half of the current smokers reported starting to smoke in college. This discrepancy between current smokers initiation rates in this stud y and those of Yale students may be attributable to differences in the studies methodologies. The survey done at Yale University used an email format, whereas this study surveyed students in intact classrooms at five universities. More smokers who had b egun smoking in college may have responded to their survey than smokers who had begun smoking prior to college. Another possibility may be that Yale students may not be representative of students from other campuses, and that they have an unusually high p ercentage of students who start smoking after coming to college. When Do Most College Student Begin Smoking? Although the percentage of current smokers in this study who started smoking in college was inconsistent with the percentage in Debernardos study findings from both studies indicated that the majority of smoking initiation occurred during the first two years of college. Approximately 80% of the students at Yale University who reported starting to smoke in college began smoking during their freshma n or sophomore year. Results of this study confirmed these findings. For example, 89.4% of the students in this study who reported first smoking two or more cigarettes during college started this behavior during their first two years of college.

PAGE 89

75 Other descriptive results from this study further pinpoint when college student tend to begin smoking. These data clearly indicate that the first six months of college is an especially vulnerable time for potential smokers. Among students who had ever smoked on two or more days in any month during college, 56.9% had started smoking during their first six months. Furthermore, 60% had first smoked on 20 days during any one month, and 50% first smoked 30 consecutive days during this time. Result of smoking initi ation behavior among freshmen, sophomores, and juniors, corroborate that freshman year of college is a critical time for potential smokers. For example, 41% of juniors who first smoked on two days during any one month during college reached this frequency of smoking during their first six months of college, and 54% had reached this frequency by the end of their first year of college. Moreover, a total of 72% had first started smoking this often by the end of their second year. The implications of these f indings by year in school further support a recommendation to first target prevention efforts to freshmen, and then to sophomores. Although students who begin smoking in college represent approximately 7 10% of all college students, separating smoking init iation behavior by time in college is useful for several reasons. First, these data pinpoint when the majority of smoking initiation and progression occurs. Second, practitioners can use these results to time their prevention efforts and to wisely use li mited resources. Third, these data may help practitioners tailor prevention messages to at risk audiences, primarily freshmen and sophomores. Correlations Between Covariates and Smoking Initiation So far, this chapter has addressed current smoking behavio r and smoking initiation results. The percentage of college students in the study who currently smoke, how often, and how much were discussed, and implications were described. Smoking initiation

PAGE 90

76 results revealed that a relatively small percentage of coll ege students begin smoking in college, but much of this behavior occurs during the first two years of college. The first six months of college appears to be a particularly vulnerable time for students. This chapter will now focus on possible reasons why s ome college students begin to smoke in college. This section will address the results of the associations between five suspected covariates and the outcome variable (grade point average, parents smoking status, susceptibility to smoke, alcohol consumptio n, and living situation). These five variables were not of primary interest in the study; rather, they were measured to be used as control variables if necessary. Four of the five variables had empirical support that linked them to smoking initiation be havior primarily among a high school population. Three of these variables (grade point average, parents smoking status, and susceptibility to smoke) have been identified as predicators of smoking initiation among adolescents. Binge drinking during colle ge has been associated with college smoking status. The fifth variable, living situation, did not have empirical support for smoking initiation, however, data were collected showing whether a participant lived alone or with others because of the social na ture of the three explanatory variables. Of these five potential covariates, only susceptibility to begin smoking during freshman year ( r =0.40), and alcohol consumption during freshman year ( r =0.20), were moderately correlated with smoking initiation. The magnitude of the correlations between the other three suspected covariates (parents smoking status, grade point average, and living situation) and smoking initiation failed to reach the inclusion criterion of .10 or greater. Two explanations may be offered as to why parents smoking status

PAGE 91

77 was not influential among this sample. First, findings from the adolescent literature indicate that parents smoking status is more influential in smoking initiation during early adolescence versus late adolescen ce. Therefore, if parents were to influence college students smoking initiation behavior, this most likely would have occurred prior to college. Second, over 95% of the participants did not live with their parents during their freshman year of college, and hence, their parents probably had minimal influence over this behavior. Like with parental influence, grade point average was included as a potential covariate because of evidence showing a strong association among high school student smoking initiat ion behavior. One possible reason why GPA was not associated with smoking initiation in this study may be due to a lack of variability among college GPAs when compared to high school GPAs. In this study, 87% of participants reported a first semester GPA of 2.5 or higher; and 65% reported a first semester GPA of 3.0 or higher. The lack of spread in the GPA scores among this sample could have undoubtedly affected this result. Likewise, a lack of variability of participants living situation probably infl uenced the results of the correlation between this variable and the outcome variable because nine out of ten participants lived with other students during their freshman year of college. Research Question Results The three explanatory variables examined i n this study were chosen based on social cognitive theory and empirical support showing an association with cigarette smoking initiation among middle and high school populations. Multiple logistic regression analysis was used to answer each of the three research questions. Only one of

PAGE 92

78 the three explanatory variables proved to be significant at a 0.05 level. The results for each question are discussed below. Research Question One Does the perceived prevalence of peers smoking behavior during freshman y ear of college increase the risk of smoking initiation during the first two years of college? Perceived prevalence of peers smoking behavior was categorized as a personal determinant within the social cognitive theory framework. Furthermore, a fair amoun t of evidence was presented linking this variable to smoking initiation among adolescent populations. Although perceived prevalence of peers smoking behavior has both theoretical and empirical support, the results from this study showed that it was not a significant factor of smoking initiation among college students. The empirical support for this variable comes predominantly from children and adolescent populations. No study thus far has linked this variable influences smoking initiation behavior amo ng a college population. Although much evidence exists showing that college students overestimate the percentage of college students who smoke, none has attempted to use inferential statistics to link this misperception with smoking initiation behavior. Thus one plausible explanation for a lack of significance for this variable among a college population is that this finding is valid. In contrast, another possible explanation may be that this result is not valid because this variable could have been inacc urately measured. Of all items on the 77 item survey, students expressed the most confusion about items measuring perceived prevalence of peers smoking behavior. Several students written comments during instrument development expressed difficulty in g uessing the number of peers who smoked during freshman year or questioned the accuracy of their estimations. So although the reliability

