DIFFERENCES AMONG SMOKERS, NONSMOKERS AND FORMER SMOKERS
AN INVESTIGATION OF HEALTH CARE WORKERS
by
Nicholas V. Cascarelli
Submitted in Partial Fulfillment ofthe Requirements
For the Degree of
Master in Health and Human Services
In the
Bitonte College ofHealth and Human Services
THE YOUNGSTOWN STATE UNIVERSITY
June 2000
DIFFERNCES AMONG SMOKERS, NONSMOKERS AND FORMER SMOKERS
AN INVESTIGATION OF HEALTHCARE WORKERS
Nicholas V. Cascarelli
I hereby release the thesis to the public. I understand this thesis will be housed at the
Circulation Desk ofthe University library and will be available for public access. I also
authorized the University or other individuals to make copies ofthis thesis as needed for
scholarly research.
Signature:
Nicholas V. Cascarelli, Stu ent
Approvals: "11 J,. J_
~ ~=f#'tJUlfA'
Dr. Carol Mikanowicz, Thesis Advisor
r. Joseph Waldron, Committee Member
.t.
Date
Date
Abstract
The study consisted oftwo hundred and six employees from two hospitals. The
38-item questionnaire asked respondents to answer questions regarding behaviors,
attitudes, and knowledge concerning smoking cigarettes in general and workplace
smoking policies. The percentage ofthis study that reported they were smokers was
15.5%. A linear regression analysis was performed to assess the predictability ofbeliefs
about smoking in general and workplace smoking policies as a function ofactual
smoking behavior. Both regression analyses displayed moderate strength r2 = .29 on
general beliefs and knowledge about smoking and r2 = .18 on workplace smoking
policies. ANOVAs were conducted to evaluate differences among smokers, never
smokers and former smokers in terms ofknowledge and beliefs about smoking in general
(p< .01) and in attitudes about workplace smoking policies (p= .019). Cigarette smokers
were more likely to have beliefs that would favor smoking behavior than nonsmokers or
former smokers. They also tended to have less knowledge ofthe effects ofsmoking. In
terms ofworkplace smoking policy, there were differences between smokers and
nonsmokers, however the differences were not as great as they were in terms ofgeneral
knowledge and beliefs.
III
Table ofContents
Abstract iii
List ofTables vi
List ofFigures vii
CHAPTER I INTRODUCTION 1
Purpose ofthe Study 1
Hypotheses ofthe Study 1
Delimitations (parameters ofstudy) 2
Assumptions ofthe study 2
Operational Definitions 3
Limitations ofthe Study 4
Summary 4
CHAPTER II LITERATURE REVIEW 6
Smoking Prevalence 6
The Effects ofSmoking 8
Background and Origins ofThe Health BeliefModel 9
The Health BeliefModel 10
The Relationship between Cigarette Smoking and the Health BeliefModel 11
The Problem ofETS (Environmental Tobacco Smoke) 15
The Relationship between Smoking and the Workplace 15
Summary 19
CHAPTER III METHODS 20
Research Design 20
Subjects 21
Instrument 21
Collection ofData 22
Statistical Treatment ofData 22
Summary 23
CHAPTER IV RESULTS OF THE ANALYSIS ,25
Profile ofthe subjects 25
Results for hypothesis 1 ,34
Results for hypothesis 2 ,36
CHAPTER V SUMMARY, CONCLUSIONS AND RECOMMENDAnONS 42
Summary ,42
Conclusions '44
Recommendations 45
iv
BIBLIOGRAPHy 47
Appendix A: Smoking in the Workplace Questionnaire 50
Appendix B: Human Subjects Committee Approval 54
Appendix C: Letter ofPermission from Employer. 56
Appendix D: Letter ofConsent 58
v
List ofTables
Table 1 Number and Percentage ofHealthcare Workers'
Smoking Behavior by Age 26
Table 2 Number and Percentage ofHealthcare Workers'
Smoking Behavior by Gender 27
Table 3 Number and Percentage ofHealthcare Workers'
Smoking Behavior by Education 29
Table 4 Number and Percentage ofHealthcare Workers'
Smoking Behavior by Race 31
Table 5 Reliability Analysis ofTotal Smoking Inventory Items 32
Table 6 ANOVA Hypothesis 1 36
Table 7 Post hoc Comparison Tukey's HSD 37
Table 8 ANOVA Hypothesis 2 40
Table 9 Post hoc Comparison Dunnett's C 41
vi
List ofFigures
Figure I The Relationship Between Smoking and the
Total Smoking Inventory Scale 35
Figure 2 The Relationship Between Smoking and the
Workplace Permissiveness Scale 39
vii
CHAPTER I
INTRODUCTION
Cigarette smoking is a major health problem for our nation. It is the leading
preventable cause ofdisease and death in the United States. Smoking prevalence is
significantly above the 15 percent that Healthy People 2000 objectives had set forth.
