Effect of prevention education on smoking relapse for navy recruits


Material Information

Effect of prevention education on smoking relapse for navy recruits
Physical Description:
ix, 163 leaves : ill. ; 29 cm.
Pokorski, Thomas Lee, 1953-
Publication Date:


Subjects / Keywords:
Tobacco habit -- Treatment   ( lcsh )
Tobacco habit -- Prevention   ( lcsh )
Naval art and science -- Tobacco use   ( lcsh )
Navies -- Tobacco use   ( lcsh )
Sea-power -- Tobacco use   ( lcsh )
Smoking cessation   ( lcsh )
bibliography   ( marcgt )
non-fiction   ( marcgt )


Thesis (Ph. D.)--University of Florida, 1992.
Includes bibliographical references (leaves 151-161).
Statement of Responsibility:
by Thomas Lee Pokorski.
General Note:
General Note:

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
oclc - 27804077
System ID:

Full Text








I wish to express my sincere gratitude to Dr. W. William

Chen, chairman of my dissertation committee. His guidance

during the past three years, especially during the

dissertation process, has been invaluable. I would also like

to extend my appreciation to Dr. R. Morgan Pigg, Dr. Claudia

Probart, Dr. Steve Dorman, and Dr. Marc Branch for serving as

members of my doctoral committee. Without the guidance and

assistance of my committee I would never have been able to

complete this work, nor retain my sanity. I also appreciate

the statistical assistance provided by Dr. David Miller, and

the help of Dr. Roger Bertholf in cotinine analysis.

I wish to thank my parents, Pat and Joe Pokorski, for

lifelong support of all my endeavors. I'd like to

acknowledge the patience my children, Mike and Nicole, have

shown over the past few years while having to deal with a

"part-time" dad. I would especially like to thank my wife

Liz for her understanding, patience, editorial assistance,

and completion of over 600 scantron forms.

I'd like to recognize the assistance provided by the

Navy's Office of Health and Physical Readiness. Without the

financial assistance provided by this office the study could

not have been as comprehensive. I'd like also to thank the

American Cancer Society, especially Roberta Moss and Marsha

Nenno, for providing the shirts, pamphlets, and posters used

in the study.

I would like to thank Captain Kathleen M. Bruyere, U.S.

Navy, and her fine staff at Recruit Training Center, Orlando,

Florida. The assistance I received while conducting this

study was tremendous. I am especially grateful for all the

efforts of PHCM Breece, without which this study could not

have been done.

Finally, I would like to thank all the fine Navy

recruits who participated in this project and made the study

the success it was.


ACKNOWLEDGEMENTS ........................................... ii

LIST OF TABLES ............................................vi

ABSTRACT ................................................... vii

INTRODUCTION ..................... ........... .. ...... ....... 1

Statement of the Research Problem .............................6
Purpose of the Study.................................... 7
Hypotheses ................ .................. ............. 8
Significance of the Study................................ 8
Delimitations ... ....................................... 11
Assumptions ......... .................................. 13
Definition of Terms .................................. ..... 13

REVIEW OF LITERATURE ....................................... 16

Problems Related to Tobacco Use........................17
Tobacco Use in the U.S ...................... ........ 18
Military Tobacco Use and Related Problems...............19
Navy Tobacco Use ......................... ....... .24
Smoking Relapse ............................... .....26
Smoking Policy and Effect on Smoking Cessation...........36
Other Relapse Research .. ............................. 39
Smoking Prevention.... ................ ... ................ 39
Educational Intervention .............................45
Validation of Smoking Status Self-Report................47
Bogus Pipeline ....................................... 48
Biochemical Tests ........................ ..............49
Concluding Statement....................... ............ 52

METHODS AND MATERIALS .................................... 55

Introduction.................... .......... .........55
Subjects .... .............. ..... ...................55
Instruments ........................................... 59
Self-Report Validation Techniques s................... 62
Procedures ............................................64
Data Preparation.............................................75
Analysis Plan.... ........................ ..................76

RESULTS AND DISCUSSION .................................... 79

Introduction. .........................................79
Population Description................................. 79
Results...................... ......... .... ................. 82
Discussion .............................................. 95


Summary........................ .......................... 107
Conclusions........ ...................... ..... ............ 110
Recommendations .....................................113










INTAKE SURVEY ...............................116

GRADUATION SURVEY ........................ 123

THREE MONTH FOLLOW-UP SURVEY ................. 126


CONSENT STATEMENT .......................... 132

EDUCATION CURRICULUM ................ ....... 133

SHAM TREATMENT CURRICULUM ................... 146

QUIT SMOKING WALLET CARD .....................148


LIST OF REFERENCES ................ .......... ..... ........ 151

BIOGRAPHICAL SKETCH ........................................ 162



Table 1. Population comparisons .............................. 23

Table 2. Study group comparison ........................ 58

Table 3. Test-retest reliability ......................... 61

Table 4. Treatment group assignments .................... 65

Table 5. Repeated measures design ........................ 77

Table 6. Chi-square frequency table ...................... 77

Table 7. Percent survey completion at each measurement
period .................................................. 80

Table 8. Percentage of smokers in non-respondent group
and original study population.......................... 81

Table 9. Percent of current smokers at each measurement
period .................................................. 82

Table 10. Percent of current smokers by gender and
corresponding relapse rates................ ............ 83

Table 11. Nicotine tolerance levels reported with
corresponding relapse rates ........................... 84

Table 12. Percentage smoking initiation ................. 86

Table 13. Attitude scores at each measurement period
and differences pre to post for all subjects............ 88

Table 14. Attitude scores at each measurement period
and differences pre to post for smokers................ 89

Table 15. Knowledge scores at each measurement period
and differences observed ................................. ... 91

Table 16. Percent of respondents indicating intention
not to smoke in the future ........................... 92

Table 17. Cigarette consumption ........................ 93

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




August, 1992

Chairman: Wei William Chen, PhD
Major Department: Health and Human Performance

This study examined impact of an education program on

smoking knowledge, attitudes, and intentions of Navy recruits

undergoing 8 weeks of training in a normal no-smoking

environment. Also examined were smoking relapse and

cigarette consumption subsequent to recruit graduation.

Specifically the study was designed to (1) add to literature

information on effective educational smoking

prevention/cessation techniques; (2) measure effect of

current no-smoking policy at Recruit Training Center,

Orlando, FL; (3) examine differences in observed levels of

smoking knowledge, attitudes, intentions, and relapse after

intervention; (4) examine smoking initiation rate differences

(for non-smokers prior to recruit training); and (5) examine

variables affecting smoking relapse.


This study included 496 recruits (357 men and 139

women), in seven companies. Companies were randomly assigned

to 4 treatment conditions: (1) policy only comparison; (2)

policy plus education; (3) policy plus education, with a

booster; and (4) policy plus sham treatment. Tobacco use

knowledge/behavior surveys were administered to all subjects

at the beginning of recruit training, at graduation, and

three months after graduation. Subjects in both education

groups received a three-hour education intervention. The

booster group received an additional hour-long booster

program at the end of the training cycle. Cotinine analyses

of randomly selected urine samples were performed

concurrently with pre- and postsurveys.

Data analysis was accomplished using repeated measures

analysis of variance, frequency tables incorporating

Pearson's chi-square statistic, and categorical modeling

procedures (SAS PROC CATMOD). Results indicated significant

educational effects for smoking knowledge and attitude

scores, reduced violation of liberty no-smoking rules, and

cigarette consumption for relapsers. No significant

treatment group differences were noted for relapse rate,

smoking initiation, or smoking intentions.

In summary, the education program used, though not

significantly affecting smoking relapse in the short term,

may positively affect future cessation attempts. The best

predictors of smoking relapse in Navy recruits were nicotine

addiction and heavy prior cigarette consumption. Study


results recommend that smoking prevention/cessation education

be part of recruit training, but further research is needed

to identify more effective ways of reaching the heavier,

addicted smokers entering the Navy.


Cigarette smoking can be linked to 1 in 6 deaths in the

United States each year. Surgeons' General for the past 10

years named cigarette smoking as the most important

preventable cause of death in society (U.S. Department of

Health and Human Services [USDHHS], 1989, 1990).

Specifically, cigarette smoking has been linked to the three

leading causes of death in the U.S. (heart disease, cancer,

and cerebrovascular disease). Likewise, involuntary smoking

causes many of the same diseases associated with active

smokers (USDHHS, 1986). The annual cost to society in

disease, death, and absenteeism related to smoking has been

estimated in excess of $50 billion (Fielding, 1986).

In response to this problem, health professionals have

provided tobacco prevention and smoking cessation programs

for many years. These efforts have shown encouraging results

in the general population. Smoking prevalence has dropped

from 40% in 1965 to 29% in 1987 (USDHHS, 1989). The Surgeon

General reports that while over 50 million Americans continue

to smoke, more than 90 million would be smoking in the

absence of recent changes in the smoking and health

environment (USDHHS, 1989). Smoking cessation produces major

and immediate health benefits for smokers of all ages

(USDHHS, 1990a) but it is not an easy task. Eighty percent

of current smokers indicate they would like to quit, and two-

thirds have made at least one serious attempt to quit

(USDHHS, 1989), but 80-85% of those who try to quit smoking

each year relapse within 12 months of cessation (O'Connell,

1990; Schwartz, 1987).

Certain sub-groups of the general U.S. population have a

higher smoking prevalence. Military personnel are more

likely than civilians to be smokers (Bray, Marsden, &

Peterson, 1991). Though the prevalence of smoking in the

military declined from 51% in 1980 to 41% in 1988, the rate

remains much higher than in the general population (29%).

Bray et al. (1991) indicate the gap is actually less when the

civilian population is standardized to reflect sociode-

mographic distribution of the military (44% military vs. 39%

civilian prevalence), but the difference remains significant.

The population the military primarily draws on for new

members has shown a smoking prevalence leveling off in the

past decade. Smoking prevalence among teen-agers did not

decline during the 1980's and studies have shown smoking

prevalence in high school seniors at 17-32% (Johnston,

O'Malley, & Bachman, 1991; Center for Disease Control [CDC],

1991a) with rates much higher for high school dropouts.

The Surgeon General lists military personnel as one

population to be targeted for prevention and cessation

interventions (USDHHS, 1989). Many factors contribute to

high smoking prevalence in the military. Tradition of

smoking in the military, living conditions, demographic

population makeup, low cigarette cost in military exchanges,

and advertising targeted to young military members represent

factors hypothesized as causes for high military smoking


Smoking prevalence in the Navy (44%) slightly exceeded

the military average in a 1988 survey (Bray et al., 1988).

It has been estimated smoking costs the Navy nearly $25

million a year in higher health care and insurance costs, and

an additional $140 million in lost wages due to absenteeism

(Zolton, 1992). Efforts are underway in the Navy to reduce

the smoking prevalence and the Navy Surgeon General has set a

goal of a smoke-free Navy by the year 2000 (Nelson & Roth,

1991). Surveys of incoming Navy recruits, though, show a

smoking prevalence of between 28% and 40% (Cronan & Conway,

1988; Grochmal, 1990). If the Navy wishes to attain the goal

of being smoke-free not only will current smoking members

need assistance stopping, but the number of new smoking

members must be reduced and eventually eliminated. Current

Navy smoking policies have not yet proven effective in

lowering smoking prevalence to civilian levels. Policy

efforts thus far have only reduced smoking rates from 54% in

1980 to 44% in 1988 (Bray et al., 1988).

A smoking ban at all Recruit Training Commands (RTCs)

seems effective in stopping smoking temporarily for recruits

who enter as smokers (Commander E. Reeves, personal

communication, May 14, 1991). This policy stemmed from study

results showing Navy recruits actually starting to smoke

while in basic training (Cronan et al., 1988). Though a near

100% effective cessation rate is encouraging, smoking relapse

after recruit training is suspected to be very high. A study

to determine current relapse rates is underway at the Naval

Health Research Center (T. L. Conway, personal communication,

May 9, 1991).

The specific problem addressed in the present study

involved finding ways to decrease smoking relapse as recruits

leave boot camp and enter the general Navy population. If

relapse of recruit smoking can be prevented, the number of

new smokers entering the general Navy population can be

reduced. Navy recruits represent a segment of the U.S.

population highly resistant to smoking cessation: young

smokers (Ferrence, 1989). They are usually between 17 and 30

years old with a mean of about 19 years (Cronan, Conway, &

Hervig, 1989; Chief of Naval Technical Training [CNTECHTRA],

1991). Young smokers tend to show less concern about

potential hazards of smoking, though they show a high

awareness of the health risks (Tuakli, Smith, & Heaton,

1990). Therefore, it is a challenge to find effective

methods of lasting smoking cessation for Navy recruits.

Smoking development progresses through several stages

over a course of two or more years (Leventhal, Baker,

Brandon, & Fleming, 1989). Many smoking recruits may not

have reached the final stage of dependence and regular use.

For these individuals smoking relapse rates after recruit

training may be affected several ways: 1) they may have never

considered themselves smokers and may not start again; 2) the

eight week gap may represent a brief lapse in their

development cycle, and they may continue to smoke whether or

not they see themselves as future smokers; or 3) they may

think they can stop just as easily in the future, feel immune

to possible risks, be influenced by smoking friends they

subsequently meet, and eventually progress to the last stage

of development.