PAGE 93

79 coefficient of this scale was adequate (0.88), future investigation of how to best measure this variable may be warr anted. Research Question Two Does the smoking related behavior of friends and roommates during freshman year of college increase the risk of smoking initiation during the first two years of college? Findings from a multiple logistic regression analysis indicated that the smoking related behavior of friends and roommates was a significant factor of smoking initiation during the first two years of college. This relationship remained even when controlling for susceptibility to smoke and for alcohol consump tion during freshman year. Due to the strong theoretical and research support for the influence of this variable on smoking initiation among high school students, this finding was not surprising. Smoking related behavior of friends and roommates had the strongest theoretical support among the three explanatory variables. It was classified not only as an environmental determinant of smoking initiation within social cognitive theory, but it also incorporated two key concepts within the theory observationa l learning and direct reinforcement. Seeing others smoke is a form of observational learning, or modeling, and according to social cognitive theory, increases the likelihood of similar behavior. The smoking status for four types of relationships (friends roommates, sorority sisters/fraternity brothers, and romantic partners) was part of how this variable was measured, and the score increased as the number of friends who smoked increased. In addition to observational learning, this variable also incorp orated the concept of direct reinforcement by measuring friends offers and encouragement to begin smoking. Like observational learning, as direct reinforcement increases for a particular behavior, the likelihood of the behavior occurring or continuing al so increases. Eight out of the

PAGE 94

80 twelve items on this scale measured either friends offers to smoke or friends encouragement to smoke. Hence, a participant who had received many offers and encouragement to smoke would have a higher composite score for th is variable than a participant who had received relatively few offers to smoke. In addition to strong theoretical support, friends smoking related behavior had the strongest empirical support among the three explanatory variables. Longitudinal studies of high school students provide strong evidence linking friends behavior and smoking initiation. Furthermore, in a small, cross sectional study of college students, participants who were susceptible to smoke in the upcoming year had indicated that friend s smoking behavior might motivate them to start smoking. They selected this as a potential reason to begin smoking more frequently than image, stress level, depression, or expression of independence. This finding confirms that smoking initiation is, at least in part, an interpersonal phenomenon, and therefore, prevention efforts should address interpersonal communication issues embedded within the behavior. Rather than focusing on the health consequences or dangers or costs of smoking, interventions sho uld concentrate on the development of resistance or refusal skills. Health communication theory and planning models would be especially useful in developing these types of interventions. For example, a mass communication theme like Dont let your friend s get you hooked on cigarettes could be tested on a college population for appeal and effectiveness. Research Question Three Does sociability during freshman year of college increase the risk of smoking initiation during the first two years of college?

PAGE 95

81 A limited number of studies provide evidence that sociability influences smoking initiation among young people. Two longitudinal studies found extroversion to be positively associated with smoking initiation among adolescents of both genders; whereas, an other found sociability to be predictive of smoking initiation only among adolescent girls. A study of male and female college students indicated that extroversion was positively associated with smoking initiation. Although the association between socia bility and smoking initiation has modest empirical support among adolescents, the results from this study indicated that it is not a significant factor of smoking initiation among college students. Two explanations may be plausible for this finding. Firs t, the finding may be valid. Of the three explanatory variables in the study, sociability has the weakest amount of evidence linking it to smoking initiation. Furthermore, these findings are mixed about its influence for each gender. One study showed a significant association for both genders, whereas another found such an association only among girls. Second, this finding may be invalid due to a lack of variability of sociability scores for this population. The participants, as a whole, were a fairly social which is not surprising considering the nature of undergraduate student life and culture. Further investigation of this variable among college students is needed to corroborate this particular finding. Strengths and Weaknesses The purpose of this study was to investigate cigarette smoking initiation behavior among a college population. Prior to this investigation, data was limited and questions remained about the size of this problem on college campuses. Furthermore, little was known about the re asons why college students might start to smoke during their college years. Like any study, this investigation had both strengths and weaknesses.

PAGE 96

82 One strength of this study was expanding the research knowledge in this area by providing information on ho w many students start smoking in college and when they are most vulnerable to begin. Data for smoking initiation were collected by age as well as point of time in college. Findings by age were consistent with those found in a large, national study. Resu lts also confirmed a previous report indicating that smoking initiation is most likely to occur during the first two years of college. This study suggests that the first year of college is an especially vulnerable time to start smoking. These results hav e practical implications for anti smoking prevention efforts. Health practitioners can apply them to at risk groups and can strategically time the delivery of anti smoking interventions. Given that health promotion resources are limited, this information is valuable. Another contribution of this study was investigating several reasons why college students may begin to smoke. Results suggest that sociability, perceptions of peers smoking behavior, GPA, living situation, and parents smoking status are no t significant factors in starting to smoke in college. In contrast, friends behavior, susceptibility to smoke during freshman year, and alcohol consumption during freshman year were significantly associated with starting to smoke during freshman or sopho more year of college. These findings serve as a starting point for more research on this issue and for application to current campus smoking prevention efforts. Although this study had several strengths, its findings must be interpreted in the context o f several limitations. First, the study was based on self reported data and required participants to recall information from their past. The extent to which recall bias influenced these findings is unknown. Second, the data are cross sectional, and ther efore,

PAGE 97

83 certainty about the direction of causation between the significant explanatory variable and the outcome variable cannot be confirmed. What cannot be determined from these data are whether persons who initiated smoking in college did so because thei r friends behavior influenced them, or whether these smoking initiators sought friends and roommates who tended to smoke, and subsequently, offered and encouraged them also smoke. Third, the sample was a convenience sample, rather than a random sample. This limits the studys findings to its sample and not a broader college population. Conclusions This study examined whether three social factors affected cigarette smoking initiation among college undergraduate students: (1) perceived prevalence of p eers smoking behavior, (2) smoking related behavior of friends and roommates, and (3) sociability. Moreover, this study determined the percentage of college students who began smoking in college, identified at risk populations, and pinpointed when colleg e students were most vulnerable to begin smoking. Results from this study indicated that 26.8% of participants smoked during the 30 days preceding the survey. Unfortunately, this falls far short of the Healthy Campus 2010 goal of 12%, and consequently, c onfirms the need for anti smoking efforts on college campuses. A total of 9.3% of all participants began smoking cigarettes in college, with many starting to smoke, and progressing in frequency, during their first two years of college. These data indica te that the first six months are a particularly vulnerable time for potential smokers. The specificity of these findings can help college health practitioners implement their smoking prevention efforts at the most appropriate time during the academic year

PAGE 98

84 Only two of the five suspected covariates were moderately correlated with smoking initiation during the first two years of college. These covariates included susceptibility to smoke during freshman year and alcohol consumption during freshman year. In contrast, grade point average during freshman year, parents smoking status prior to freshman year, and students living situation during freshman year failed to be moderately correlated with smoking initiation during the first two years of college. Alth ough susceptibility and alcohol consumption were not the variables of interest for this study, their relationship to smoking initiation warrants further research. Of the three explanatory variables of this study, smoking related behavior of friends and ro ommates was the only significant factor of cigarette smoking initiation during college. This relationship remained even after controlling for the other two explanatory variables and the two covariates. In contrast, perceived prevalence of peers smoking behavior, as well as sociability, were not significantly associated with smoking initiation during the first two years of college. Recommendations Although this study had some limitations, its findings can guide future practice and research. The follow ing recommendations are offered to help practitioners implement anti smoking initiatives for college students and for researchers who are interested in investigating cigarette smoking initiation among a college population. Recommendations for College Healt h Practice Based on this studys findings, the following recommendations are provided for college health practice. 1. Develop anti smoking interventions specifically (a) tailored to susceptible freshmen, (b) designed to address friends behaviors and influen ce on the smoking initiation process, (c) designed to help susceptible freshmen make a firm