In the United States, 48 million adults smoke which is approximately 25% ofthe adult
population (American Lung Association, 1999).
Smoking occurs in both genders with 25 million men and 23 million women smoking
on a daily basis. Education is a good predictor ofsmoking behavior demonstrating that
smoking rates are highest among those individuals who have 9-11 years ofeducation.
Poverty is a second predictor as people who are at or below the poverty level have higher
smoking rates. Smoking causes both health and economic consequences as 90% ofall
lung cancer can be attributed to smoking and $2 for every pack ofcigarettes bought goes
to the medical expenditures associated with that pack.
Purpose ofthe Study
The purpose ofthis study was to measure smoking behavior among a working
population in a healthcare setting in Mahoning and Trumbull Counties in Northeast Ohio.
Those aspects would include knowledge, attitudes, and behaviors associated with
cigarette smoking in general and in the workplace. The relationship ofdemographic
factors such as age, gender, race, and level ofeducation will also be examined to look for
trends as they relate differences in beliefs, behaviors or knowledge.
Hypotheses ofthe Study
Hypothesis 1. There are significant differences among smokers,
nonsmokers, and former smokers in terms ofhealth beliefs about smoking.
Hypothesis 2. There are significant differences between smokers,
nonsmokers, and former smokers in terms ofattitudes toward workplace
smoking policies
Delimitations (parameters ofstudy)
1. Geographical - The participants work in either Mahoning or Trumbull
Counties ofNortheastem Ohio.
2. Institutional type - Hospital settings
3. Age - All respondents are 18 and over.
4. Gender - Most ofthe respondents were female. (79.1%)
5. Race - Ethnic minorities were underrepresented. (10.7%)
6. Education - Most ofthe respondents had a bachelor's degree or
higher. (53.4%)
2
Assumptions ofthe study
1. Smoking is perceived by the general public as not a healthy behavior
choice.
2. Smoking is an addiction.
3. The participants filling out the questionnaire understand the definition
ofsmoker.
4. The questionnaire will reflect the knowledge, attitudes, and behaviors
ofthe respondents with a reasonable amount ofaccuracy.
5. Participants will be honest in answering survey questions.
Operational Definitions.
The following terms are defined as used in this study.
Smoker - A person who has smoked within the last six months.
Former Smoker - A person who no longer smokes but used to smoke more than
six months ago.
Nonsmoker - A person who has never smoked.
Psychological addiction - Addiction to cigarettes that can be attributed, at least
partially, to psychological and/or social coping mechanisms ofthe individual.
Physiological Addiction - Addiction to cigarettes that can be attributed, at least
partially, to a physical need for nicotine by the body to maintain homeostasis.
Total Ban - This is a workplace smoking policy whereby smoking is not permitted
at all.
Partial Ban - This is a workplace smoking policy where smoking is permitted in
designated areas.
3
Limitations ofthe Study
1. The sample was one ofconvenience. Only the hospitals that gave
permission took part in the study. Members ofminority races, men,
elderly, and unemployed are underrepresented or not represented in the
study. Thus the generalizability ofthe results is limited.
2. This study represents the first time this instrument was used. Although
reliability tests were run on the majority ofthe questions, only replication
ofthe study would solidify reliability.
3. The opinions on the smoking policy questions may be skewed by
workplace policies already in existence.
4. The definitions ofsmoker, former smoker, or nonsmoker were not
defined in the questionnaire.
Summary
Smoking cigarettes is a health behavior that causes and contributes to many diseases
not only to the smoker, but also to the surrounding people in the same environment. In
addition, a person who smokes also has to cope with the addiction to nicotine and the
negative stigma that is attached to a person who smokes. The workplace is often the
place that addresses the problem ofenvironmental tobacco smoke by the implementation
ofa smoking policy.
4
Chapter II is a review ofthe literature that discusses the theories used as the foundation
for the study. It also presents some ofthe prevailing facts about smoking behavior in
general and presents the existing literature that examines workplace smoking policies.