Recruits who progress to the last stage of smoking

development before entering the Navy should relapse like any

other ex-smoker in the cessation population. However, one

would expect these recruits to exhibit some relapse

differences in that most have not really quit smoking

voluntarily; they quit, not by choice, but to conform to

existing policy. Since smoking relapse rates have never been

studied for this type of population it is difficult to

predict exactly what will effect reduction of relapse.

The transtheoretical model of behavior change, as

applied to smoking cessation (Prochaska & DiClemente, 1983),

shows that smokers cycle through various stages (pre-

contemplation, contemplation, action, and maintenance) in

their quit attempts. Progression through these stages

requires voluntary incorporation of a number of processes of

change. Smoking recruits are placed into the action stage

and may revert to their pre-recruit stage upon graduation and

reinstitution of free-choice concerning tobacco use.

Physical as well as psychological factors are involved

in relapse (Carmody, 1990; Shiffman, Read, Maltese, Rapkin, &

Jarvik, 1985). For recruits, physical withdrawal symptoms

are gone after eight weeks of training, as peak physical

withdrawal symptoms of tobacco appear to last for about 1 2

weeks (USDHHS, 1988). Therefore, if recruits relapse to

smoking it will probably be from psychological factors, a

belief they never quit, or a willful intention to start

again. Each of these areas were addressed in the present


Statement of the Research Problem

The military population has been targeted as a group in

need of effective smoking prevention and cessation efforts.

Efforts thus far have not succeeded in reducing smoking

prevalence to civilian levels. An attempt has been made to

promote cessation and discourage initiation by banning

tobacco use for all personnel at recruit training centers.

No study results have yet been published concerning the

effect of the policy on smoking after training, and no

efforts have been made to find ways to decrease smoking

relapse for graduating recruits.

In general, military smoking prevalence can be reduced

by addressing current service members and new recruits.

Cessation and prevention efforts for current service members

is vital, but if 40% of new recruits are smokers, the problem

will perpetuate. Specifically, the present study focuses on

reducing the number of smokers in the recruit population. To

do this, the effect of current smoking policy at Recruit

Training Center, Orlando, FL, was examined. Relapse after

graduation, smoking attitudes, and future smoking intentions

were assessed. Changes in smoking relapse and initiation were

then examined for groups exposed to various intervention


Purpose of the Study

The purpose of this study was to examine the impact of

an education program on smoking knowledge, attitudes,

intentions, and on smoking relapse rate and subsequent

cigarette consumption. Specifically the study was designed


(1) measure effect of the no-smoking policy at

Recruit Training Center, Orlando, FL, on recruit smoking

knowledge, attitudes, intentions, and relapse;

(2) examine differences in observed levels of recruit

smoking knowledge, attitudes, intentions, and relapse

for various treatment conditions;

(3) examine smoking initiation rate differences (for

non-smokers prior to recruit training) for various

treatment conditions; and

(4) examine variables affecting smoking relapse.


The null hypotheses relating to Navy recruits after

eight weeks in a smoking restricted environment are that no

differences will be observed among treatment conditions for

(1) smoking relapse rate,

(2) smoking attitudes,

(3) smoking knowledge,

(4) smoking intentions,

(5) cigarette consumption, and

(6) smoking initiation after graduation (for

never/former smokers prior to recruit training).

Significance of the Study

Results of this study should assist Navy smoking

cessation efforts by finding effective means of reducing

smoking prevalence in personnel entering the service. This

study was unique in that it was the first to examine smoking

intervention while including the Navy female recruit

population. Secondarily the study was designed to contribute

to literature information on: a) successful techniques to

prevent relapse in young ex-smokers, and b) successful

techniques to prevent smoking initiation in non-smokers

during employment transition.

This study will provide further evidence to support or

refute the theory that policy alone is insufficient to effect

permanent change on smoking decisions. The theory of

prevention education as an effective measure for behavior

change also was tested. This study should prove useful for

determining if subject tailored smoking prevention classes,

in the context of a strict no-smoking policy, reduce

recidivism after the policy is removed.

Since this study was the first smoking relapse

prevention research conducted with military recruits it

should produce new knowledge applicable to all military

training commands. The results will provide a basis for

further research in discovering program variables which

increase success, and subject variables which determine who

is suited for particular interventions. Portions of the

results may also prove applicable to behavior change of other

detrimental health habits (eg. drug and alcohol use, poor

nutrition habits, seat belt usage).

Smoking relapse for recruits will be affected by many

factors they encounter after leaving boot camp. The present

study cannot change any of these factors. However, the

interventions were designed to help participants recognize

and cope with many of these factors. For recruits who

believe they have quit, establishing a firm intention to

remain quit is very important. This aspect of prevention

stems from the theory of reasoned action which states the

best predictor of behavior is a person's intention to perform

the behavior (Ajzen & Fishbein, 1980).

For recruits who believe they have not quit, or are

determined to start again, interventions used in this study

may still be of benefit. If they can realize they have quit

smoking, and the quitting was a beneficial achievement for

them, they may be able to enter the behavior change cycle at

the maintenance stage. If this measure fails, the program

still may be of benefit if it moves smokers from the pre-

contemplation stage toward the contemplation stage in which

they may give serious consideration to quitting at a later

date. If so, these individuals may be more receptive to

Navy-wide smoking policies and programs they will be exposed

to throughout their careers.

Research indicates light smokers often make the

transition to heavy smoking during the first two years after

high school (Johnston et al., 1991). If this transition can

be prevented in Navy accessions, receptiveness to later

cessation efforts also may be assisted. Successful

recidivism prevention and earlier cessation will certainly

boost progress toward the goal of a smoke free Navy by the

year 2000.

Another group that will benefit from the smoking relapse

program includes recruits who have never smoked, especially

those who may already be susceptible to social influences.

According to the National Institutes of Health, information

about the health and social consequences of smoking proves

critical for cessation (Glynn, Boyd, & Gruman, 1991). This

same information has been shown effective in smoking

prevention programs (Flay, 1985; Glynn, 1989). Therefore, a

comprehensive smoking education program addressing the needs

of all recruits, should be effective in reducing the number

of smokers entering the Navy population and assisting future

prevention and cessation efforts.


(1) The curriculum used was submitted for review to

several experts in the fields of health education and

smoking prevention/cessation; the curriculum was

subsequently adjusted to include most of the suggestions


(2) The sample was randomly selected and found to be

representative of a six month sampling of all recruits

at RTC Orlando and all three Navy recruit commands.

(3) Companies involved in the study were randomly

assigned to the four treatment groups.

(4) Cotinine analysis of urine was accomplished on

random samples of participants in conjunction with the

intake and graduation surveys to verify self-reported

tobacco use, and a bogus pipeline procedure was used to

increase validity of self-reported tobacco use.

(5) Companies which had companion companies (combined

classroom periods) received the same study treatment.

(6) Follow-up surveys were mailed to participants'

command, whenever possible, in an effort to increase



(1) Because previous research with the study population

was minimal, techniques used successfully with other

populations were adapted for the study.

(2) The original study population was reduced to its

final number due to unavoidable attrition resulting from

medical problems, reading aptitude, drug test results,

voluntary drops, and a number of other factors.

(3) Only recruits from one of the three Navy recruit

training facilities were used.

(4) Several potential threats to internal validity

remained beyond experimental control including company

commander influence, and seasonality differences in


(5) Cotinine analysis was not accomplished for each

subject nor for each survey period.

(6) Addresses for follow-up mailings were obtained from

a Navy computerized personnel tracking network which may

be less than 100% accurate.


(1) Recruit company commanders uniformly followed no-

smoking policies set for the command.

(2) The sample population were uniformly exposed to

outside variables which might affect treatment outcome.

(3) Subjects in each treatment group self-reported

tobacco usage with the same degree of truthfulness.

(4) Subjects in differing treatment groups had no

interaction affecting results.

Definition of Terms

The following are working definitions for selected terms

used in this study.

Smoking relapse/recidivism was defined as at least one

puff per day for seven days. This definition was

recommended by a conference of the National Heart, Lung,

and Blood Institute (Carmody, 1990). Relapse does not

necessarily mean a return to the previous smoking rate.

Current smokers were defined for this study as

individuals who have smoked at least 100 cigarettes in

their lifetime and have smoked a cigarette, cigar, or

pipe within 2 months prior to entering the Navy (intake

survey); or since graduation from recruit training (3

month follow-up survey).

Current smokeless tobacco users were defined as having

categorized themselves as users of chewing tobacco or

snuff and have used at least 3 days in the past 12


Former smokers were defined as once current smokers, but

having not smoked a cigarette, cigar, or pipe in the

last 2 months.

Former smokeless tobacco users were defined as once

current users, but having not used chewing tobacco or

snuff in the last 2 months.

Never smokers were defined as having smoked less than

100 cigarettes in their lifetime, indicated they never

regularly smoked, and characterized themselves as having

never used cigarettes, cigars, or pipes.

Never users of smokeless tobacco were defined as having

categorized themselves as never using chewing tobacco or

snuff on a regular basis.

Light smoking was considered using less than one

cigarette per day.

Moderate smoking was considered using from 1 19

cigarettes per day.

Heavy smoking was considered using a pack, or more, of

cigarettes per day.

Smoking relapse rate was determined by dividing the

number of participants (categorized as current smokers

at intake) indicating they started smoking again after

boot camp, by the total number of smokers remaining in

the study.
S start
total smokers

Boot camp is a common term used for basic recruit

training in the military. Generally the training

received immediately upon entering the military, which

lasts about 8 weeks.

Sham treatment refers to a placebo treatment similar to

the main intervention only in amount of attention paid

to participants. The material presented is only vaguely

related to the study goal.

Command refers to the actual component of the Navy to

which participants were assigned after recruit training.

Knowledge gain refers to observed change in smoking

knowledge scores between pretest and posttest.


WHEN loves grows cool, thy fire still warms me;

When friends are fled, thy presence charms me,

If thou art full, though purse be bare,

I smoke, and cast away all my care

-German Smoking Song


The purpose of this chapter is to provide a summary of

what has been learned from other studies regarding smoking

prevention and smoking relapse prevention. Insight as to how

the current study will contribute to a fuller understanding

of the issues is also presented. The chapter starts with a

look at ill effects of tobacco, and tobacco use prevalence

rates in the U.S. population, and military sub-groups.

Several theories of smoking relapse are then reviewed

followed by an examination of how these theories have been

used in the past as basis for research in relapse prevention.

Research on the effects of smoking policy and miscellaneous

relapse research are then examined. Research on smoking

prevention, as it relates to this study, is then reviewed.

Prevention methods which current research has shown effective

are presented, followed by an overview of recent research in

ways of increasing truthfulness of smoking self-report, and

means to verify smoking self-report. A concluding statement

ends this chapter, in which the theoretical framework is

related to the actual experimental design of the study.

Problems Related to Tobacco Use

Cigarette smoking remains a major problem in U. S.

society today. A recent report (CDC, 1991b) says more than

434,000 deaths, and an estimated 1,199,000 years of potential

life lost (YPLL) before age 65, were caused by cigarette

smoking in 1988 in the United States. In 1964, the Surgeon

General reported (Public Health Service [PHS], 1964)

cigarette smoking causes lung and laryngeal cancer in men,

and causes bronchitis. The 1989 Surgeon General's report on

the health consequences of smoking (USDHHS, 1989) showed

cigarette smoking also causes coronary heart disease,

cerebral vascular disease (stroke), atherosclerotic

peripheral vascular disease, lung and laryngeal cancer in

women, oral cancer, esophageal cancer, chronic obstructive

pulmonary disease, intrauterine growth retardation, and low

birth weight babies. Smoking was also found to be a

contributing factor for cancers of the bladder, pancreas, and

kidney; and associated with cancers of the stomach and

uterine cervix (USDHHS, 1989). Research also has established

involuntary smoking as a cause of disease, including lung

cancer, in healthy non-smokers (USDHHS, 1986).

Cigarette smoking is the number one preventable cause of

death in the U.S. today (USDHHS, 1989). Smoking causes more

premature deaths than cocaine, heroin, alcohol, fire,

automobile accidents, homicide, and suicide combined (USDHHS,

1990). It has been estimated that smoking costs society over

$50 billion annually in smoking related diseases, death, and

absenteeism (Fielding, 1986). However, more than 50 million

Americans continue to smoke (USDHHS, 1990b).

Tobacco Use in the U.S.

Smoking prevalence in the adult U.S. population was

about 29% in 1987, which is down from the 40% prevalence seen

in 1965 (USDHHS, 1989). Unfortunately, this decrease in

smoking is not seen uniformly throughout the population. The

Surgeon General reported the prevalence of smoking in women

has declined more slowly than men, and smoking rates will be

about equal for men and women in the mid-1990's (USDHHS,

1989). The report also states female adolescents not

planning on higher education show a much higher smoking

initiation rate than for male adolescents. A higher smoking

prevalence was also reported among black and Hispanic men

than white men. More disturbing is the fact that smoking

prevalence among teenagers has not declined over the past

decade. Johnston, et al. (1991), reporting on their 15th

annual national survey of American high school seniors,

stated that in 1989 29.4% of seniors were current smokers (30

day prevalence of cigarette use), which was the same

proportion as the class of 1981. High school dropouts have

prevalence rates for all types of drugs, including

cigarettes, substantially higher than in-school students

(Johnston, O'Malley, & Bachman, 1989).