PAGE 99

85 commitment not to smoke, and (d) implemented within six months from when the majority of freshmen begin college, typically during fall semester. 2. Address the miscon ception that once individuals graduate from high school, they are no longer vulnerable to start smoking. 3. Use The College Student Smoking Survey in part or in whole, to identify the number and characteristics of college students who are likely to begin s moking. 4. Apply the findings from this study to inform decisions about the relative importance of this health issue on college campuses and the amount of resources to allocate to smoking prevention. 5. Create health interventions and messages that address the relationship between cigarette smoking initiation and alcohol consumption. Recommendations for Future Research In addition to recommendations for college health practices, the following recommendations are provided for research on smoking initiation among undergraduate college students. 1. Replicate this study using a longitudinal research design and a random sample. 2. Conduct both quantitative and qualitative studies about the nature of friends offers and encouragement to smoke. 3. Conduct research on the effecti veness of an anti smoking interventions targeted to, and implement for, traditional aged freshmen during their first six months of college. 4. Investigate the nature of alcohol consumption and smoking initiation during the first semester of college. 5. Conduct quantitative and qualitative research on susceptibility to start smoking during freshman year of college. This chapter focused on a discussion of a multi campus study designed to determine how many traditional aged, undergraduate students start smoking ci garettes in college, when they are most vulnerable to begin, and whether three social factors put them at greater risk to start smoking in college. These findings expand the limited body of literature in the area of smoking initiation among college studen ts and should be

PAGE 100

86 corroborated by future research. In the mean time, however, college health practitioners can use these findings to guide the development and delivery of their smoking prevention efforts. Such efforts could help at risk students remain sm oke free during their college years, and ultimately, could help to achieve the Healthy Campus 2010 goal of reducing the percentage of college smokers to 12% or less by the year 2010.

PAGE 101

87 APPENDIX A LIST OF EXPERT PANEL MEMBERS 1. Dr. Jill Varnes Professor, Department of Health Science Education University of Florida PO Box 118210 Gainesville, FL 32611 8210 2. Dr. William Chen Department Chair, Department of Health Science Education U niversity of Florida PO Box 118210 Gainesville, FL 32611 8210 3. Dr. Sadie Sanders Visiting Professor, Department of Health Science Education University of Florida PO Box 118210 Gainesville, FL 32611 8210 4. Dr. Phillip Barkley Director, Student Heal th Care Center University of Florida PO Box 117500 Gainesville, FL 32611 7500 5. Joanne Auth Former Coordinator for Education/Training Programs Student Health Care Center University of Florida PO Box 117500 Gainesville, FL 32611 7500 6. Dr. Jane Jone s Professor, School of Health Promotion and Human Development University of Wisconsin Stevens Point Stevens Point, WI 54481

PAGE 102

APPENDIX B INSTITUTIONAL REVIEW BOARD APPLICATIONS AND APP ROVAL LETTERS

PAGE 103

89

PAGE 104

90

PAGE 105

91

PAGE 106

92

PAGE 107

93

PAGE 108

94

PAGE 109

95

PAGE 110

96

PAGE 111

97

PAGE 112

98

PAGE 113

99

PAGE 114

100

PAGE 115

101

PAGE 116

102

PAGE 117

103

PAGE 118

104

PAGE 119

105

PAGE 120

106

PAGE 121

107

PAGE 122

108

PAGE 123

109

PAGE 124

110

PAGE 125

111

PAGE 126

112

PAGE 127

113

PAGE 128

114

PAGE 129

115

PAGE 130

116

PAGE 131

117

PAGE 132

118

PAGE 133

APPENDIX C PARTICIPANT CONSENT LETTERS

PAGE 134

120 October 7, 2002 Dear UF Student: I am a doctoral student at the University of Florida. As part of my dissertation research, I am recruiting first and second year undergraduate students, at least 18 years old, t o help me create a college student smoking survey. I would like both smokers and non smokers to give input. Participants will help create a well designed survey that will be distributed on five college campuses in January. The purpose of the final study will be to learn about college students smoking patterns. The results will guide smoking prevention and cessation programs. Your willingness to give input on a draft of the survey is not part of your grade for this class and is voluntary. Your decision w ill not affect your grade in any way. You are free to withdraw your consent and discontinue participation at any time without consequence. Your participation will take about 30 minutes. You do not have to answer any question you do not wish to answer. P articipation involves no anticipated risks and your responses will be anonymous. If you have any questions, please contact me at emmeree@ufl.edu or (352) 392 0583, ext. 1409. You may also contact my supervisor, Dr. Jill Varnes, at varnes@aa.ufl.edu or (3 52) 392 0583 ext. 1309 or 392 2404. If you have any questions regarding your rights as a research participant, please contact the University of Florida Institutional Review Board, University of Florida, Box 112250, Gainesville, FL 32611 2250; phone (352) 392 0433. If you are willing to participate, please sign this letter and return it to your instructor, Terri Mitchell, by the end of class on Oct. 9. (A second copy of this letter is provided for your records). If more than 10 students agree to partic ipate, I will randomly select 10 to give input. If you are selected, I will email you a draft of the survey with detailed instructions. To ensure your anonymity, do not return any survey materials to me via email and do not put your name anywhere on the m aterials. Return them to Terri Mitchell by Monday, Oct. 14. She will give you a 45 minute long distance phone card for your participation. If you choose to withdraw from the study at any time, inform Ms. Mitchell. She will still give you your phone card Final study results will also be provided if you request them below. Thanks for your consideration! Jane Emmere, M.Ed., C.H.E.S. Project Director I have read the procedure described above for the College Student Smoking Survey. I attest that I am at least 18 years old, voluntarily agree to participate, and have received a copy of this project description. _______________________________ ___________________________ ________ Participants Name (Print) Participants Signature Date Email Address: (Print neatly) ____________________________________________ Number of semesters at UF (post high school): ______________________________ Check here if you would like to receive a copy of the final manuscript.