Chapter III presents the methods and procedures used in the study and sets guidelines
for data analysis.
Chapter IV presents the results ofthe data analysis. There is also a discussion ofthe
results.
Chapter V is a summary ofthe findings, discussions, limitations and recommendations
for further research.
5
CHAPTER II
LITERATURE REVIEW
The review ofthe literature consists ofgeneral information about smoking prevalence,
the effects ofsmoking, background ofthe Health BeliefModel, the constructs ofthe
Health BeliefModel, the relationship ofcigarette smoking and the Health BeliefModel,
the problem ofEnvironmental Tobacco Smoke, and the relationship ofsmoking and the
workplace.
Smoking Prevalence
Twenty-five million men (27.6%) and 23 million women (22.1 %) Americans smoke
on a daily basis. However, by the end of2000, it is predicted that these figures will be
reversed and more women will smoke than men. Another theme in the literature is that
education is a good predictor ofsmoking behavior. People who have 9-11 years of
education have the highest smoking rates (35.4 %) whereas smoking rates were lowest
among adults who have 16 or more years ofeducation (11.6 %). Smoking rates are
higher among individuals who live below the poverty level (33.3 %) than those who live
at or above the poverty level (24.6 %). From a historical perspective, smoker's aged 25 
44 usually had the highest smoking prevalence ofany age group. However, smokers age
18-24 are now the age group with highest prevalence at 28.7 %. This could be explained
by previously high rates ofunderage smokers keeping the habit into adulthood. Figures
for the other age subdivisions include: 25 to 44 years ofage, 28.6 %; 45 to 64 years, 25.5
%; 65 years and older, 13.0 %. (CDC,1998)
The distribution geographically ofsmoking prevalence in the United States shows a
great deal ofvariation from state to state. Kentucky and West Virginia have the highest
6
smoking rates, over 30 % for men and women. Utah and California are the only states
below 20 %. They are 17.1 % and 19.5 % respectively. In general, adults in the southern
region ofthe United States have higher rates ofsmoking than adults in the western part of
the country, which have the lowest rates (Shopland, Hartmann, Gibson, Mueller, Kessler,
& Lynn, 1996). The authors qualified their findings by noting that some ofwhat they
found could be attributed, at least partially, to the differences in the smoking behavior
between men and women and among diverse racial and ethnic populations in various
areas ofthe country.
The majority ofsmokers are ofCaucasian non-Hispanic ancestry, but the rates ofthis
group are decreasing faster than minority groups. There is no single factor that
determines patterns ofcigarette smoking among racial/ethnic minority groups. The
patterns ofuse are the result ofcomplex interactions ofmultiple factors, such as
socioeconomic status, cultural characteristics, stress, biological aspects, varying
differences ofcommunities to mount effective anti-smoking campaigns. Smoking rates
are generally declining for most groups. However, African-American and Hispanic
American adolescents have shown an increase in rates since the early 1990's.
Consequently, cigarette manufacturers have increasingly targeted certain races such as
African-Americans or Hispanic Americans. The modes used to implement this strategy
were through advertisement in magazines, garnering community loyalty by hiring
community members, providing communities with revenues, and sponsoring cultural,
athletic, entertainment events targeted at these groups (DHHS, 1998). Previously,
billboards were strategically placed in minority communities, but recent legislation
prohibits this visibility.
7
Effects ofSmoking
The Centers for Disease Control and Prevention estimated over each year as a result of
smoking cigarettes. (American Lung Association, 1999) They also reported that
smoking-related diseases cost the United States at least 97 billion dollars per year - 50
billion is direct healthcare costs (hospital care, physicians, and medications) and the
remainder is money in lost productivity. Also in this report, nearly halfofthe 50 billion
in direct healthcare costs are being paid by public funds. For every pack ofcigarettes sold
in the U.S., $2 is needed for medical expenditures associated with the purchase ofthat
pack.
The Surgeon General's Report in 1989 identified at least 43 carcinogens. The report
also indicated that approximately 90 % oflung cancer cases are attributed to smoking.