Military Tobacco Use and Related Problems

Tobacco use in the military has been quite prevalent for

many years. Tobacco has been part of the traditional rations

given to soldiers during war time (Ferrence, 1989). This

practice started the smoking habit in hundreds of thousands

of Americans and began an association of cigarettes with the

military that persists today. Very little research can be

found in the literature on military smoking cessation and

none can be found on military smoking relapse prevention.

However, this is not a reflection of the magnitude of the

problem. A recent survey of the military found over 40% of

all members are smokers (Bray, et al., 1988). Marsden, Bray,

and Herbold (1988) showed the number of illnesses reported by

U.S. military personnel is significantly related to smoking,

as well as to other substance abuses, such as alcohol. Their

study utilized data from the 1985 worldwide survey of alcohol

and non-medical drug use among military personnel. They

found the number of healthy behaviors practiced was inversely

related to use of alcohol, drugs, and tobacco. Substance

users, particularly heavy users, were more likely to have

experienced ill health during the past 12 months. When

controlling for background variables, which also related to

ill health, number of illnesses was still significantly

related to reported drinking level, drug use, and smoking.

Smokers of one or more packs per day averaged 0.22 more

illnesses per year than non-smokers. Ballweg and Bray (1989)

also reported smokers were more likely than non-smokers to

describe their health as fair or poor. Though the military

smoking rate is high, it reflects a significant decrease from

1980 when 51% of military members smoked (Bray, et al.,


Many explanations have been hypothesized to account for

the high military smoking prevalence seen today. The

following are descriptions of some of these hypotheses.

-Distinctive military conditions such as relocation

overseas, family separation, or a greater perceived

acceptance of use may foster higher rates of use (Bray

et al., 1991).

-Military personnel are not demographically

representative of the general population (this will be

discussed at length later in this chapter).

-The military has been selecting individuals who are

predisposed to smoke (Gardner, 1991). Smoking can be

predicted by pay-grade, race/ethnicity, education,

service branch, age, poor health practices, and higher

stress at work (Bray, Marsden, Guess, & Herbold, 1989;

Carroll, Lednar, & Carter, 1989).

-Tobacco products are available at much lower prices at

military outlets than in the civilian market (Cronan &

Conway, 1988; Nelson & Roth, 1991). Many programs are

funded by tobacco sales revenue. Fiscal 1988 saw $102.1

million in tobacco sales by Navy exchanges (McBeth,

1989) and in 1989 $20 million of tobacco profits were

given back to Morale Welfare and Recreation funds.

Fiscal 1991 sales topped $126 million (Steigman, 1992).

-Military life brings an increase in interpersonal

communication and social participation due to higher

density in living quarters and increased group

activities. These conditions are conducive to

initiation/continuance of the smoking behavior

(Ferrence, 1989).

-The military encourages cohesiveness and uniformity,

and members may begin smoking to become like each other,

and "one of the group" (Cronan & Conway, 1988).

-Work breaks and other opportunities to relax are often

paired with opportunities to smoke (Cronan & Conway,


-Civilian publications aimed at the military populations

contain a large amount of sponsorship/advertisement from

the tobacco industry (Nelson & Roth, 1991).

Because of the nature of military employment, any

substance abuse can often cause consequences of graver impact

than in the civilian community. Greater responsibilities are

placed in the hands of younger people in the military than

would ever be allowed in the civilian work force. A twenty

year old may be responsible for actual steerage of a large

naval vessel, for preparation of a multimillion dollar

aircraft for flight, or for emergency medical treatment of a

seriously injured person. If that individual's performance

is hindered because of use of any psychoactive substance,

including tobacco, lives and expensive hardware can be placed

in jeopardy. The U.S. Surgeon General has concluded nicotine

is a psychoactive substance (USDHHS, 1988).

The safety aspects of even legal substances such as

tobacco can be crucial. Smoking is restricted at certain

times on ships. The smoking lamp is either lit or out on

board a naval ship. This refers to a practice years ago

aboard naval vessels in which the captain would order a lamp

lit for sailors to light their cigarettes when their duties

had ended for the day. The lamp is gone but the term is

still used to tell when it is safe or not to smoke. A lit

cigarette can be a hazard around fuel, ordinance, or any

volatile substance. On submarines the air must be

recirculated, and cigarette smoke is a significant

contributor to atmospheric contaminants (Scali, 1989).

The military population is not demographically

representative of the general population, which may account

for some of the difference in smoking prevalence. Table 1

presents a breakdown of some of these differences.

Table 1. Population comparisons

Active Duty Total U.S.

Military Population














40 & over

72% (1,557)*

20% (426)

8% (85)





17.3% (376)

32.4% (704)

17.8% (387)

14.9% (324)

10.5% (228)

7.2% (157)











6.1% (15,054)

8.1% (19,981)

7.2% (17,701)

8.9% (21,878)

7.8% (19,194)

37.6% (92,518)

* Numbers are represented in thousands.
Source: U. S. Bureau of the Census, Statistical abstract of the
United States: 1990 (110th edition) Washington DC & Military Market,
October 1989, pp. 46-52.

Service members are younger, most are male, and the

majority of recruits come from the lower socioeconomic

levels. These are factors also associated with higher

smoking prevalence (Bray et al., 1991). Bray et al. (1991)

reported, controlling for key sociodemographic

characteristics known to be associated with substance abuse,

military personnel are in general still substantially more

likely than civilians to be smokers and heavy smokers. They

also reported differences among military and civilian women

are more pronounced than among military and civilian men.

The rates for men differed by 4 and 7 points respectively,

while the rates for women differed by 13 and 14 points.

Navy Tobacco Use

The most recent Worldwide survey of substance abuse and

health behaviors among military personnel (Bray, et al.,

1988) showed prevalence of any smoking in the Navy was 43.8%

in 1988. A 1986 survey of shipboard personnel indicated

49.8% smoked (Conway, & Cronan, 1988). Smokeless tobacco is

used by 16.1% of Naval personnel. A 1987 Navy report (Cronan

& Conway, 1988) found about 28% of incoming recruits were

regular smokers with another 13% claiming to be former


An informal tobacco use survey was conducted by the

dental staff at the Orlando Recruit Inprocessing Facility.

The survey was designed by Lieutenant Commander David

Grochmal, Dental Corps, and was accomplished as part of the

dental screening process. The survey was conducted from

March to October, 1990 and included over 13,000 recruits.

This survey found 39.7% (34.1% male and 56.7% female)

reported they were smokers before entering the Navy. Also,

7.4% of the men indicated they were smokeless tobacco users

(Grochmal, 1990).

In 1988 a survey of high school seniors and young adults

(Johnston, et al., 1989) found an 18% prevalence of daily

smoking for seniors and a 29% rate of current (casual)

smokers. This same study reported a drastically higher

smoking rate in respondents 1 4 years past high school who

were not in college. Other studies have shown that smoking

rates are higher for seniors who do not plan on pursuing

higher education (Glynn, 1990; Johnston et al., 1991). Since

the majority of the recruit population comes from this group,

the higher recruit smoking prevalence is not surprising.

Though the vast majority of Navy recruits have a high school

diploma (or equivalent), some have obtained it in a non-

traditional manner, and thus are part of a group at much

higher risk to smoke.

Smoking Relapse

"To cease smoking is the easiest thing I ever did; I

ought to know because I've done it a thousand times"

-Mark Twain

Most relapse does not occur immediately, but rather long

after signs of physiological abstinence have disappeared

(Jarvik & Hatsukami, 1989). Research shows approximately 70%

of successful quitters relapse within 3 months, and an

additional 10-15% relapse between 3 and 12 months after

quitting (Schwartz, 1987; Shiffman et al., 1985). Several

theories have been offered to explain why relapse rates are

so high. Shiffman et al. (1986) have suggested three types

of variables influence relapse proneness: a) enduring

personal characteristics, b) background variables, and c)

precipitants. Craving for nicotine and the positive

reinforcement nicotine brings has also been theorized as

reason for relapse (Carmody, 1990). Gritz, Carr, and Marcus

(1991) conducted a study on tobacco withdrawal syndrome in

unaided quitters. They found tobacco dependence to be a good

indicator of smoking relapse. Higher Fagerstr6m tolerance

scores predicted higher probability of relapse. Their other

main finding was higher cigarette consumption at enrollment

predicted higher short and long term relapse.

Theory of Reasoned Action

One model which has been used to describe likelihood of
behavior change has been termed the theory of reasoned action

(Ajzen & Fishbein, 1980). This theory states the best

predictor of behavior is a person's intention to perform the

behavior. Behavioral intention is seen as a function of two

determinants; the person's attitude toward the behavior, and

the perceived expectation of important others with regard to

performance of the behavior (subjective norm). In a review

of this theory Sutton (1989) said, "In other words, a

person's behavioral intentions, and hence behavior, depend

ultimately on the person's belief concerning; (a) the

possible consequences of the behavior, and (b) the

expectations of important others. It follows that in order

to change behavior it is necessary to change the underlying

beliefs" (p 291). The theory goes on to explain that

external variables can influence intentions, and thus

behavior, but only by influencing attitude toward behavior or

the subjective norm, or the relative importance of the two

components. Since starting or stopping cigarette smoking is

a behavior that falls in the domain of this theory (a

behavior that can be regarded as a decision) intention to

stop/reduce smoking should be a good predictor of behavior.

Transtheoretical Model (Stages of Change)

Prochaska and DiClemente first applied their stages of

change model to self change of smoking habits (Prochaska &

DiClemente, 1983). They believe most behavior involves

repetitive and habitual actions which are quite resistant to

extinction. Further, behavior change requires movement

through discrete stages in order to achieve maintained

cessation or initiation. The five stages of change an

individual cycles through in attempting behavior changes are:






Because changing smoking behavior is highly prone to

relapse, it is theorized that individuals tend to move

through these stages in a cyclical fashion (Prochaska &

DiClemente, 1982). Successful change often requires repeated

recycling through these stages. Research suggests the

average smoker cycles three to four times through the stages

before attaining long-term abstinence (Prochaska &

DiClemente, 1984; Marlatt, Curry, & Gordon, 1988). Those in

relapse and pre-contemplation do not have enough motivation

to change their smoking habits. Those in the contemplation

stage are aware of a need to change, but have not yet taken

action. The key to successful change lies in exiting the

maintenance stage after replacing the behavior in their

lifestyle. If the intervention method is tailored to the

individual's stage of change, the chances of success can be

enhanced. Rimer (1990) says people in the process of change

must have access to interventions which start at their stage

in the change process. Because of this, she says

intervention should include a range of options to reach

people in all stages. Intervention programs may not succeed

in behavior change, but may succeed in moving the individual

to a higher stage, which is important (Prochaska &

DiClemente, 1983). No studies have as yet been reported

concerning any education effect on this recycling phenomenon.

Social Cognitive Theory

Bandura's (1989) social cognitive theory revolves around

smokers' beliefs about ability to exercise control over

events which control their lives. Bandura sees self-efficacy

as influencing thought patterns, motivation, level of stress,

and selection of environments; thus affecting smoking

cessation outcome expectations.

All of these theories have commonalties, and show that

smoking relapse is individualistic and complex. Recruits who

smoked before entering the Navy have been placed in the

action stage of change while in recruit training, even though

they were probably in a lower stage before they came in. No

previous studies have reported or predicted what effect this

will have. It is not known what effect this will have on

their intentions to stay quit. Certainly it will show they

are capable of not smoking for an extended period of time.

What is known is young adult smokers do quit, or at least

attempt to quit, smoking. Thirteen percent of recruits

surveyed indicated they were former smokers (Cronan & Conway,

1988). A prospective cohort study from pre-adolescence to

young adulthood (Swan, 1991) found about 14% of 21 22 year

olds reported being ex-smokers. Another study (Ershler,

Leventhal, Fleming, & Glynn, 1989) states 52% of 6 12 grade

smokers had indicated at least one quit attempt. A survey by

Tuakli et al. (1990) indicated 69% of the 12 20 year old

smokers had tried to quit. The top reasons for quitting

were: 1) didn't enjoy smoking, 2) health concerns, 3)

pregnancy, 4) costs, and 5) parents or friends asked. This

indicates even young smokers are aware of the risks of

smoking, and some may have progressed through the cycle of

changing their smoking behavior.

Relapse is defined in the literature as a smoking

episode after a period of voluntary abstinence that is

followed by a return to continuous smoking (O'Connell, 1990).

To include recruit graduates in this category may be
stretching this definition, but since we now have an all

voluntary armed forces, all recruits enter the Navy of their

own free will. All recruits may not enter with the intention

to permanently quit smoking though, which may affect smoking

relapse seen after graduation.

Relapse Prevention

The topic of relapse prevention has received a lot of

research attention recently. Since most smoking cessation

programs report from 50-80% abstinence at the end of

treatment (Schwartz, 1987) improvement of long-term quit

rates is a logical next step in smoking cessation research.