PAGE 135

121 Dear Student: I am a doctoral s tudent at the University of Florida working on my dissertation research. You can help me learn more about college students smoking behavior by completing a short survey as a part of a multi campus research project. The purpose of this study is to find o ut whether social factors influence college students smoking behavior. Although we have a great deal of reliable information about the smoking behavior of high school students, information about college students is lacking. Should you choose to participa te, the information you provide will assist health professionals develop smoking prevention and cessation services on college campuses. This study is not part of your grade for this class. Your instructor has agreed to set aside 15 minutes of class time for those who choose to participate. Your participation is voluntary and your decision will not affect your grade in this class in any way. You are free to withdraw your consent and discontinue participation at any time without consequence. The survey wi ll take 10 15 minutes. You do not have to answer any question you do not wish to answer. Participation involves no anticipated risks and your responses will be anonymous. If you choose not to participate, please sit quietly while others complete the surve y. If you have any questions, please contact me at emmeree@ufl.edu or (352) 392 0583, ext. 1409. You may also contact my supervisor, Dr. Jill Varnes, at varnes@aa.ufl.edu or (352) 392 0583 ext. 1309 or 392 2404. If you have any questions regarding yo ur rights as a research participant, please contact the University of Florida Institutional Review Board, University of Florida, Box 112250, Gainesville, FL 32611 2250; phone 352 392 0433. If you intend to participate, please sign this letter. By signi ng below, you give me permission to report your anonymous responses as aggregated data in the final manuscript that I will submit to my faculty supervisory committee. To ensure your anonymity, give your instructor this letter before you fill out a questio nnaire and do not put your name anywhere on the questionnaire or the scan tron form. A second copy of this letter is provided for your records. Study results will be provided to all participants who request them. Thanks for your consideration! Jane Em mere, M.Ed., C.H.E.S. Project Director I have read the procedure described above for the College Student Smoking Survey. I attest that I am at least 18 years old, voluntarily agree to participate, and have received a copy of this project description. _________________________________ ______________ Participants Signature Date If you would like to receive a copy of the final manuscript submitted to the instructor, neatly PRINT your name and either an email address or a complete mailing address bel ow:

PAGE 136

122 Social Factors of Cigarette Smoking Initiation Among College Undergraduate Students The purpose of this research is to find out whether social factors influence college students cigarette smoking initiation behavior. Very little research has been done about the reasons why individuals begin to smoke cigarettes while attending college. Findings from this study may lead to information that will assist health professionals create smoking prevention programs on college campuses. For this project, you will be asked to fill out a questionnaire about cigarette smoking behavior (or your intentions to smoke if you do not smoke). This study is not part of your grade for this class. Your participation is voluntary and you must be at least 18 years old to pa rticipate. Refusal to participate will not affect your grade in this class in any way. You are free to withdraw your consent and discontinue participation at any time for any reason without penalty or prejudice from the investigator. The survey will take about 15 minutes. Your responses will remain anonymous, and you do not have to answer any question you do not wish to answer. Do not put your name or social security number anywhere on the survey. The foreseeable risks or ill effects from participating in this study are minimal. No compensation or benefits for your participation will be provided, other than study results will be provided if you request them below. If you choose not to participate, please sit quietly while others complete the survey. If you have questions about your rights as a research participant, please contact Ms. Sandra Smith, Coordinator of Research Compliance, Office of Academic Research and Sponsored Programs, Ball State University, Muncie, IN 47306, (765) 285 5070. If you int end to participate, please sign one of the copies of this letter and keep the second copy for your records. To ensure your anonymity, the signed copy will be collected separately from your answer form. By signing below, you give the principal investigator permission to report your anonymous responses as aggregated data in a final manuscript to be submitted to my faculty supervisory committee. I, ________________________, agree to participate in this research project titled Social Factors of Cigarette S moking Among College Undergraduate Students. I have read and understand the explanation provided to me. I have had all of my questions answered to my satisfaction, and I voluntarily agree to participate in this study. I attest that I am at least 18 yea rs old and have been given a copy of this consent form. ( O v e r )

PAGE 137

123 _______________________________ ________ Participants Signature Date If you would like to receive a copy of the final results of this study, neatly PRINT your name and either an email address or a complete mailing address below: Principal Investigator: Jane Emmere, Ph.D. Candidate Faculty Sponsor: Department of Health Science Education Dr. Rebecca Brey, Associate Professor University of Florida Dept. of Physiology and Health Science Gainesville, FL 32611 8210 Telephone: (352) 392 05893 Email: emmeree@ufl.edu Edition Date: January 2003 Faculty Sponsor: Dr. Rebecca Brey, Associate Professor Department of Physiology and Health Science Ball State University Muncie, IN 47306 Telephone: (765) 2 85 5961 Email: rbrey@bsu.edu

PAGE 138

124 Information Sheet Social Factors of Cigarette Smoking Initiation Among College Undergraduate Students I am a doctoral student at the University of Florida working on my dissertation research. You can help me learn more about college students cigarette smoking behavior by completing a short survey as a part of a multi campus research project. The purpose of this study is to find out whether social factors influence college students smoking initiation behavior. Should you choose to participate, the information you provide will assist health professionals with smoking prevention services on college campuses. Approximately 1,200 college students from five universities are expected to participate in this project. You were chosen to participate because you are enrolled in Pol Sci 207, however, this study is not part of your grade in this class. Your class was se lected because it is a general education elective at TAMU and historically contains male and female undergraduate students of a variety of races and academic majors. This study is not part of your grade for this class. Your instructor has agreed to set a side 15 minutes of class time for those who choose to participate. Your participation is voluntary and your decision will not affect your grade in this class in any way. You are free to withdraw your consent and discontinue participation at any time witho ut penalty or prejudice. You do not have to answer any question you do not wish to answer. The survey is anonymous and will take about 15 minutes. Do not put your name or social security number anywhere on the survey. You must be at least 18 years old to participate. The foreseeable risks or ill effects from participating in this study are minimal. There is no compensation or benefits for your participation, however, study results will be provided to any participant or non participant who emails me at emmeree@ufl.edu If you choose not to participate, or you are under 18 years old, please sit quietly while others complete the survey. You are welcome to ask the survey staff questions any time during the survey process. If you have any questions after the survey has been administered, please contact me at PO Box 118210, Gainesville, FL 32611 8210, (352) 392 0583, ext. 1409, emmeree@ufl.edu You may also contact Dr. Steve Dorman, my research supervisor at TAMU at the Department of Health and Kinesiol ogy, College Station, TX 77843, sdorman@hlkn.tamu.edu (979) 845 3124. This research study has been reviewed and approved by the Institutional Review Board Human Subjects in Research at Texas A&M University. For research related problems or questions reg arding subjects rights, contact the Institutional Review Board through Dr. Michael W. Buckley, Director of Support Services, Office of the Vice President for Research at (979) 458 4067. Thank you for your consideration! Jane Emmere, M.Ed., C.H.E.S. P rincipal Investigator Date: _____________________