(DHHS, 1990) Smoking also can be an important factor in the development ofother
types ofcancers such as cancer ofthe esophagus, gastrointestinal tract, urinary tract, and
cervix. Smoking causes most cases ofemphysema and chronic bronchitis. Smoking
contributes to diseases ofthe circulatory system including arteriosclerosis, coronary heart
disease, and stroke. In addition to these major life threatening illnesses, smoking is at
least partially responsible for many other conditions and disorders such as infertility,
slow wound healing, impotence, peptic ulcer, ectopic pregnancy, and bone density
deficiencies in women (Hahn & Payne, 1998)
8
Background and Origins ofThe Health BeliefModel
The Health BeliefModel (HBM) was truly one ofthe first theories developed
specifically to explain health behavior. It's efforts grew out ofa need to explain people's
unwillingness to become involved with programs that would help to detect or prevent
disease. The Health BeliefModel's origins can be traced back to the 1950's as a
synthesis ofexisting theories from social psychology. The two theories used to
formulate the Health BeliefModel were Stimulus-Response (S-R) Theory and Cognitive
Theory.(Glanz, Lewis & Rimer, 1997)
Stimulus-Response Theory is based on the constructs ofreinforcements and rewards to
explain the behavior ofindividuals. Here learning is a result ofeither a reduction in
behavioral drives due to the intervention ofreinforcements or an increase in certain
behaviors due to the presence ofa reward ifthe proper behavior has taken place. One of
the problems with S-R theory is that it does not address the ability ofhumans to reason.
The Cognitive Theory, however, does address the mental processes ofhumans in terms of
their decision-making. This theory is based on the notion that behavior depends on the
individual's perception ofprobability that a particular action will achieve a particular
outcome.
The Health BeliefModel borrows more heavily on Cognitive theory. This is not to say
that S-R theory is unimportant. S-R Theory lays the groundwork for the Health Belief
Model in that reinforcements and rewards serve to influence expectations or hypotheses.
The Cognitive theory influences behavior change more directly.
9
The Health BeliefModel
The Health BeliefModel was used in this research as a theoretical base. This model
poses many constructs and their relationships to one another in determining the likelihood
to bring about a desired behavioral change.
The first component ofthe Health BeliefModel involves those perceptions of
individuals. The two major constructs within this component are perceived susceptibility
and perceived severity. Perceived susceptibility is simply one's opinion ofchances of
becoming ill. The perceived severity construct is the individual's perception ofthe
seriousness ofan illness. (Strecher & Rosenstock, 1997)
The second component ofthe Health BeliefModel involves those factors that modify
the ability to make the desired behavior change. Key factors within this component
include both ascribed and achieved characteristics ofthe individual. In turn these
characteristics lead to cues to action. Cues to action are another construct ofthe Health
BeliefModel that are either internal or external triggers ofindividual action. (Strecher &
Rosenstock, 1997)
The final component ofthe Health BeliefModel is the likelihood ofaction. This
component is an assimilation ofthe previous two components and their interaction with a
person's perception ofbenefits and barriers to behavior change. A construct that
pervades this component ofthe Health BeliefModel is self-efficacy. It is a measure ofan
individual's ability to take action. In essence, ifthe person perceives by making a
behavior change that the benefits are greater than the barriers to making the change, the
greater degree ofself-efficacy that person will have and the likelihood ofbehavior
change is increased. (Bandura, 1977)
IO
The Relationship between Cigarette Smoking and the Health BeliefModel
The literature on smoking in terms ofthe Health BeliefModel usually focuses on one
component ofthe Health BeliefModel. (Strecher & Rosenstock, 1997) It is rare to find
studies that encompass all the various aspects ofthe HBM.
Perhaps the reasoning for many studies not using all aspects ofthe HBM relates to the
concepts ofperceived susceptibility and perceived severity. It has been assumed that the
differences between smokers' and nonsmokers' attitudes about threat are not statistically
significant. In a 1992 study by Brownson, 83% ofcurrent smokers and 91 % of
nonsmokers felt smoking was harmful. This study also found that both groups were
similar on benefits ofquitting.
Another paradigm in looking at beliefs ofsmoking is to examine differences in beliefs
and benefits ofsmokers and former smokers. One study found that fewer smokers
(27.9%) than former smokers (42.1 %) accepted that smoking causes disease (Chapman,
Wong, & Smith, 1993). The diseases the authors ofthis study referred to were heart
disease, poor circulation, bronchitis, lung cancer, and stroke.
The ability to get past perceived barriers to quitting is what determines the likelihood of
behavioral change. The difficulty with quitting smoking is that there is a significant
presence ofboth psychological and physiological barriers that interfere with smoker's
success.