In reviewing current issues in preventing smoking relapse,

Carmody (1990) found smoking relapse is influenced by a host

of variables, including: physiological factors, withdrawal

symptoms, conditioning processes, stress and negative mood

states, level of commitment and motivation to remain

abstinent, social support, weight gain, and repertoire of

cognitive and behavioral coping strategies in response to

temptations to smoke. He concluded that the primary

theoretical framework for relapse prevention is social

learning theory.

Mazur (1986) points out that social learning theory adds

the principles of observational learning and imitation to

traditional principles of classical and operant conditioning.

He suggests Bandura's social learning theory can help account

for the acquisition of many addictive behaviors including

smoking. Others (Carmody, 1990; Perry, Baranowski, & Parcel,

1990; Rimer, 1990) believe this theory is particularly

valuable to health educators because it illuminates the

dynamics of individual behavior and gives direction for

intervention design in behavior change programs. Bandura

(1989) says people make causal contributions to their own

motivation and action within a system of triadic reciprocal

causation. The model shows action, personal factors, and

environmental events operating as environmental determinants.

In the model, people partly determine the nature of their

environment and are influenced by it.

The preceding theories lend well to the multi-component

interventions thought to be important in relapse prevention

strategies. They can easily be applied to multi-level change

strategies because of inclusion of environmental, personal,

and behavioral constructs.

Social cognitive theory is an outgrowth of social

learning theory (Bandura, 1989). This theory appears to be

useful as a model in studying smoking relapse (Carmody,

1990). People's beliefs about their ability to exercise

control over events that affect their lives are central to

the theory. Wojcik (1988) found the strongest prediction of

abstinence versus relapse in a sample of 75 smokers

attempting to quit on their own was self-efficacy.

Shiffman et al. (1985) emphasize education, assessment,

and coping-skills training in their approach to relapse

prevention. They also found combining cognitive and

behavioral coping enhanced effectiveness of preventing

relapse. Seven types of behavioral coping were listed:


-engaging in a distractive activity



-engaging in a physical activity


-engaging in any other activity.

Cognitive coping strategies were:


-self-punitive thoughts

-positive health consequences

-negative health consequences

-distracting thoughts

-intent to delay

-other self-talk.

People who expect smoking cessation to require active coping

have been shown to be more successful in maintenance

(Shiffman et al., 1986). They determined it is probably most

useful to teach ex-smokers a broad repertoire of coping

responses to prevent relapse.

Other approaches to relapse prevention include

motivational enhancement, social support, coping skills

training, reinforcement, lifestyle balancing, and the use of

various pharmacological agents (Carmody, 1990). Several

studies have shown that use of refusal skill training

techniques is beneficial in maintenance of smoking cessation

(Curry, Marlatt, & Gordon, 1987; Curry, Murray, Gordon, &

Baer, 1988; Shiffman et al., 1985; Stevens & Hollis, 1989).

In a randomized study using skills training and social

support as relapse prevention methods, Stevens and Hollis

(1989) found skills training significantly decreased relapse

rates, while social support showed no significant differences

from a no-treatment control.

In addition to skills training, the National Institutes

of Health (NIH) say information about the health and social

consequences of smoking is critical for cessation (Glynn et

al., 1991). Gibbons, McGovern, and Lando (1991) showed the

utility of including information on the risks of smoking as

well as the benefits of quitting, in a study of relapse in a

smoking cessation clinic. The NIH study states high risk

perception declines significantly over time for those who

fail in their quit attempt. Perception is an important

factor related to cessation commitment and efforts to

counteract this tendency to decline may reduce relapse rates,

or may help relapsers maintain cessation motivation to try

again. Several studies suggest booster sessions may help

maintain motivation to stay quit when used as part of a

multicomponent strategy (Brink, Simons-Morton, Harvey,

Parcel, & Tiernan, 1988; Brownell et al., 1986). Brownell et

al. (1986), suggest booster sessions may be useful in the

perception that one is part of a continuing process of

change. They also point to the usefulness of giving self-

help manuals to those who have quit, as another means of

enhancing vigilance in maintaining non-smoking. They believe

use of self-help manuals may be especially important in the

late maintenance stages of quitting. They state that

although initial abstinence rates from self-help books and

brochures are low, such materials may be effective in keeping

recent quitters from returning to smoking. Curry et al.

(1988) also found self-help materials were helpful in relapse

prevention, for those who used the materials. Motivation is

another factor which has been cited by many studies as

critical in maintaining cessation (Brandon, Tiffany,

Obremski, & Baker, 1990; Brownell et al., 1986; Carmody,

1990; Glynn et al., 1991; Shiffman et al., 1985).

A task force on interventions to prevent relapse

(Brownell et al., 1986) suggest that sustained smoking

cessation requires the modification of a range of social

skills. The removal of an ingrained personal and social

habit leaves a void that must be filled to prevent relapse.

They suggest the void be filled by new reinforcers, new

social skills, or general lifestyle changes. New, more

health oriented reinforcers, may include; exercise,

relaxation training, and meditation. An effort must be made

to find alternatives for which smoking was previously


Smoking Policy and Effect on Smoking Cessation

Navy recruit training lasts 8 weeks. The current

smoking policy does not allowed anyone to smoke at the

command (CNTECHTRA Instruction 5100.6A, 1991). This includes

recruits, instructors, and all staff members. Therefore, the

recruits who smoked prior to entry, graduate boot camp as ex-

smokers who have successfully quit for 2 months. This 100%

cessation rate, with zero recidivism for 2 months, is far

better than any other cessation program reported in the

literature. However, no other quasi-involuntary cessation

programs have been reported and very little information is

available concerning relapse after leaving this type of

environment. Cronan et al. (1989) compared smoking

intervention techniques in recruit training (the current

smoking policies had not yet been instituted). Four groups

were followed through their training cycle. One group

received an educational intervention, another feedback from

health risk evaluations, and a third was designated no

smoking and were not allowed to smoke during training. A

fourth group was followed as a control. The no smoking

group, although showing the best cessation at the end of

training (measured after their first opportunity to smoke),

still reported a 75% relapse rate.

The only other study examining relapse rate in this area

is ongoing (T. L. Conway, personal communication, June 6,

1991) and results are not yet available. This study began

last year to look at tobacco use in all new Navy accessions.

Participants were surveyed at the beginning of training, at

graduation, and at 1 month follow-up. Interventions were not


The current Navy policy is not as restrictive overall as

it is for recruits, but does meet the Department of Defense

directives. Tobacco use is not permitted in: 1) auditoriums,

conference rooms, classrooms, or libraries; 2) elevators; 3)

official buses, vans, and shuttle vehicles; and 4)

gymnasiums, child care and youth activity centers (SECNAVINST

5100.13A, July 17, 1986). Other limitations are placed on

working areas, eating facilities, aircraft, living quarters,

and medical facilities, but guidelines are left to the

individual Commanding Officer. In contrast though, sales of

tobacco products are still discounted from civilian prices at

Navy Exchange facilities, although they may not be purchased

at medical and dental facilities (USDHHS, 1990b).

There is no evidence that policy alone affects smoking

prevalence. In fact, the 1989 Surgeon General's report

(USDHHS, 1989) reported insufficient research has been

undertaken to determine what effect, if any, smoking

restrictions have had on smoking prevalence. Two studies did

examine a no-smoking policy which has been successful in

decreasing consumption of cigarettes in settings affected by

the policy, but not prevalence rates (Pentz et al., 1989;

Petersen et al., 1988). Policy combined with education does

seem to be more beneficial than policy alone. Pentz et al.

(1989) found a policy which emphasized adolescent education

to prevent and stop smoking, rather than regulating it by

punishment, was significantly related to lower recent

smoking. Education, designed to reorient the ex-smoker's

attitudes toward more realistic expectations about quitting

smoking, is an important first step in relapse prevention

(Carmody, 1990).

A survey to assess short-term impact of Army smoking

policies (Carroll et al., 1989) reported no influence of

policy on decision to quit, though a consumption drop was

reported in some areas. This study also showed intention to

quit is related to knowledge of smoking-related diseases.

The current recruit smoking policy lacks support of a strong

prevention education component (CNTECHTRA, 1991a). Results

of the current recruit policy study may reflect those of the

Army survey because of this fact. The Navy has a long way to

go to meet its goal of being smoke free by the year 2000

(Zolton, 1991). New restrictions are being proposed, but

Navy guidelines must include positive approaches to help

individuals make the decision not to smoke. Vice Admiral

Zimble, the Navy Surgeon General, said: "We want a

positively motivated force that chooses not to smoke"

(Nelson, 1991). Health educators also realize coercion of

positive health behavior is not only unethical, but is

usually not a very successful tactic. The challenge, in this

study of prevention education supplemented with strict

policy, will be to increase the likelihood that recruits will

favorably respond to the education treatments.

Other Relapse Research

Though much is known about relapse in self-quitters and

program aided cessation, very little is known about relapse

in those who involuntarily quit smoking. One related area is

relapse of women who quit smoking during pregnancy. These

women quit voluntarily, but usually not for themselves.

Rather they quit so as to not affect the health of their

babies. Relapse in these women is usually quite high

(McBride & Pirie, 1990; Mullen, Quinn, & Ershoff, 1990) and

has been reported as high as 70% at one year post-cessation

(Fingerhut, Kleinman, & Kendrick, 1990) even though smoking

at this time may still have an indirect effect on the baby.

Several studies have been conducted on smoking cessation for

pregnant women, but none have specifically targeted relapse


Another area where involuntary smoking cessation may

occur is in jails/prisons, but no information could be found

in the literature on this subject.

Smoking Prevention

As more and more adult smokers quit, in both the

civilian and military populations, efforts to prevent the

onset of smoking become the next logical step towards the

goal of a smoke-free Navy/society. The Surgeon General

(USDHHS, 1989) reported children and adolescents hold the key

to progress toward curbing tobacco use in future generations.

This report also said there is a growing recognition that

prevention and cessation efforts need to target specific

populations with a high smoking prevalence and a high risk of

smoking related disease. The military population is listed

as one of the groups that needs to be targeted.

Many approaches have been tried over the years, both in

school and community settings, to prevent adolescents from

beginning to smoke. Early campaigns to increase information

and create fear of long term consequences did not succeed.

Although succeeding in changing knowledge, beliefs, and

attitudes, very few programs consistently reduced the onset

of smoking (Flay, 1985). In a review of psychosocial

approaches to smoking prevention, Flay (1985) examined the

success of such programs and the theory involved in their

evolution. He states these programs incorporate complexities

of the psychosocial process of becoming a smoker into the

design of prevention interventions. He lists the following

components as part of a life/social skills approach to

smoking prevention:

-long term consequences of cigarette smoking

-the prevalence of smoking

-correcting students perception of social norms

concerning smoking

-teaching students about the social influences to smoke

-providing behavioral skills to resist influences

-enhancing personal skills and/or self-esteem

-providing general social competence skills.

Flay cites the social inoculation theory as the main

influence for this approach, with added theoretical

bolstering from the attitude change theory and the social

learning theory. Among the program components he felt were

important, based on successful studies, were:

-information on the immediate effects of smoking

-correction about the misrepresentation about the

prevalence of smoking

-discussions of family and media influences on smoking;

and how to deal with them

-a public commitment procedure.

The health belief model is one of the few social

psychological models to be developed expressly to understand

health behaviors (Rimer, 1990). One of the critical

dimensions of this model concerns the failure to believe in

the possibility of having pathology in the absence of

symptoms. This is particularly true with cigarette smokers,

especially young ones. Many of the health problems of

smoking are long term (USDHHS, 1989; USDHHS, 1987). Lung

cancer, emphysema, and cardiovascular diseases do not develop

within the first few years of smoking. Young smokers show a

good knowledge of the connection of smoking with these

diseases (Banspach, Lefebvre, & Carleton, 1989; Johnston et

al., 1989). However, because they are not experiencing any

of the disease symptoms, they feel these effects are a long

way off, and they will be able to stop smoking well before

disease onset. In general, adolescents feel invulnerable to

accidents and chronic diseases (Glanz, Lewis, & Rimer, 1990).

One of the axioms of health education is that knowledge

is necessary, but not sufficient for behavior change.

Information is necessary but not sufficient for knowledge

(Rimer, 1990). Therefore, teaching young students about the

long term effects of smoking may be a good tactic, but by

itself, will not win the prevention war.

In order for people to quit smoking for health reasons,

they must believe cessation will benefit their health, and

also that they are capable of quitting (Rosenstock, 1990).

Knowledge about the risks of smoking is better now than it

ever has been. The 1989 Surgeon General's report (USDHHS,

1989) states the proportion of adults who believed cigarette

smoking increases the risk of emphysema and chronic

bronchitis rose from 50% in 1964 to 81% (chronic bronchitis)

and 89% (emphysema) in 1986. These proportions increased

among current smokers from 42% in 1964 to 75% (chronic

bronchitis) and 85% (emphysema) in 1986. Johnston et al.