PAGE 139

125 COLLEGE STUDENT SMOKING SURVEY CONSENT FORM I, _________________________________, agree to participate in the research study titled "Social Factors of Cigarette Smoking Initiation Among College Student s" conducted by Jane Emmere from the Department of Health Science Education at the University of Florida under the direction of Dr. Stuart Fors, Department of Health Promotion and Behavior at the University of Georgia, (706) 542 4365. I understand that I do not have to take part in this study and that I can stop taking part at any time without giving any reason, and without penalty. My participation or non participation in this survey will have no impact on my grades. I can ask to have all of the informati on about me returned to me, removed from the research records, or destroyed. The reason for this study is to find out how many college students started smoking cigarettes before and after college and possible reasons why they start in college. If I cho ose to participate, I will be one of approximately 1,200 other college students from five universities who are expected to participate. I will not benefit directly from this research, however, participation in this study may lead to information that will help health professionals create smoking prevention programs for college students. I can request a copy of the results of this study by providing my email and/or my U.S. mailing address at the bottom of this form. If I volunteer to take part in this stud y, I will be asked to do the following things: 1) Sign, date, and return this consent form separately from my answer form. 2) Refrain from putting my name or any other personal identifier on my survey or answer form. 3) Answer questions about cigarette smoking be havior that will take 10 15 minutes. No risk is expected. Some questions may cause mild anxiety, however, this is rare. I understand that I do not have to answer any question that causes me discomfort and that there is no penalty for leaving any questio n blank. I can choose to withdraw my participation in this research at any time without penalty or prejudice. My responses and the results of this participation will be anonymous. The only persons who will have access to my anonymous data are the prim ary investigators graduate faculty committee members or faculty advisors. The survey staff will answer any further questions about the research study now or while I take the survey. The primary investigator can be reached by telephone at (352) 392 0583, or by email at emmeree@ufl.edu. I understand that I am agreeing by my signature on this form to take part in this research project and understand that I will receive a signed copy of this consent form for my records. I attest that I am at least 18 years old. Jane Emmere _______________________ __________ Name of Researcher Signature Date Telephone: (352) 392 0583, ext. 1409 Email: emmeree@ufl.edu ___________ ____________ __________ Name of Participant Signature Date I would like a copy of the results of this research. Please neatly print your name and provide an email address and/or a U.S. mailing address below: Additional questions or problems regard ing your rights as a research participant should be addressed to Chris A. Joseph, Ph.D. Human Subjects Office, University of Georgia, 606A Boyd Graduate Studies Research Center, Athens, Georgia 30602 7411; Telephone (706) 542 3199; E Mail Address IRB@uga.e du.

PAGE 140

126 Spring 2003 Dear Student: I am a doctoral student at the University of Florida working on my dissertation research. You can help me learn more about college students cigarette smoking behavior by completing a short survey as a part of a multi cam pus research project involving over 1,000 participants. The purpose of this study is to find out whether social factors influence college students smoking initiation behavior. Should you choose to participate, the information you provide will assist heal th professionals with smoking prevention services on college campuses. This study is not part of your grade for this class. Your instructor has agreed to set aside 15 minutes of class time for those who choose to participate. Your participation is volunt ary and you must be at least 18 years old to participate. Refusal to participate will not affect your grade in this class in any way. You are free to withdraw your consent and discontinue participation at any time without penalty. The survey will take ab out 15 minutes. Your responses will remain anonymous, and you do not have to answer any question you do not wish to answer. Do not put your name or social security number anywhere on the survey. Participation involves no more than minimal risk. No compen sation or benefits for your participation will be provided, other than study results will be provided to anybody who requests them below. If you choose not to participate, please sit quietly while others complete the survey. If you have any questions, please contact me at emmeree@ufl.edu University of Florida, PO Box 118210, Gainesville, FL 32611 8210or (352) 392 0583, ext. 1409. You may also contact my faculty advisor at the University of Alabama, Dr. James Eddy, Chair of the Department of Health S cience at jeddy@bama.ua.edu or (205) 348 2956. If you intend to participate, please sign one of the copies of this letter and keep the second copy for your records. To ensure your anonymity, the signed copy will be collected separately from your answer fo rm. By signing below, you give me permission to report your anonymous responses as aggregated data in a final manuscript to be submitted to my faculty supervisory committee. Thanks for your consideration! Jane Emmere, M.Ed., C.H.E.S. Principal Invest igator I understand that this research study has been reviewed and approved by the Institutional Review Board at the University of Alabama. For research related problems or questions regarding participants rights, I can contact the Chair of the Institu tional Review Board at the University of Alabama at (205) 348 5152. I have read and understand the explanation provided to me. I have had all of my questions answered to my satisfaction, and I voluntarily agree to participate in this study. I attest that I am at least 18 years old and have been given a copy of this consent form. _____________________ _________ Participants Signature Date If you would like to receive a copy of the final results of this study, neatly PRINT your name and either an ema il address or a complete mailing address below:

PAGE 141

APPENDIX D THE COLLEGE STUDENT SMOKING SURVEY

PAGE 142

128

PAGE 143

129

PAGE 144

130

PAGE 145

131

PAGE 146

132

PAGE 147

133

PAGE 148

134

PAGE 149

135

PAGE 150

13 6

PAGE 151

137 APPENDIX E FINAL SURVEY SCALE A NALYSES Table E 1. Reliability Analysis for Perceived Prevalence of Peers Smoking Behavior Scale Item M SD Tried 1.294 1.118 Whole 1.122 .993 First Six Months 1.129 .995 Second Six Months 1.143 1.054 Socialize 1.800 1.0 79 Alcohol Consumption 1.904 1.091 Mildly Intoxicated 1.576 1.006 Moderately Intoxicated 1.714 1.084 Extremely Intoxicated 1.881 1.307 Number of items: 9 Number of response choices: 5 Composite data: n = 1,224 Mean = 13.561 Variance = 48.688 SD = 6.9 78 Table E 2. Item Total Statistics for Perceived Prevalence of Peers Smoking Behavior Scale Item Corrected Item Total Correlation Alpha if Item Deleted Tried .549 .874 Whole .590 .871 First Six Months .590 .871 Second Six Months .569 .872 Socializ e .572 .872 Alcohol Consumption .687 .862 Mildly Intoxicated .705 .861 Moderately Intoxicated .719 .859 Extremely Intoxicated .667 .865 n = 1,224 Coefficient Alpha = .880