Congress and the media have recently given special attention to the addictive properties
ofcigarettes. From a physiological standpoint, the administration ofnicotine for a few
11
weeks can bring about changes in the central nervous system and endocrine system that
are indicative ofphysiological dependence (Henningfield & Keenan, 1993). Another
aspect that is commonly associated with this type ofdependence is a withdrawal response
ofthe body when the individual experiences a period ofprolonged absence ofthe drug in
the body. Evidence for the physiological dependence aspect to smoking can be supported
by a study conducted by Royce, Hymnowitz, Corbett, Hartwell, & Orlandi in 1993. This
study investigated the differences between African-Americans and Caucasians in terms of
ability to quit smoking. The premise ofthe study took in two assumptions. The first was
that smoking is more socially acceptable among Caucasians than among African 
Americans. The second assumption was that African-Americans smoke fewer cigarettes
per day than Caucasians on the average. However, when it came to quitting, Caucasians
were more successful than African-Americans. Some ofthis could be explained by the
disproportionate distribution ofAfrican-Americans in lower socioeconomic strata. This
explanation could be used to explain psychological dependence. A more compelling
indicator ofthe difficulty African-Americans have quitting is pharmacological in origin.
African-Americans, although they smoke less, have the tendency to prefer menthol
cigarettes. Menthol cigarettes have both higher tar and higher nicotine dosages.
Consequently, the tolerance to nicotine is higher and a more traumatic withdrawal
response would occur ifa person were to attempt to quit. (Royce, Hymnowitz, Corbett,
Hartwell, & Orlandi, 1993)
12
The psychological dependence on cigarettes is another aspect ofbarriers to behavior
change that also has been studied at some length in the literature. In a 1994 study
conducted by Stretcher, most ofthe barriers were brought about from the person
experiencing fear. Although related to physiological dependence, the fear ofstress or
anxiety from the absence ofsmoking is psychological. This stress is caused by the
anticipation ofwithdrawal symptoms that will in all likelihood occur.
Another fear commonly perceived by people who are thinking about quitting smoking
is the fear ofgaining weight. Using cigarettes has long been associated with weight
control and as some research indicates a reason why people, usually women more than
men, begin to smoke.
Peer pressure is a commonly held explanation for adolescent smoking. A 1997 study
conducted by Wolfson, Forster, Claxton, and Murray found that close to 75% of
adolescents in the 8th, 9th, and 10th grades obtained cigarettes from friends or family
members. In many studies determining predictors for adult smoking, adolescent smoking
is commonly found to be the best predictor. (Paavola, Vartiainen, & Puska, 1996). The
final fear ofthe Strecher study was that fear offailure and relapse exists especially in
situations where others are present to encourage relapse.
Very closely related to psychological dependence is a concept the Health BeliefModel
referred to as self-efficacy. This construct ofthe HBM is commonly believed by many
researchers to be the strongest predictor ofbehavior change. (Strecher & Rosenstock,
1997)
13
Although much ofthe research shows the use ofonly a portion ofthe Health Belief
Model as a theoretical base for smoking behavior, one study used the entire model.
(Mikanowicz, Fitzgerald, Leslie, & Altman, 1999) In this research, employees were
administered a questionnaire that addressed the various constructs ofthe HBM. In
addressing perceived benefits ofsmoking behavior, 31 % oftobacco users felt smoking
helped control weight and 82% felt tobacco use helped reduce tension. The researchers
showed that a higher percentage ofsmokers perceived barriers to quitting than
nonsmokers and former smokers. In terms ofperceived susceptibility, a higher
percentage ofemployees who never used tobacco products were more likely to believe
smoking contributes to lung cancer and heart disease. However, in contrast to the 1992
Brownson study, the former users perceived less susceptibility than even current
smokers. In evaluating the cue to action component ofthe HBM, 63% ofsmokers
believed smoking was harmful and wanted to quit. However, only 41 % actually wanted
to enroll in a program to quit.
14
The Problem ofETS (Environmental Tobacco Smoke)
Environmental tobacco smoke or "secondhand smoke" has been classified as a human
(Group A) carcinogen by the Environmental Protection Agency. In a study by the EPA,
second hand smoke was estimated to cause 37,000 deaths from heart disease and 13,000
deaths from cancers each year in the U.S. (EPA, 1992). Although this study has some
detractors, it is still widely supported by the American Lung Association and other health
organizations.