(1989) reported that 70% of young adults perceive regular

pack-a-day cigarette smoking as entailing high risk. Despite

these findings, and despite the fact that 80% of all smokers

indicate a desire to quit, over 50 million Americans still


Not all adolescents who try a cigarette become regular

smokers; not all who start smoking regularly continue to do

so their entire life. Once an individual has experimented

with smoking, other factors come into play to determine

whether the behavior will become a continuing habit. New

smokers do not become regular smokers immediately. They pass

through the phases of stabilization, acceleration, or

deceleration of their smoking habits (Epstein, Grunberg,

Lichtenstein, & Evans, 1989). Horn (1979) describes the

influence of three groups of factors which bias this choice:

the cost/benefits evaluation of the behavior; common

stereotypes that characterize perception of behavior; and

psychological factors characterizing both personal structure

and personality integration factors, particularly as they

relate to societal demands. The cost of smoking includes the

harmful effects on health, financial cost to the individual

or society, and more recently, a feeling of discrimination in

most public places.

Benefits a smoker may perceive vary even more than the

disadvantages (Horn, 1979). Peer acceptance or facilitation

of social interaction places high on the list for

adolescents. Smokers often smoke to reduce stress, enhance

pleasure, improve their concentration, gain satisfaction from

handling the cigarette, or for numerous other reasons. For

many smokers the benefits do seem to outweigh the costs. The

perceptual stereotypes referred to by Horn develop from a

mystique about what smoking and smokers are like, and why

people smoke. Horn states that, the greater the role played

by these superficial and inaccurate beliefs about the

behavior, the more difficult it becomes for an individual to

develop a sound decision-making process. These stereotypes

are perpetuated by cigarette advertisements. Horn's third

group of factors contains a variety of patterns for

psychological forces which may help determine personal

choices for health behavior, particularly behaviors which

reflect a conflict engendered in individuals between the

demands of society and their own inner desires.

The population entering the Navy consists of current

smokers, former smokers, occasional smokers, and non-smokers

(Cronan & Conway, 1988; Cronan et al. 1989). Studies have

found nearly all smokers start smoking in their teenage years

(USDHHS, 1987; USDHHS, 1989). However, the National Cancer

Institute (NCI) now feel there is a second period where

individuals are at risk for starting smoking. Recent

research indicates there may be another smoking uptake period

in the early 20's as employment transition takes place (T. J.

Glynn, personal communication, September 16, 1991). Cronan

et al. (1989) found that some recruits were starting to smoke

during recruit training (this study was done before the

current no-smoking policy was instituted). It is possible,

since the institution of the no-smoking policy in recruit

training, that smoking initiation may now be taking place in

the training environment after recruit graduation. Because

of possible smoking uptake during employment transition, it

is important that prevention efforts also be presented during

this period. Cronan & Conway (1988) concluded the Navy must

develop and evaluate programs directed toward preventing

personnel from smoking. They felt prevention programs should

probably be implemented as early as recruit training.

Educational Intervention

Because of the mix in smoking status among recruits, a

similarly mixed intervention/prevention strategy may prove

most effective. In a meta-analysis of 39 smoking cessation

interventions Kottke, Battista, DeFriese, and Brekke (1988)

found the most effective programs employed more than one

modality for motivating behavior change. Cronan et al.

(1989) evaluated smoking interventions in recruit training.

They found a one-hour education program significantly reduced

the number of recruits who started smoking during training,

but did not increase the number of smokers who quit. Their

educational intervention stressed the hazards of smoking and

techniques for stopping.

In 1987, the NCI convened an expert advisory panel to

assess the question "What are the essential elements of a

school-based smoking prevention program?" (Glynn, 1989). The

panel concluded that existing programs have been consistently

effective in delaying the onset of smoking. Programs with a

smoking-only focus have about an equal effect as those with a

multicomponent focus. They felt minimum smoking prevention

components should include:

-information about social consequences and short term

physiological effects of tobacco use

-information about social influences on tobacco use,

especially peer, parent, and media influences

-training in refusal skills, including modeling and

practice of resistance skills.

They felt the minimum length of school-based programs should

be two, 5 session blocks with booster sessions in subsequent

years. Doctor Glynn, of the NCI, stated a shorter program

(2-3 hours) would probably be effective with individuals who

had already demonstrated to themselves that they could stop

smoking (T. J. Glynn, personal communication, June 18, 1991).

When used as part of a cessation program, education

intervention may last from 1-4 hours, depending on the length

of the overall program. No studies have been reported in

which length of the education segment was varied, while

attempting to decrease relapse rates or prevent smoking.

The Waterloo Smoking Prevention Project (Flay, 1985)

found previous experience with smoking did not preclude

prevention of further smoking. The study results showed the

program influenced current experimenters to quit smoking, and

remain quit. The prevention curriculum was six hours long

and was designed to influence attitudes towards tobacco and

the acquisition of social skills. The program had three main

components: 1) provide information on the effects of

smoking, 2) focus on social influences to smoke (family,

peer, media), and 3) decision to integrate information on

smoking to individual's own smoking environment and a public

commitment procedure about their decision of whether or not

to smoke. Three booster sessions were included in the


Banspach et al. (1989) assessed two smoking prevention

programs: 1) a five-lesson, video presentation on resisting

pressures to smoke, and 2) a four-lesson teacher led

discussion of smoking and advertising. They found both

programs had a similar positive effect on preventing smoking

in students. The effects seen were significantly different

from those observed in a control group. They concluded that

increasing a student's awareness about the messages they

receive can be enhanced by either a video presentation with

discussion, or a teacher led discussion with student


Validation of Smoking Status Self-Report

Cigarette smoking behavior is most often measured by

self-report of frequency and intensity. Researchers find

this procedure simple, efficient, and inexpensive. Murray,

O'Connell, Schmid, and Perry (1987) state because many

adolescents smoke infrequently and episodically, it may be

difficult for them to characterize their usual pattern of

smoking. Adolescents may also try to mislead the

investigator and underreport smoking behavior out of

embarrassment or a desire to please. Self-reports of smoking

status therefore, may not always be accurate, particularly

where smokers feel pressure to give up smoking but may not

have achieved it (Jarvis, Tunstall-Pedoe, Feyerabend, Vesey,

& Saloojee, 1987; Murray et al., 1987). In a study to

validate self-report claims, Jarvis et al. (1987) reported

19% of cotinine confirmed smokers claimed to be non-smokers.

However, other studies found only 2% of smokers claiming to

be non-smokers when not actively persuaded to give up smoking

(Lee, 1987; Thompson, Stone, Nanchahal, & Wald, 1990). For

this reason researchers try various techniques to increase

truthfulness of reporting and to validate self-reports of


Bogus Pipeline

One strategy shown effective in increasing validity of

smoking self-reports has been termed the bogus pipeline

method. The rationale for this approach is that individuals

should be more likely to disclose their actual smoking

behavior if they believe the researcher has an independent

and objective means of measuring smoking behavior. Evans,

Hansen, and Mittelmark (1977) were the first to show efficacy

of this procedure with smoking research. Murray et al.

(1987) examined eleven papers on this subject and found five

reported a pipeline effect, while six did not. They

conducted a study to determine effectiveness of the approach

and effectiveness of different methods of message delivery.

Their data supported the hypothesis that adolescent smokers

are more willing to disclose cigarette use under pipeline

conditions. They concluded it is both prudent and

conservative to employ as credible a pipeline procedure as

possible, and use of this procedure offers the best assurance

of a valid assessment.

Biochemical Tests

A number of biochemical markers can be used to validate

smoking self-reports including measures of thiocyanate,

carbon monoxide, nicotine, and cotinine (Fitzpatrick, 1991;

Jarvis et al., 1987; Sepkovic & Haley, 1985). Levels of

carbon monoxide and thiocyanate (a metabolite of hydrogen

cyanide) are easier and less expensive to determine, but may

be raised through exposures unrelated to smoking, such as

auto emissions and diet (USDHHS, 1988; Jarvis et al., 1987).

Neither can they be used as a marker for smokeless tobacco.

All tobacco products contain substantial amounts of nicotine,

which is absorbed from tobacco smoke in the lungs, and from

smokeless tobacco in the mouth and nose. Nicotine is

metabolized in the liver and lungs, and cleared from the body

by the kidneys (USDHHS, 1988). Measures based on nicotine

have the advantage of being tobacco specific (all forms of

tobacco), but require extensive laboratory instrumentation.

Nicotine is a good marker of tobacco use, but its short half-

life in serum (0.5 2 hours) makes it a poor choice.

Nicotine is metabolized to cotinine, trans-3-hydroxycotinine,

and nicotine-l-N-oxide; all of which can theoretically be

used as markers (USDHHS, 1988).

Most investigators agree that cotinine is the analyte of

choice because it is tobacco specific and has a relatively

long half-life (Haley, Axelrad, & Tilton, 1983; Jarvis et

al., 1987; Langone, Cook, Bjercke, & Lifschitz, 1988). The

half-life of cotinine is on average 20 hours (Jarvis et al.,

1987; Langone et al., 1988) which makes sample timing in

relation to cigarette smoking less critical. Cotinine

concentrations in blood are generally tenfold greater than

nicotine, and as a result less sensitive analytical

methodology may be acceptable. Cotinine is metabolized, only

17% is excreted unchanged, but levels may persist for up to

seven days after cessation of habitual smoking (USDHHS,


Cotinine samples from blood, saliva, and urine are

equally applicable to the whole range of issues requiring

estimates of nicotine exposure from tobacco smoking (Jarvis

et al., 1987). Choice of fluid may depend more on practical

considerations (eg. cost, collection availability) than

pharmokinetic considerations. Because concentrations of

cotinine in urine are tenfold to hundredfold greater than

concentrations in plasma or saliva, a variety of analytic

techniques meet sensitivity requirements (USDHHS, 1988).

Without use of a 24 hour urine sample though, normalization

using creatine excretion is often used (Haley, Colosimo,

Axelrad, Harris, & Sepkovic, 1989). This normalization

requires additional analysis and thus higher costs. Haley et

al. (1989) concluded the additional analyses may not be

necessary when simple validation of nonsmoking status in a

smoking cessation program is the endpoint.

Several analytical techniques have been used to

determine cotinine concentration in biological fluids. The

most frequent used are radioimmunoassay, liquid

chromatography, and gas chromatography (Skarping, Willers, &

Dalene, 1988). The choice depends on the biological fluid to

be assayed; the need for sensitivity, precision, and

accuracy; and economic considerations (USDHHS, 1988).

Immunoassay techniques are simpler, generally require smaller

samples, and may be less expensive. They have been

determined to be best suited for smoking/non-smoking


Non-smokers can show low levels of cotinine from

nicotine exposure to environmental tobacco smoke (eg. car,

home, restaurants). Studies have shown passive cotinine

concentrations as high as 32 ng/ml, and active cotinine

concentrations as low as 44 ng/ml (Wall, Johnson, Jacob, &

Benowitz, 1988). Sensitivity for non-smoker analysis is

generally set in the range of 1-25 ng/ml (Thompson et al.,

1990), however cutoff sensitivity and specificity set for

discriminating true has also been set at 50 ng/ml (Jarvis et


al., 1987). Other studies have also set non-smoking values

at 50 ng/ml (Fitzpatrick, 1991; Jarvis et al., 1987).

Fitzpatrick sums up nicely the benefits of cotinine


The cotinine assay used today can determine either
active or passive exposure to tobacco smoke. It is
specific for exposure to nicotine. It can use urine,
which like saliva, is easy to collect, but it can also
use serum. It has a long half-life of 16 hours so that
a person cannot easily prepare for the test by not
smoking for short periods--ideally, an assay should be
able to detect an individual who has not smoked for 48
hours. It is inexpensive, so that it may be performed
routinely in large screening programs. (Fitzpatrick,
1991, p.11)

There is no biochemical test which can be expected to

yield perfect separation of non-smokers passive smokers, and

active smokers under field conditions. Wall et al. (1988)

give several reasons for this: intersubject variance in

nicotine metabolism, time of day of sample collection,

underreporting of active smoking, adjustment of cigarette

consumption for nicotine content, and perhaps over or

underreporting of passive smoke exposure. However, the use

of biochemical analyses definitely increase the validity of

smoking self-reports.

Concluding Statement

Because of the impact of health, health care costs, and

readiness, the Navy Surgeon General has set a goal of a

Smoke-Free Navy by the year 2000. Prevention efforts in new

accessions, as well as cessation efforts throughout the

ranks, are critical in meeting this goal. Currently the Navy

recruit population shows a higher smoking prevalence than the

general U. S. population. The no-smoking policy in Navy

recruit training provides a perfect opportunity to allow

smokers to stop, but many factors prevent those who quit from

staying quit. Current research reveals many approaches that

show promise in helping these individuals to stay quit, and

non-smokers from ever starting.

Smoking is becoming more and more socially unacceptable,

both in general society, and in the military. However, if

smokers are expected to quit, they must be provided

assistance. Setting policies that provide an environment

conducive to not smoking are beneficial, but alone don't show

high cessation rates. Education programs are an essential

adjunct to no-smoking policies. Quitting smoking is very

difficult and many people cannot do so on their own. Smoking

prevention and smoking relapse prevention programs must be

developed with the characteristics of the Naval population in

mind. They should draw from existing theory and current

research. Programs should provide information designed to

assist individuals in changing behavior intentions to more

health oriented ones. They need to recognize that all

smokers/non-smokers are in different stages of behavior

change and offer varied approaches to reach the entire

audience. This is especially true when groups are to be

addressed and smokers cannot be separated out. Motivation is

crucial in any smoking cessation effort; programs must be

designed around providing motivational incentives and

building self-efficacy.