PAGE 152

138 Table E 3. Reliability Analysis for Smoking Related Behavior of Friends and R oommates Scale Item M SD Friends .693 .738 Partner .398 .701 Roommate .481 .830 Greek .746 .722 Offer best friend .664 .947 Offer roommate .414 .854 Offer Greek .640 .923 Offer partner .368 .819 Encourage best friend .278 .623 Encourage roommate .193 .562 Encourage Greek .229 .610 Encourage partner .165 .519 Number of items: 12 Number of response choices: 4 Composite data: n = 503 Mean = 5.264 Variance = 42.091 SD = 6.488 Table E 4. Item Total Statistics for Smoking Related Behavior of Frien ds and Roommates Scale Item Corrected Item Total Correlation Alpha if Item Deleted Friends .693 .901 Partner .588 .912 Roommate .611 .912 Greek .594 .912 Offer best friend .753 .905 Offer roommate .715 .907 Offer Greek .734 .906 Offer partner .729 .906 Encourage best friend .682 .909 Encourage roommate .598 .912 Encourage Greek .643 .910 Encourage partner .661 .911 n = 503 Coefficient Alpha = .916

PAGE 153

139 Table E 5. Reliability Analysis for Sociability Scale Item M SD Importance .763 .257 Perceptio n .703 .263 Bars .511 .378 Party .588 .361 Free time .742 .275 Ease .761 .245 Alone .744 .278 Frequency socialize .657 .263 Frequency party .395 .261 Number of items: 9 Number of response choices: 4 Composite data: n = 1,231 Mean = 5.863 Variance = 2.956 SD = 1.719 Table E 6. Item Total Statistics for Sociability Scale Item Corrected Item Total Correlation Alpha if Item Deleted Importance .666 .808 Perception .516 .823 Bars .524 .826 Party .669 .805 Free time .560 .818 Ease .526 .822 Alon e .401 .835 Frequency socialize .570 .817 Frequency party .546 .820 n = 1,231 Coefficient Alpha = .836

PAGE 154

140 Table E 7. Reliability Analysis for Susceptibility to Smoke Scale Item M SD Upcoming year .318 .613 Same gender .261 .520 Opposite gender .272 .5 34 Boy/girlfriend .246 .477 Date .251 .493 Roommate .240 .484 Greek .262 .533 Mildly Intoxicated .461 .745 Moderately Intoxicated .523 .792 Extremely Intoxicated .611 .836 Number of items: 10 Number of response choices: 4 Composite data: n = 622 M ean = 3.44 Variance = 26.65 SD = 5.162 Table E 8. Item Total Statistics for Susceptibility to Smoke Scale Item Corrected Item Total Correlation Alpha if Item Deleted Upcoming year .813 .947 Same gender .870 .946 Opposite gender .868 .945 Boy/girlfrie nd .820 .948 Date .803 .948 Roommate .875 .946 Greek .830 .947 Mildly Intoxicated .832 .947 Moderately Intoxicated .814 .949 Extremely Intoxicated .711 .955 n = 622 Coefficient Alpha = .953

PAGE 155

141 Table E 9. Reliability Analysis for Alcohol Consumption Scale Item M SD Drinking Frequency 1.750 1.588 Moderately Intoxicated 1.480 1.445 Extremely Intoxicated .817 1.126 Frequency Five or More 1.216 1.470 Number of items: 4 Number of response choices: 6 Composite data: n = 1,233 Mean = 5.261 Variance = 27.753 SD = 5.268 Table E 10. Item Total Statistics for Alcohol Consumption Scale Item Corrected Item Total Correlation Alpha if Item Deleted Drinking Frequency .897 .929 Moderately Intoxicated .912 .918 Extremely Intoxicated .824 .951 Frequency Five or More .896 .923 n = 1,233 Coefficient Alpha = .947

PAGE 156

142 APPENDIX F LIST OF FACULTY ADVI SORS 1. Ball State University Dr. Rebecca A. Brey Associate Professor Department of Physiology and Health Science Muncie, IN 47306 765.285.3758 rbrey@bsu.edu 2. Texas A & M University Dr. Steve Dorman Professor and Dept. Chair Department of Health and Kinesiology College Station, TX 77843 979.845.3124 sdorman@tamu.edu 3. University of Alabama Dr. James Eddy Department Chair Department of Health Science 200A Foster Tuscaloosa, AL 35487 0311 205.348.2956 jeddy@bama.ua.e du 4. University of Florida Dr. Jill Varnes Professor Department of Health Science Education PO Box 118210 Gainesville, FL 32611 8210 352.392.0583 varnes@aa.ufl.edu 5. University of Georgia Dr. Stuart Fors Department Chair Department of Health Promo tion and Behavior Athens, GA 30602 6522 706.542.4365 stufors@arches.uga.edu

PAGE 157

143 APPENDIX G DATA COLLECTION PROTOCOL Data Collection Instructions 1. Read the script out loud a couple times before the class. 2. Have your assistant(s) distribute the surveys while you read the script. 3. If participants have questions about the wording or resp onse choices of an item, do not try to interpret the question for them, just say: Select the best response for you, or leave it blank if you are unsure about it. 4. Thank the students for their participation. 5. Return all completed surveys to me as soon as a ll surveys at your school have been completed (you can recycle any extra surveys).

PAGE 158

144 Data Collection Script: Hello. Im _______________________ and am helping a Ph.D. candidate from the University of Florida with her dissertation research. We are passing out a packet with an information sheet explaining of her study and a short survey. Please do not begin the survey until I go over the instructions with the entire class. Itll only take me a minute. The principal investigator of this study is Jane Emme re (pronounced Em er ay). Her study examines why people begin to smoke cigarettes in college. Jane needs input from nonsmokers, current smokers, and former smokers. The yellow cover page is yours to read and keep for your records. The survey t akes most people 10 minutes or less to complete. Now, Id like to briefly go over a few important points about completing the survey. Since the pages will be scanned, its extremely important that you completely bubble in the response circles. Check marks or xs over the circle wont scan well. Dont put your name anywhere on the survey, and please do not make any extraneous marks or comments near any response bubble. This will mess up the scanning. Theres a place for comments at the end of the sur vey. You can use a pencil or a black or blue pen. If you make a mistake, cross through it and then bubble in your new response. Choose one response for each item. Skip any question that youre uncomfortable answering. Are there any questions about com pleting the survey? OK. When youre done with the survey, please bring them to the _________ (front of the classroom, back of the classroom, or both). Thank you!