Health behavior change programs should be guided by
social learning theory to encompass environmental, personal,

and behavioral constructs. Participants must decide for

themselves that smoking is potentially detrimental, recognize

the pressures on them to smoke, and develop their own ways to

cope with these pressures. Alternatives must be presented in

order for informed choices to be made. Recruit smoking

education programs should build on what many recruits will

have already had in their formal education. Misconceptions

about tobacco use must be corrected. Long-term consequences

of smoking should be discussed, but short-term effects also

need to be stressed. Current Navy policy, and future policy

direction, should be discussed to give the new Navy member a

picture of what will be expected of them. Recruits who

entered as smokers have to believe they have actually quit

smoking, if they are expected to stay quit after graduation.

Recruit training is a period when a complete lifestyle

transformation is taking place. Recruits are making

wholesale changes in the ways they live, work, and socialize.

They are taking on added responsibilities, receiving personal

and financial independence, and adjusting many health

practices. They should be encouraged at this time to make

decisions, in all areas, that will lead to self-fulfillment

and better health.



This chapter describes the methods employed to conduct

the study. Subjects are described with a discussion of the

sampling methods used. The survey instruments are described

and validity and reliability are discussed. A detailed

description of the study procedures is given including survey

administration, educational interventions, and the booster

program. A discussion of self-report validation techniques

incorporated in the study is then given. Finally, a

description of how data were handled is presented along with

the plan for data analysis.


Seven companies of recruits from the Navy Recruit

Training Center (RTC) in Orlando, FL were used in this study.

Five companies were male and two female. Participants were

357 men and 139 women. The mean age of participants was

19.26 years (standard deviation [SD] = 2.27). Over 99% held

a high school degree or equivalent, 22% had some college or

were trade/technical school graduates, and 2% had a four year

college degree. The mean Armed Forces Qualification Test

(AFQT) score was 65.76 (SD=18.62). The AFQT score provides a

measure of general trainability of recruits. The Navy

requires a minimum AFQT score of 17 with a high school

degree, 31 for other credential holders, and 50 without a

high school diploma (Foley & Rucker, 1989). Seventy-eight

percent of participants were white, 15% were black, 5% were

Hispanic, and 2% were from other ethnic backgrounds.

Recruit training is an intensive 8 week program designed

to indoctrinate the new members as to the ways of Navy life

and prepare them for military duties. The mission of RTC

Orlando is to conduct a training program which will effect a

smooth transition from civilian to Navy life; foster

patriotic behavior; affirm the dignity of the individual;

encourage high standards of personal responsibility, conduct,

manners, and morals; create a desire for self-improvement and

advancement; provide the recruit with knowledge and skills

basic to all naval personnel; develop pride in the unit and

in the Navy and a desire to observe appropriate naval

customs, ceremonies, and traditions; and provide the

Department of the Navy with personnel possessing an effective

level of physical fitness. The mission is carried out

through an intense training and educational program.

The educational program includes health/hygiene classes,

and classes on substance abuse. The current curriculum

however, includes no specific tobacco-use prevention

information (CNTECHTRA, 1991a). There are approximately 75 -

85 recruits in each company and about 5 classes start every

week. A recent tobacco use survey showed that about 28% of

entering recruits are smokers (Cronan & Conway, 1988), which

indicates about 23 members of each study company should be


The subject sample was drawn from the total recruit

population. Because of time constraints a purely random

sample of all companies was not feasible. One company,

chosen randomly from a weekly starting group, was used as a

pilot group and went through the research program before the

main study group. The other 7 companies were chosen from all

those starting during adjacent weeks in September 1991.

At RTC, companies are sectioned into training groups.

Each training group contains all companies starting during a

given week (usually 6-7 companies). The training group

progresses through recruit training at the same rate and

graduate on Friday of their eighth week. Companies are

filled as recruits arrive at RTC, without segregation except

by gender. This random assignment to companies allowed for a

random recruit sampling using any training group. Training

Group 47 was selected because of projection for 5 male and 2

female companies. This training group started training

during the week of September 9, 1991. However, the second

female company anticipated did not fill until the following

week, thus the second female company used was part of

Training Group 48.

In an effort to determine if the study population was

representative of the entire Navy recruit population

demographic data were obtained from Naval Technical Training

Command data bases (CNTECHTRA, 1991b). Table 2 is a

comparison of the study group to 6 months of all recruits

going through RTC Orlando, and all 3 Navy RTC's for several


Table 2. Study group comparison
Education Average Average Percent
level AFQT* Age Minority

All RTCs 6 mo 11.69 61.04 19.51 19.37
RTC Orlando 12.07 67.89 19.42 17.8

T-1 12.05 66.07 19.48 19.12

T-2 12.09 68.48 19.41 14.94

T-3 12.02 64.78 19.40 23.48

Sham 11.95 60.57 20.10 26.58

*Armed Forces Qualification
to all services

Test (AFQT) is required for enlisted entry

This table shows the companies chosen for the study were

fairly representative of all companies going through RTC

Orlando, and the rest of Navy Recruit Training Centers.


Smoking history and relapse information were collected

via pencil and paper survey. Surveys were conducted during

the recruits first week of training (P-4 day), after

graduation (8-3 day), and 3 months after graduation. The

surveys used were modified versions of those by the Naval

Health Research Center for their "new accession tobacco use



The Horn-Waingrow Smoking Motives Questionnaire (Girdano

& Dusek, 1988) was added to the intake survey (see Appendix

A) to examine type of smokers who are best helped by this

program. The Fagerstr6m nicotine tolerance questionnaire

(Fagerstr6m, 1978) was also added to examine nicotine

addiction. Questions were added to assess knowledge of and

attitudes toward tobacco use.

The graduation survey (see Appendix B) was much shorter

than the intake survey. Most of the questions were repeats

from the first survey with only a few added to assess change

in knowledge and attitudes. The 3 month follow-up survey

(see Appendix C) was a modified version of the Naval Health

Research Center's 12 month follow-up survey with questions

added to measure change in attitude and knowledge.

Validity and reliability

A test-retest reliability analysis was accomplished with

the intake survey. An extra company (N=84) was chosen to be

a pilot. This company was administered the intake survey on

their P-4 day. They were encouraged to bring any questions

or concerns to the attention of the researcher. All problems

were noted and incorporated into the briefing for the main

study companies. This group completed the same intake survey

three days later. Key questions to be used in the main study

analysis were analyzed here as part of an instrument

reliability test-retest procedure. The questions were

examined for frequency of non-agreement, then Spearman's and

Pearson's correlations were performed to compute reliability.

Table 3 summarizes the results of the instrument reliability

measurements. Nine of the thirteen items tested showed a

correlation of .80 or better (Spearman) and six of twelve

showed a correlation of .80 or better (Pearson). Questions

that showed poor agreement were explained in the briefing to

the main study companies and the wording was changed on some

for subsequent survey versions. In the study analysis

Questions 10 and 11 were scored as correct if responses below

and above exact percentages were answered. Question 5 showed

a low correlation mainly with non-smokers. For this reason

attitude scores were analyzed for all subjects and for

smokers only.

Table 3. Test-retest reliability
Frequency of
Question # Non-agreement Spearman's R Pearson's R

1 10% .93 .90

2 12% .92 .89

3 10% .90 .95

4 9% .75 .75

5 17% .51 .51

6 27% .79 .76

7 34% .65 .62

8 14% .71 .71

9 17% .82 .68

10 38% .80 .79

11 31% .89 .88

12 17% .86 .86

13 25% .92 .93
Question/Variable Key:
1. Smoking status at entry
2. Smokeless tobacco use at entry
3. During the last 30 days prior to entering the Navy how many
cigarettes did you usually smoke on a typical day when you
smoked cigarettes
4. Do you believe cigarette smoking is related to heart disease
5. Are you concerned about the health effects of cigarette
6. How important is it to you to be a non-user of tobacco
7. Over the next year, how likely is it that you'll stop smoking,
if you now smoke
8. In the future, do you see yourself as someone who smokes
9. Would you be in favor of being placed in a smoke-free work
10. What percent of the U.S. population do you believe are current
11. What percent of all members of the Navy do you believe are
current smokers
12. Should tobacco companies be allowed to target advertising
toward certain groups in order to recruit new customers
13. Fagerstrom nicotine tolerance level

This project, including the instruments and consent

statement, was reviewed and approved by the University of

Florida Institutional Review Board (see Appendix D). The

instruments were destroyed after data analysis to satisfy

Institutional Review Board requirements.

Self-Report Validation Techniques

During the instruction period prior to completion of the

intake survey, participants were briefed on possible self-

report cross checks using urine tests. It was explained that

some of their urine samples submitted upon entry would be

screened for a substance which would tell us their smoking

status. They were also told future random drug screen

samples also might be used to validate follow-up self

reports. Subjects were assured the results would in no way

negatively affect them, but that this procedure was a

requirement of an academic research committee. It was

anticipated this bogus pipeline approach would increase

truthfulness of self-reporting for all three surveys.

Cotinine analysis of urine specimens was accomplished

for a random sampling of participants in the intake and

graduation surveys. Urine specimens provided as part of the

Navy drug screening program were used in the study. All

recruits are required to provide a urine sample within hours

of reporting to recruit training. These samples are tested

locally by means of radioimmunoassay and positive samples are

sent to Navy drug screening laboratories for gas

chromatography-mass spectrometry screening. The samples

testing negative locally are normally disposed of on site.

Normal RTC procedures require that the day after uncontrolled

liberty (after recruit graduation) one company be randomly

selected for a unit drug screen. The same procedures are

again followed for sample collection and testing. For the

purposes of the present study only drug negative samples

provided the basis for a sampling pool.

Fifty urine samples were selected randomly (20 female

and 30 male) by drug screen lab personnel, from participating

companies upon arrival at Recruit Training Command Orlando.

Approximately 10 ml were transferred to a storage vial and

frozen. The same procedure was followed for the company

chosen for the post-liberty recruit urinalysis. Forty-two

samples were collected at this time. All samples were kept

frozen until analyses could be conducted.

Cotinine analysis was conducted utilizing Florescence

Polarization Immunoassay technology. An Abbott TDX automated

fluorescence analyzer (instrument #2, serial L30179) at the

Shands Toxicology laboratory in Gainesville, Florida was

used. Cotinine reagent was purchased from Abbott

Laboratories and was within freshness parameters. The

analyzer was calibrated by Shands lab personnel, according to

the operations manual. Low, medium, and high controls were

run after calibration and results were within acceptable

limits. Each batch analysis run included at least one

control, and all control results were within acceptable


Results were reported in nanograms per milliliter

(ng/ml). The sensitivity of this analysis, defined as the

lowest measurable concentration which can be distinguished

from zero with 95% confidence, was determined by the

manufacturer to be 50 ng/ml. Indication of smoking was then

set at any result greater than 50 ng/ml. Results were

compared against the self-reported smoking reported on

respective surveys.

Results of cotinine analyses were in 100% agreement with

self-reported tobacco use at pretest, and 96% agreement with

self-reported tobacco use at posttest. This reflects only a

1% misrepresentation of smoking status overall. One subject

indicated no tobacco use but tested high for cotinine; two

others indicated tobacco use but tested low for cotinine.

Self-reported tobacco use and cotinine chemical analyses were

highly correlated (r=.91).


After the 7 companies were selected they were randomly

assigned to control or treatment categories. Treatment group

assignment was made before any of the survey data was

processed. One extra company was selected to serve as a

pilot group. This group was used to assist in analysis of

instrument validity and reliability, curriculum

effectiveness, and to uncover any unforeseen problems. Table

4 presents an overview of treatment group assignments.

Table 4. Treatment group assignments

Pilot T1 T2 T3 T4

Company C-175 n=84 C-184 n=81 C-181 n=83 C-183 n=82 C-185 n=79

Number K-079 n=80 C-182 n=88 K-078 n=84
Notes: T1 = Policy only
T2 = Policy + education
T3 = Policy + education + booster
T4 = Policy + sham treatment
Pilot = Policy + education

Intake surveys

All participating companies completed consent forms and

intake surveys on their P-4 day of training (1st week at RTC)

in the same classroom, at the same time of day. Participants

were briefed on what they would be asked to do in the project

and asked to read the consent statement (see Appendix E).

The researcher read the statement aloud to make sure the

subjects knew their participation was voluntary. Information

confidentiality was assured at this point. The intake survey

asked for full name and social security number, but in an
effort to assure confidentiality, the follow-up surveys

requested only the last 4 digits of the social security

number as identification. After the explanation, questions

were taken about the project and the consent form.

Participants were then asked to sign the consent forms if

they agreed to participate in the study. The signatures were

witnessed, dated, serialized, and retained for a period of 5

years. The statement portions of the consent forms were

collected and given to the company yeomen (recruit in charge

of all company paperwork) for later re-distribution to the

participants. The intake surveys (see Appendix A) were then

completed by the participants. The pilot group participants

were administered the same intake survey approximately 72

hours later.