PAGE 159

145 LIST OF REFERENCES Alreck, P. L., & Settle, R. B. (1995). The survey research handbook (2 nd ed.). Boston: Irwin/McGraw Hill. American Association for Health Education. (2001). Report of the 2000 Joint Committee on Health Education and Promotion Terminolog y. American Journal of Health Education, 32 90 103. American College Health Association. (2000). Position statement on tobacco on college and university campuses Baltimore: Author. American College Health Association (2002). National College Health Asses sment web summary Retrieved from http://www.acha.org/projects_programs/ nacha_sampledata_public.cfm on February, 14, 2002. Bandura, A. (2002). Social cognitive theory of mass communication. In J. Bryant & D. Zillmann (Eds.), Media effects: Advances in th eory and research (2 nd ed., pp. 121 153). Mahwah, NJ: Lawrence Erlbaum. Baranowski, T., Perry, C. L., & Parcel, G. S. (1997). How individuals, environments, and health behaviors interact: Social cognitive theory. In K. Glanz, F. M. Lewis, & B. K. Rimer (E ds.), Health behavior and health education : Theory, research, and practice (2 nd ed., pp. 153 178). San Francisco: Jossey Bass. Bauman, K. E., Botvin, G. J., Botvin, E. M., & Baker, E. (1992). Normative expectations and the behavior of significant others: A n integration of traditions in research on adolescents cigarette smoking. Psychological Reports, 71, 568 570. Buss A. H., & Plomin, R. (1986). The EAS approach to temperament. In R. Plomin & J. Dunn (Eds.), The study of temperament: Changes, continuiti es, and challenges (pp. 67 77). Hillsdale, NJ: Lawrence Erlbaum. Carvajal, S.C., Wiatrek, D. E., Evans, R. I., Knee, C. R., & Nash, S. G. (2000). Psychosocial determinants of the onset and escalation of smoking: Cross sectional and prospective findings i n multiethnic middle school samples. Journal of Adolescent Health, 27 255 265.

PAGE 160

146 Centers for Disease Control and Prevention. (1997, November 14). Youth risk behavior surveillance: National college health risk behavior survey, United States, 1995. [Electr onic version]. CDC Surveillance Summaries, Morbidity and Mortality Weekly Report 46 (No. SS 6) Retrieved from http://www.cdc.gov on February 15, 2002. Centers for Disease Control and Prevention. (1998, May 22). Selected cigarette smoking initiation and quitting behaviors among high school students United States, 1997 [Electronic version]. Morbidity and Mortality Weekly Report 47 386 389. Retrieved from http://www.cdc.gov on February 15, 2002. Centers for Disease Control and Prevention. (1999). Chronic diseases and their risk factors: The nations leading causes of death [Electronic version]. Retrieved from http://www.cdc.gov/nccdphp/statbook/statbook.htm on February 15, 2002. Centers for Disease Control and Prevention. (2000, December 22). Reducing tobacco use: A report of the Surgeon General [Electronic version]. Morbidity and Mortality Weekly Report 49 1 27. Retrieved from http://www.cdc.gov on February 15, 2002. Centers for Disease Control and Prevention. (2001, November 2). Youth tobacco surveillance United States, 2000. [Electronic version]. CDC Surveillance Summaries, Morbidity and Mortality Weekly Report 50 (No. SS 4) Retrieved from http://www.cdc.gov on February 15, 2002. Cen ters for Disease Control and Prevention. (2002, April 12). Annual smoking attributable mortality, years of potential life lost, and economic costs United States, 1995 1999 [Electronic version]. Morbidity and Mortality Weekly Report, 51 300 303. Retrieved f rom http://www.cdc.gov on October 15, 2002. Centers for Disease Control and Prevention. (n.d.) Exposure to environmental tobacco smoke and cotinine levels Fact sheet. [Electronic version]. Retrieved from http://www.cdc.gov/tobacco/research_data/ environm ental/factsheet_ets.htm on March 20, 2003. Cherry, N., & Kiernan, K. (1976). Personality scores and smoking behavior: A longitudinal study. British Journal of Preventive and Social Medicine, 30 123 131. Choi, W. S., Pierce, J. P., Gilpin, E. A., Farkas, A J., & Berry, C. C. (1997). Which adolescent experimenters progress to established smoking in the United States? American Journal of Preventive Medicine, 13 385 391. Cohen, J. (1977). Statistical power analysis for the behavioral sciences New York: A cademic Press. Conrad, K. M., Flay, B. R., & Hill, D. (1992). Why children start smoking cigarettes: Predictors of onset. British Journal of Addiction, 87, 1711 1724.

PAGE 161

147 Debernardo, R. L., Aldinger, C. E., Dawood, O. R., Hanson, R. E., Lee, S., & Rinaldi, S. R. (1999). An e mail assessment of undergraduates attitudes toward smoking. Journal of American College Health, 48, 61 66. Distefan, J. M., Gilpin, E. A., Choi, W. S., & Pierce, J. P. (1998). Parental influences predict adolescent smoking in the United St ates, 1989 1993 [Electronic version] Journal of Adolescent Health, 22 466 474. Duncan, T. E., Tildesley, E., Duncan, S. C., & Hops, H. (1995). The consistency of family and peer influences on the development of substance use in adolescence. Addiction, 9 0, 1647 1660. Emmons, K. M., Wechsler, H., Dowdall, G., & Abraham, M. (1998). Predictors of smoking among U.S. college students. American Journal of Public Health, 88, 104 107. Everett, S. A., Husten, C. G., Kann, L., Warren, C. W., Sharp, D., & Crossett, L. (1999). Smoking initiation and smoking patterns among U.S. college students. Journal of American College Health, 48, 55 60. Flay, B. R. (1993). Youth tobacco use: Risks, patterns, and control. In C. T. Orleans & J. Slade (Eds.), Nicotine addiction: Prin ciples and management (pp. 132 183). New York: Springer Publishing Company. Flay, B. R., Hu, F. B., & Richardson, J. (1998). Psychosocial predictors of different stages of cigarette smoking among high school students. Preventive Medicine, 27, (5), A9 A18. Hestick, H., Perrino, S. C., Rhodes, W. A., & Sydnor, K. D. (2001). Trial and lifetime smoking risks among African American college students. Journal of American College Health, 49, 213 219. Hosmer, D. W., & Lemeshow, S. (1980). A goodness of fit test for the multiple logistic regression model. Communications in Statistics A10 1043 1069. International Personality Item Pool (2001). A scientific collaboration for the development of advanced measures of personality traits and other individual differences Retrieved from http://ipip.ori.org/ on September 18, 2002. Jackson, C. (1997). Initial and experimental stages of tobacco and alcohol use during late childhood: Relation to peer, parent, and personal risk factors. Addictive Behaviors, 22, 685 698. Jackson C. (1998). Cognitive susceptibility to smoking and initiation of smoking during childhood: A longitudinal study. Preventive Medicine, 27 129 134.