Education intervention

The education intervention phase was conducted for

participants on their 2-2 day of training. The curriculum

(see Appendix F) consisted of lecture, video-tapes, visual

aids, demonstrations, and group discussions. Participants

were encouraged to ask questions and participation was

solicited during large group discussions.

The main lecture topic areas were:

*Navy tobacco use policies

*myths about tobacco use

*physiological effects of smoking (short and long


*reasons for choosing to use/not use tobacco

*benefits of not starting/resuming tobacco use

after RTC graduation.

The videos used were:

*"The Performance Edge", from the U.S. Department

of Health and Human Services. It shows the effect

of tobacco and alcohol on performance. A

motivational pitch for non-use is given

*"The Feminine Mistake", from Pyramid films. Shows

the physiological effects of smoking, benefits of

quitting, and long term hazards of smoking

*"Clearing the Air", from the U.S. Navy. Shows

the effects of smoking, Navy tobacco use policies,

and strategies to quit and stay quit.

The videos used were previewed by 12 recruits 7 weeks before

the education intervention began. Positive feedback was

received for their use in the curriculum.

Posters were displayed throughout the classroom. Poster

placement was the same for each lecture period. Those used

were from the American Cancer Society, and included:

*Animals-"It looks as stupid when you do it"


*Animals-"Butts are gross" #F-652/90

*"Smoking is very debonair" #2163.03-LE

*"Smoking is very sophisticated" #2163.02-LE

*"Smoking is very glamorous" #2163.01-LE

*"Smoking doesn't work" #2418.04-LE

*"Life, the only race you don't win by finishing

first...Don't smoke" #2122.LE

*"Are you a draggin lady?"

*"12 things to do instead of smoking" #2106.LE.

Other visual aids and demonstrations were used during

the lecture periods and the breaks. The following teaching

aids were utilized:

*Consequences of smoking--Display which

graphically depicts detrimental physical effects of

smoking, such as emphysema, lung cancer, and heart


*Second-hand smoke demonstrator--Device which

shows the effect of sitting in a closed room with

smokers, by filtering smoke passively released from

a smoked cigarette

*Mr. Grossmouth--Display which shows the physical

effects of smokeless tobacco to the mouth, tongue,

teeth, and palate

*American Cancer Society pamphlet "50 questions

about smoking".

The education intervention was taught in the same

classroom each time and by the same instructor (the

researcher). Classes started at 7:00 a.m. and ran until

10:00 a.m. Students were told to relax and that the material

was not testable, but they were free to take notes. The

temperature of the room was maintained at 780 F and the

students were allowed to stand, or walk to an in-classroom

drinking fountain, if they got drowsy. Two 15 minute breaks

were given, during which students were asked to discuss

topics for subsequent group discussions. The students were

also encouraged to look at the displays and posters during

the breaks. No unusual external events took place during any

of the class periods.

The entire treatment program was designed to motivate

recruits (both smokers and non-smokers) to not initiate

smoking after graduation. A recent personal communication

with Dr T. J. Glynn (September 16, 1991), of the National

Institutes of Health indicated that the primary component of

this program should be motivation to maintain nonsmoking

status, and to want to do so. The benefits of remaining non-

smokers were stressed, and ex-smokers were challenged to make

the decision to not start again. The program was designed to

instill confidence in the recruit's ability to remain smoke

free. It was stressed that the Navy is anti-smoke and not

anti-smoker, and only the individual can make the choice of

being smoke-free. Every effort was made to ensure the

intervention program was not viewed as coercive.

The intervention curriculum was presented to the

supervisory committee, and to several outside experts for

review. The committee was composed of four health educators

and one experimental behavioral psychologist. The outside

reviewers included:

Carlo DiClemente, Ph D

Behavioral Researcher

University of Houston

Thomas J. Glynn, Ph D

Chief, Cancer Prevention and Control

Extramural Research Branch

National Institutes of Health

National Cancer Institute

Terry L. Conway, Ph D

Health Psychology Department

Naval Health Research Center

All reviewers stated they believed the program to be a sound

design. Many suggestions were given, and most were

incorporated into the existing plan.

Sham treatment

A sham (placebo) treatment was conducted for one company

to analyze the effect of any intervention on the study

population. The idea was to present an educational

intervention, irrelevant to smoking prevention/cessation, but

prefaced by an explanation that the class is part of an

overall program designed to help recruits remain tobacco-free

after graduation from RTC.

This intervention was accomplished on the company's 2.2

day of training, but only for a 40 minute period. The

curriculum (see Appendix G) consisted of a short lecture

about the Navy's commitment to maintain a healthy, productive

workforce, Navy smoking policy, and goals for tobacco use.

Tobacco use was tied in to weight control by the common theme

of heart disease. A 24 minute Navy video entitled "Weight

and fat control/nutrition education" was then shown.

A question and answer period followed the movie. The

class was concluded by a statement about the Navy's

commitment to a healthy workforce and the fact that recruits

should try to make healthy choices in all aspects of their



Two companies were given a booster class at the end of

recruit training to reinforce key points of the education

intervention, and to further motivate them not to use tobacco

after graduation. This class was presented on the Friday

before the recruits' departure from RTC. Their formal

graduation ceremony (Pass and Review) was held Friday morning

and the recruits were allowed to go on uncontrolled off-base

liberty starting that evening. The booster class lasted

approximately 45 minutes and was presented mid-afternoon to

both companies in this treatment group.

The booster class incorporated several strategies to

reinforce prior training and increase motivation. Two

pamphlets were distributed to each recruit:

*Smart Move: A Stop Smoking Guide ACS #2515LE

*Clearing the Air: How to Quit Smoking and Quit

for Keeps National Cancer Institute NIH

Publication No. 89-1647.

These pamphlets are comprehensive guides to quitting smoking.

They cover much of the information which was given in class

and contain many suggestions to help avoid relapse. A wallet

card (see Appendix H) was also given to the recruits. This

card lists 5 reasons to remain tobacco-free and 5 ways to

avoid smoking.

The first part of the class period involved going

through the materials distributed. Those who had never

smoked were asked to pass the materials on to a friend or

relative who does smoke, or save it for a smoking shipmate

they may meet later. The ex-smokers were asked to keep the

material for future reference and to place the card in their

purse of wallet for reference as necessary after graduation.

A nineteen minute video (Hazards of Tobacco) was then

shown. This video was made by Captain David Moyer, a Navy

physician who has taught the material in the video to over

50,000 Navy and Marine Corps recruits at training centers in

San Diego, California. This video also covered many of the

same points the education intervention covered, as well as

some different reasons for the recruit to not start/resume

tobacco use.

The recruits were then asked to sign a pledge to assist

the Navy in reaching its goal of being smoke-free by the year

2000. Each recruit received a certificate of participation

in the tobacco prevention program (see Appendix I), but

signing the pledge was voluntary. It was stressed that

signing was not necessarily a promise never to use tobacco,

but rather a promise to do whatever they can to help reach

the goal. The recruits were told the Navy hoped part of this

assistance would be a choice to remain tobacco-free. The

certificate was signed by the researcher, as the program

director, and the Commanding Officer of Recruit Training

Center, Orlando.

All recruits in the booster companies received a T-shirt

courtesy of the American Cancer Society. The T-shirts were

yellow and carried the logo of "Smoke-free class of 2000".

These shirts are being used in the national "Smoke-free class

of 2000" campaign. The recruits were told they are part of

the Navy's smoke-free class of 2000. The recruits were

thanked for their participation, congratulated on graduation,

and reminded of the upcoming follow-up surveys.

Follow-up surveys

Follow-up surveys were administered just before the

participants left RTC, then again 3 months after graduation.

These surveys were designed to measure current tobacco use

habits, knowledge, and attitudes.

The graduation survey (see Appendix B) was administered

by the company commanders after the uncontrolled liberty

weekend. Because some of the questions dealt with possible

rules infractions, the company commanders were asked to leave

the room while the surveys were completed. The company

yeoman collected the completed surveys, sealed them in an

envelope, and returned them to the RTC scheduling office

where the researcher collected them. The female control

company had their graduation surveys mailed to them 3 weeks

after graduation due to confusion in who was to administer.

The 3 month follow-up surveys (see Appendix C) were

mailed to participants at their current command. Those

attending schools located at Naval Training Center, Orlando

were administered the survey during one of their class

periods by a command representative. These surveys were

collected by the command Drug and Alcohol Program Advisor and

mailed to the researcher. The majority of the surveys were

mailed to the participants' Commanding Officer. The

Commanding Officers' representatives were instructed to

distribute the surveys to listed participants, collect the

completed forms, and mail them back to the researcher. One

reminder was sent to the commands which had not responded

within one month of the original mailing. Several commands

not responding to the reminder letter were phoned and

questioned concerning survey status. A number of surveys

(31) had to be re-sent to individuals that had already moved

from their second command. Some surveys (in addition to

those readdressed) were mailed, to individuals (80), with a

self-addressed government envelope and instructions for

return. Participants were not required to pay any postage.

Data Preparation

Once the surveys were completed the answers were

transferred to a 240 item general purpose answer sheet

formatted for optical mark reading. The number of questions

asked on the three surveys totalled 279. Many of the answers

were eliminated or combined so that only 168 data points were

actually coded. The answer sheets had only five response

circles per answer. Some survey questions that had more than

five responses had to be collapsed to fit this requirement.

This was done consistently for questions which were repeated

in surveys. All answers were coded by numbers one through

five (eg. No=1; Yes=2) for data entry, and a survey layout

(codebook) was composed to avoid confusion during analysis.

This layout was used during the coding process to ensure

answers were coded consistently and transferred to the

appropriate circles on the answer sheets. This process was

completed by two individuals who were versed on the

procedure. This process was chosen for a number of reasons


-multiple surveys were involved

-illogical answering could be identified before entry

-the answer sheets could be re-scanned in the event of

data file loss on the computer

-computer entry errors, due to the large data set, could

be minimized.

The completed answer sheets were scanned by a mark reader

which then compiled the data and stored a data file on a

floppy,disk and in a university computer account.

Once the data were in the computer file variables were

range checked for accuracy and improper codes were either

corrected or changed to missing data points. No attempt was

made to impute missing data points.

Data analyses were accomplished using:

-the Statistical Analysis System (SAS) on the University

of Florida VAX cluster

-SAS on the Northeast Regional Data Center computer

Virtual Telecommunications Access Method, and

-Stat View SE + Graphics statistical analysis program on

a Macintosh SE-personal computer.

Analysis Plan

This study was set up on a split plot with repeated

measures design. Table 5 is a presentation of the design

setup. Because tests of the null hypotheses consisted of

both continuous and categorical variables, different testing

procedures were used. Smoking attitudes and smoking

knowledge were examined using a general linear model (GLM)

for repeated measures analysis of variance (ANOVA). Null

hypotheses with categorical variables (relapse rate, smoking

initiation, cigarette consumption, and smoking intentions)

were examined by using frequency tables with the Pearson's

chi-square statistic. This test statistic looks at the

goodness of fit between the distribution observed and the

distribution expected (Aday, 1989).

Table 5. Repeated measures design

Pre-test Post-test 3 Month f/u

0 T 0 O0

O T2 O 0

O T3 O O

O T4 O O

Notes: T1


+ education
+ education + booster
+ sham treatment

Smoking relapse and smoking initiation were examined

according to Table 6, using differences between pretest and






Table 6.

Chi-scuare freuuencv table

Notes:D1=current smoking at pretest ana at follow-up (RELAPSE)
D2=current smoking at pretest but not at follow-up (NON-RELAPSE)
D3=no smoking at pretest or follow-up (NON-INITIATION)
D4=no smoking pretest but current smoking follow-up (INITIATION)

D1 D2 D3 D4

Smoking intentions were examined using differences between

pretest and posttest, observe effect of policy and education.

Cigarette consumption was examined using differences between

pretest and follow-up.

A number of univariate analyses procedures were used to

help describe population variables and draw conclusions about

possible future benefits in research designs with this

population. The significance level for all the tests was set

at p=.10, as the consequences of a Type I error did not

outweigh acceptance of a high probability of making a Type II

error. As this was the first study of its kind with Navy

recruits, the exploratory nature of the educational design

was considered in determining the alpha level.



The purpose of this chapter is to present the results of

data analysis and an explanation of how these results support

or do not support the hypotheses. The data presentation and

discussion are organized around individual hypotheses.

The chapter begins with a precise description of the

study population as it evolved through the study period.

Next, the actual results as they apply to each hypothesis are

presented. The last section is an interpretation of the

findings, a discussion of the implications of additional

knowledge in the smoking field, and a discussion of potential

applications of the findings.

Population Description

The study began with a total of 557 subjects in 7

recruit companies. Fourteen percent of these subjects were

lost due to normal attrition from their original recruit

company. This attrition was equally distributed across

subject companies. Therefore, the total number of subjects

who completed a pretest and graduated with their original

company was 496 (only two subjects refused to complete the

pretest survey). Seventy-four percent of this total

completed surveys at graduation (only 50% were completed in

treatment Group 1 due to an administering error in one


Follow-up surveys were mailed to 516 individuals. Of

these surveys 320 were returned completed for a response rate

of 62%. Table 7 summarizes survey completion at each


Table 7. Percent survey completion at each measurement

Treatment group

1 2 3 4 Total

Pre (140) (151) (136) (69) (496)

Post 50% (70) 85% (128) 82% (112) 81% (56) 74% (366)

F/U 57% (80) 64% (97) 60% (82) 29% (20) 56% (279)
*Notes:Actual numbers are in parentheses.
Treatment: 1 = Policy only
2 = Policy + education
3 = Policy + education + booster
4 = Policy + sham treatment

Treatment Group 4 consisted of only one recruit company and a

poor follow-up return rate of 20 should be noted.

Ten percent of the surveys were reported lost or

undeliverable to participants, and 28% were not returned (not

accounted for) at all. Forty-two of the surveys not returned

had been sent to 13 commands who did not respond to the

initial letter or the reminder, and could not be reached by

phone. The rest of the non-returned surveys had been mailed

directly to individuals. Only a 22% response rate was

obtained when participants' Commanding Officers were not

involved in administering the follow-ups. The increase in

overall follow-up response shows the utility of mailing

surveys via Commanding Officers.

Of the completed surveys at follow-up, 279 were usable

for the main study population. This represented 56% of the

original pretested population The rest were unidentifiable

as study participants. An analysis was performed to

determine if the group of non-respondents was representative

of the whole population with respect to initial smoking

status; Table 8 summarizes the findings.

Table 8. Percentage of smokers in non-respondent group and
original study population

Treatment group

1 2 3 4 Total
Non- 51.7% 46.3% 53.7% 42.9% 48.8%
Respondents (31) (25) (29) (21) (106)
At pretest 50.7% 35.8% 46.3% 39.1% 43.3%
(71) (54) (63) (27) (215)
*Note: Actual numbers in parentheses.

Non-respondents of each treatment group had a higher

percentage of smokers, but the differences from the pretest
results were not significant (X2 = 4.65, p>.l).


Smoking relapse

Participants were asked to respond to questions

regarding current smoking status, as part of each of the

three surveys. Current smoking status for the pretest

reflected status just before entering recruit training. A

summary of current smoking by treatment group, at each

measurement period, is presented in Table 9. This table also

shows the smoking relapse rate for each treatment group.

Table 9. Percent of current smokers at each measurement

Treatment group
1 2 3 4 Total
n=161 n=171 n=166 n=69 N=496
Pre 50.7% 35.8% 46.3% 39.1% 43.3%
(71) (54) (63) (27) (215)
Post 27.1% 3.8% 2.6% 18.9% 10.2%
(19) (5) (3) (11) (38)
F/U 42.5% 25.5% 33.7% 25.0% 32.7%
(34) (25) (28) (5) (92)
Rate 82.5% 66.6% 76.5% 66.6% 75.5%
*Note: Actual numbers in parentheses

While relapse rates (at follow-up) are lower among

participants in the treatment groups, results of chi-square

statistics calculated for these differences showed they were
not significantly different than the control (X2 = 1.68,

p>.l). Since the RTC smoking policy still applied to

participants at time of posttest, indications of current

smoking were considered violations of rules (smoking on

liberty) rather than relapse. The education treatment groups

(Groups 2 & 3) showed lower violation of smoking rules and

results of chi-square analysis showed these differences to be

significant (X2 = 22.08, p<.001).

Several variables were examined that may have confounded

the relapse results: participant gender, AFQT scores,

Fagerstr6m nicotine tolerance scores, age started smoking,

and cigarette consumption prior to entry. Covariance of the

continuous variables (AFQT and Fagerstr6m scores) was

initially examined by a general linear means procedure (SAS

PROC GLM) to determine if further analysis was warranted.

Categorical modeling procedures were then run

controlling for the variables of possible significance.

Table 10 summarizes the smoking status at entry and relapse

rate by gender, for the study population.

Table 10. Percent of current smokers by gender and
corresponding relapse rates

Male Female
Current smokers at 40.9% 49.6%

Relapse rate 72.6% 79.2%

The differences seen in relapse rate by gender were found not
to be significant (X2 = .634, p>.l).

The Fagerstr6m scores were transformed to categorical

form by creating two levels:

low tolerance (Fagerstr6m scores 1 6)

corresponding to low nicotine addiction

high tolerance (Fagerstr6m scores 7 11)

corresponding to high nicotine addiction.

Table 11 summarizes nicotine tolerance level reported at

entry and corresponding relapse rates.

Table 11. Nicotine tolerance levels reported with
corresponding relapse rates

Nicotine Tolerance Level

Low High

Reported at entry 74.7% 25.3%

Relapse Rate 76.1% 96.8%

The data seen in Table 11 seems to indicate that smokers

reporting a high nicotine tolerance relapsed to smoking at a

higher rate. The differences seen in relapse rate by
nicotine tolerance level were found to be significant (X2 =

3.58, p<.l).

Another category that seems to affect smoking relapse is

the number of cigarettes smoked prior to entry in recruit

training. Figure 1 summarizes the relapse rates for given

categories on smoking at entry.




i 6 o0 Relapse
Q rate


02----0 no-- -
1 2 3 4 5
Cigarettes smoked
*Note: 1 = no smoking
2 = <1 per day
3 = 1-5 per day
4 = 6-20 per day
5 = >20 per day

Figure 1. Percent relapse for corresponding entry smoking

Some relapses are found in cigarette consumption

category "none" because those indicating they quit within two

months prior to recruit entry were considered current

smokers. An obvious trend is seen in that relapse increases

with an increase in cigarette consumption. The differences

seen in relapse rate by prior cigarette consumption were

found to be significant (X2 = 35.85, p<.001). Analysis of

age smokers started and their AFQT scores did not reveal any

significant results.

To determine if relapse was affected by increased

knowledge scores an analysis of variance (ANOVA) was

conducted modeling relapse with knowledge gain. Possible

knowledge score range was zero to six. The mean knowledge

gain for non-relapsers was 0.95 (16%) and for relapsers was

0.34 (6%). This difference was significant (F = 5.09, df =

1, 77, p<.l).

In an effort to understand why recruits start smoking

again after boot camp, participants were asked to list

reasons they began smoking again. The top five reasons

relapsers listed for starting smoking after boot camp were:

"It calmed me" 81%

"I liked the taste" 56%

"Parents or friends back home smoke" 53%

"Friends smoked" 49%

"I had no desire to quit" 44%.

Smoking initiation

The other side of the smoking status analysis concerned

non/former smokers who began smoking during the three month

follow-up period. Table 12 is a summary of smoking

initiation by treatment group.

Table 12. Percentage smoking initiation

Treatment group

1 2 3 4 Total
Smoking 2.5% 7.3% 4.1% 7.1% 5.3%
started (1) (5) (2) (1) (9)
*Note: Actual numbers in parentheses

A chi-square procedure was used to analyze differences

between groups. The differences observed were found to be
not significant (X2 = 1.44, p>.l).

Smoking attitudes

The attitude score was obtained by summing adjusted

values for the following variables: concern about cigarettes,

concern about pipe smoking, concern about chewing tobacco,

concern about snuff use, the importance of being a non-

smoker, feeling about working in a smoke-free environment,

and attitude toward targeted cigarette advertising. The

range of possible scores was zero to ten (ten being the most

positive attitude toward not smoking).

Two sets of analyses were conducted for attitude

changes. Smoking attitude scores for all participants were

examined, then scores were examined for those participants

who were current smokers at entry to the Navy. Table 13

presents the attitude score means and standard deviations

observed (for all participants) at each measurement period by

treatment group.

Table 13. Attitude scores at each measurement period and
differences pre to post for all subjects

Treatment group

1 2 3 4 Total
6.31 6.37 6.33 6.23 6.32
Pre (2.88) (2.80) (2.99) (2.99) (2.89)
6.46 7.41 7.36 6.87 7.12
Post (2.93) (2.58) (2.65) (2.73) (2.71)
6.66 6.49 7.05 6.65 6.71
F/U (3.09) (2.94) (2.77) (2.89) (2.92)
Diff.** 0.25 1.05 0.91 0.89 0.83
post-pre (2.33) (2.36) (2.69) (2.87) (2.55)
*Note: -Standard deviations in parentheses
**Differences reflect only subjects completing both pre and post

Repeated measures ANOVA were performed to determine if

time, treatment group, or interaction effects were present.

A significant time effect was observed for all groups (F =

4.38, df = 6, 394, p<.l). Treatment effect differences were

not significant (F = 0.5, df = 3, 197, p>.1), but interaction

effects were significant (F = 1.89, df = 6, 394, p<.l). This

indicates the changes seen through time on the smoking

attitude scores of all subjects were different depending on

which treatment group the subject was in. Separate critical

F values were then calculated for paired attitude scores

based on the significant interaction analysis. Pretest

scores were not significantly different for any two treatment

groups, which was also the case for follow-up scores. For

this reason, only attitude differences pre to post are

reported. The only significant treatment differences seen

occurred on the posttest and are summarized below.

TRT 1 TRT 2 F = 13.36, p < .10

Posttest TRT 1 TRT 3 F = 11.37, p < .10

scores TRT 4 TRT 2 F = 3.96, p < .10

TRT 4 TRT 3 F = 2.92, p < .10

Table 14 presents the attitude score means and standard

deviations observed (for smokers at entry) at each

measurement period by treatment group.

Table 14. Attitude scores at each measurement period and
differences pre to post for smokers

Treatment group

1 2 3 4 Total
4.63 4.41 4.17 4.33 4.40
Pre (2.52) (2.44) (2.51) (3.10) (2.57)
4.95 5.89 5.84 5.75 5.63
Post (2.72) (2.47) (2.77) (2.64) (2.66)
4.77 4.23 5.02 6.0 4.77
F/U (2.83) (2.42) (2.47) (2.90) (2.61)
Diff.** 0.18 1.51 1.50 1.50 1.19
post-pre (2.36) (2.47) (3.26) (2.527) (2.76)
*Note: Standard deviations in parentheses
**Differences reflect only subjects completing both pre and post

Repeated measures ANOVA were performed to determine if

time, treatment group, or interaction effects were present.

A significant time effect was observed for all groups (F =

5.74, df = 6, 148, p<.l). Treatment effect differences were

not significant (F = 0.61, df = 3, 74, p>.l). Interaction

effects were also not significant (F = 1.38, df = 6, 148,

p>.l). Separate critical F values were calculated for paired

attitude scores. Pretest scores were not significantly

different for any two treatment groups. The only significant

differences seen for posttest and follow-up scores are

summarized below:

TRT 1 TRT 2 F = 5.59, p < .10

Posttest TRT 1 TRT 3 F = 5.13, p < .10

scores TRT 1 TRT 4 F = 6.19, p < .10

Follow-up TRT 2 TRT 3 F = 3.07, p < .10

scores TRT 2 TRT 4 F = 4.57, p < .10

Smoking Knowledge

Knowledge on certain aspects of tobacco use were

examined for all treatment groups. The knowledge score was

obtained by summing adjusted values for the following

variables: smoking effect on heart disease, smoking effect

on cancer, smoking effect on emphysema, smoking

addictiveness, smoking prevalence in U.S. population, and

smoking prevalence in Navy population. The possible range

for scores was zero to six. A score of six indicated a high

level of smoking knowledge and a score of zero indicated a

low knowledge.

Table 15 presents the mean scores and standard

deviations observed at each occasion by treatment group along

with the differences observed between occasions.

Table 15. Knowledge scores at each measurement period and
differences observed

Treatment group

1 2 3 4 Total
4.01 4.06 4.07 3.93 4.03
Pre (1.08) (1.17) (1.02) (1.13) (1.10)
4.29 4.59 4.57 4.25 4.48
Post (0.83) (0.85) (0.85) (1.00) (0.89)
4.14 4.39 4.56 4.05 4.35
F/U (1.02) (1.00) (0.93) (0.39) (0.98)
Diff.** 0.19 0.54 0.45 0.29 0.41
post-pre (1.21) (1.09) (1.11) (1.12) (1.13)
Diff.*** -0.22 -0.25 -0.07 -0.35 -0.19
f/u-post (1.24) (1.04) (1.10) (0.70) (1.07)
*Note: Standard deviations are in parentheses.
**Differences reflect only subjects completing both pre and post
***Differences reflect only subjects completing both pre and

Repeated measures ANOVA were performed to determine if

time, treatment group, or interaction effects were present.

A significant time effect was seen for the entire population

(F = 8.46, df = 6, 398, p<.l). Treatment effect was also

significant (F = 2.27, df = 3, 199, p<.l) but interaction

effect was not significant (F = 1.31, df = 6, 398, p>.l). A

knowledge difference variable (knowledge gain) was produced

by subtracting the pretest knowledge score from the posttest

knowledge score. An ANOVA (SAS PROC GLM) was conducted

modeling knowledge gain with treatment level. Post-hoc tests

revealed a significant difference only between treatment

Groups 1 and 2. A knowledge difference variable (knowledge

retention) was produced by subtracting pretest scores from

follow-up. Results of knowledge retention show the booster