PAGE 162

148 Johnston, L. D., OMalley, P. M., & Bachman, J. G. (2001). Monitoring the Future national survey results on drug use, 1975 2000. Volume II: College students and adults ages 19 40 Bethesda, MD: National Institute on Drug Abuse. Killen, J. D., Robinson, T. N., Haydel, K. F., Hayward, C., Wilson, D. M., Hammer, L. D., et al. (1997). Prospective study of risk fact ors for the initiation of cigarette smoking. Journal Consulting and Clinical Psychology, 65, 1011 1016. Mayhew, K. P., Flay, B. R., & Mott, J. A. (2000). Stages in the development of adolescent smoking [Electronic version] Drug & Alcohol Dependence, 59, (Supplement1), S61 S81. Morris, L. W. (1979). Extraversion and Introversion: An Interactional Perspective New York: John Wiley. Page, R., M. (1998). College students distorted perception of the prevalence of smoking. Psychological Report, 82, 474. Pallo nen, U. E., Prochaska, J. O., Velicer, W. F., Prokhorov, A. V., & Smith, N. F. (1998). Stages of acquisition and cessation for adolescent smoking: An empirical integration [Electronic version]. Addictive Behaviors, 23 303 324. Patrick, K., Covin, J. R., F ulop, M., Calfas, K., & Lovato, C. (1997). Health risk behaviors among California college students. American Journal of College Health, 45 265 272. Pierce, J. P., Farkas, A. J., Evans, N., Berry, C., Choi, W., Rosbrook, B., et al. (1993). Tobacco use in C alifornia 1992: A focus on preventing uptake in adolescents. Sacramento, CA: Dept of Health Services. Pierce, J. P., Choi, W. S., Gilpin, E. A., Farkas, A. J., & Merritt, R. K. (1996). Validation of susceptibility as a predictor of which adolescents take up smoking in the United States. Health Psychology, 15 355 61. Prokhorov, A. V., de Moor, C. A., Hudmon, K. S., Hu, S., Kelder, S. H., & Gritz, E. R. (2002). Predicting initiation of smoking in adolescents: Evidence for integrating the stages of change a nd susceptibility to smoking constructs [Electronic version]. Addictive Behaviors, 27 697 712. Samet, J. M. (2001). The risks of active and passive smoking. In P. Slovic (Ed.), Smoking: Risk, perception, and policy (pp. 3 28). Thousand Oaks, CA: Sage. Si pps, G. J., & Alexander, R. A. (1987). The multifactorial nature of extraversion introversion in the Myers Briggs Type Indicator and Eysenck Personality Inventory. Educational and Psychological Measurement, 47 543 552. Spielberger, C. D., & Jacobs, G. A. (1982). Personality and smoking behavior. Journal of Personality Assessment, 46 396 403.

PAGE 163

149 Stern, R. A., Prochaska, J. O., Velicer, W. F., & Elder, J. P. (1987). Stages of adolescent cigarette smoking acquisition: Measurement and sample profiles. Addictiv e Behaviors, 12, 319 329. Substance Abuse and Mental Health Services Administration. (2002). Results from the 2001 National Household Survey on Drug Abuse: Volume I: Summary of national findings [Electronic version]. Rockville, MD: Office of Applied Studie s. Retrieved from http://www.samhsa.gov on November 7, 2002. Task Force on National Health Objectives in Higher Education. (2002). Healthy campus 2010: Making it happen. (Conference Edition). Baltimore: American College Health Association. Unger, J. B., J ohnson, C. A., Stoddard, J. L., Nezami, E., & Chou, C. (1997). Identification of adolescents at risk for smoking initiation: Validation of a measure of susceptibility. Addictive Behaviors, 22 81 91. Unger, J. B., Rohrbach, L. A., Howard Pitney, B., Ritt O lson, A., & Mouttapa, M. (2001). Peer influences and susceptibility to smoking among California adolescents. Substance Use and Misuse, 36 551 571. Unger, J. B. & Rohrbach, L. A. (2002). Why do adolescents overestimate their peers smoking prevalence? Cor relates of prevalence estimates among California 8 th grade students. Journal of Youth and Adolescence, 31 147 153. U.S. Department of Health and Human Services. (1994). Preventing tobacco use among young people: A report of the Surgeon General [Electroni c version]. Retrieved from http://www.cdc.gov/tobacco/sgrpage.htm on November, 20, 2001. U.S. Department of Health and Human Services (2000 ). Healthy people 2010 [Electronic version]. Retrieved from http://www.healthypeople.gov on January 20, 2003. U.S. De partment of Health and Human Services (2001). Cigarette smoking among adults United States, 1999 [Electronic version]. Retrieved from http://www.cdc.gov/tobacco/sgrpage.htm on November, 20, 2001. U.S. Department of Health and Human Services (2001a). Targ eting tobacco use [Electronic version]. Retrieved from http://www.cdc.gov/tobacco/sgrpage.htm on November, 20, 2001. U.S. Department of Health and Human Services (2001b). Women and smoking: A report of the Surgeon General [Electronic version]. Retrieved fr om http://www.cdc.gov/tobacco/sgrpage.htm on November, 20, 2001. Wang, M. Q., Fitzhugh, E. C., Eddy, J. M., Fu, Q., & Turner, L. (1997). Social influences on adolescents smoking progress: A longitudinal analysis. American Journal of Health Behavior, 21, 1 11 117.

PAGE 164

150 Wechsler, H., Rigotti, N. A., Gledhill Hoyt, J., & Lee, H. (1998). Increased levels of cigarette use among college students. Journal American Medical Association, 280, 1673 1678. West, P., & Michell, L. (1999). Smoking and peer influence. In A. J. Goreczny & M. Hersen (Eds.), Handbook of pediatric and adolescent health psychology (pp. 179 202). Needham Heights, MA: Allyn & Bacon. West, P., Sweeting, H., & Ecob, R. (1999). Family and friends influences on the uptake of regular smoking from m id adolescence to early adulthood. Addiction, 94, 1397 1411. White, V., Hill, D., & Hopper, J. (1996). The outgoing, the rebellious, and the anxious: Are adolescent personality dimensions related to the uptake of smoking? Psychology and Health, 12 73 8 5.

PAGE 165

151 BIOGRAPHICAL SKETCH Jane Emmere is originally from the Milwaukee area. She graduated from the University of Wisconsin at Stevens Point with a bachelors degree in physical education/wellness. Jane completed her degree in 1986, and then worked at the Na tional Wellness Institute in Stevens Point. Her responsibilities there included directing the National Wellness Conference and assisting with worksite wellness projects. After working three years at the National Wellness Institute, Jane decided to pursu e a masters degree. In the fall of 1989, she moved to Athens, Georgia, enrolling in the Department of Health Promotion and Behavior. In 1991, she moved to Gainesville, Florida, to take a position at the University of Floridas Student Health Care Cente r. For over eight years, she coordinated health promotion programs and services for college students. Her areas of expertise included body image, eating disorders, stress management, peer education, weight management, and wellness. Jane received her Ph .D. in the spring of 2003 with specialties in health communication and research design. She returned to the University of Florida, Student Health Care Center, in April 2003 and serves as the Assistant Director of Health Education Programs.