Forty-eight ofthe fifty states have some restriction on smoking in public places. These
restrictions do vary from designated smoking areas to total bans. In terms ofthe
workplace, 23 states restrict smoking in the workplace in the private sector and 43 states
have restrictions in government workplaces. In August 1997, President Clinton signed an
executive order banning smoking in federal buildings. (ALA, 1999)
The Relationship between Smoking and the Workplace
Many workplaces now support some sort ofsmoking policy. Generally there are three
major categories ofsmoking policies. The first smoking policy would be a total ban
policy. As the name implies, this plan is where smoking is not permitted anywhere on
the premises. The second type is a partial ban. This policy is where smoking may be
permitted in designated areas ofthe workplace. The third type ofpolicy would be to
have no policy and smoking would be permitted anywhere at the worksite. (Mikanowicz
& Altman, 1995)
15
More than 80 % ofworkers are covered by an official workplace smoking policy
(Gerlach, Shopland, Hartman, Gibson, & Pechacek, 1997). However, less than halfare
actually protected by smoking policies that prohibit smoking in both the work area and
common areas ofthe workplace. Ofthose who work indoors, an estimated 58 million
Americans (40 million ofwhom are not smokers) are not protected by a smoke-free
workplace policy (Gerlach, Shopland, Hartman, Gibson, & Pechacek, 1997). This
research also found that white-collar workers (53.7 %) were more likely than service
workers (34.8 %) and blue-collar workers (27.4 %) to be covered by a smoke free policy.
A meta-analysis ofvarious workplaces in Australia and the United States that instituted
a total ban policy found significant results to support a perceived change in behavior.
Nineteen studies between 1986 and 1996 were evaluated in this analysis. The researchers
showed that 18 out ofthe 19 studies reported a daily decline ofsmoking rates and 17 of
the 19 studies reported a decline in smoking prevalence. With the advent ofthe smoking
ban, a reduction of602 million cigarettes consumed in Australia and a reduction of9.7
billion cigarettes per year in the United States could be attributed to a total ban on
smoking in the workplace. This reduction accounts for approximately 2% ofall
cigarettes smoked annually in each country. This study also made a projection that if
workplaces were universally smoke free, the reduction ofcigarettes consumed would
almost double to 1.17 billion in Australia and more than double to 20.7 billion in the
United States (Chapman, Borland, Scollo, Brownson, Domninello, & Woodward, 1999).
16
The United States and Australia are further along than other parts ofthe world in trying
to reduce exposure to ETS in the workplace. A study ofthe German metal industry
revealed that in that industry, only 30% ofthe employees were not allowed to smoke in
their immediate work area. (Brenner, Born, Novak, & Wanek, 1997) This study also
showed a greater social acceptance ofsmoking in that 60% ofnonsmoking blue-collar
workers and 52% ofnonsmoking white-collar employees would be bothered by passive
smoking ifsmoking were permitted in their work area. Despite this difference from the
United States and Australia, the German metal industry was shown to have one major
similarity. Smoking prevalence and smoking intensity among active smokers was
significantly lower ifthe employees worked at ajob where smoking was banned. This
study along with the Chapman et al 1999 study lends strong support for the effectiveness
ofa smoking ban in the workplace.
Some studies compare the differences between the effects oftotal bans versus partial
bans in terms ofbehavior changes among those employees who smoked. One such study
was conducted on 242 smokers who worked at places that had either total or partial ban
policies. The researchers concluded that smokers who worked at jobs that had total bans
were more likely to stop smoking during the workday than those employees who smoked
and worked at a place with only a partial ban. Italso found that smokers in partial bans
actually were more likely to increase there smoking during work hours. Although there
were many significant differences in behavior between the two groups, the attitudes about
workplace smoking policies ofboth groups were very similar, regardless ofsmoking
policy existing at their job. (Styles & Capewell, 1998)
17
Another study did a comparison ofone workplace with a total ban and another with no
smoking policy. This study was conducted at two hospitals, one was about to implement
a smoking ban (experimental group) and the other was not (control group). Both groups
were similar in age, gender, and occupational status. Both groups were evaluated 4
weeks before the smoking ban and again at 4 weeks after the smoking ban. Smokers in
the experimental group on the average reduced their smoking during work hours from
7.57 cigarettes per day before the smoking ban to 3.64 per day after the ban. There was
no significant increase in cigarettes during non-work hours and no significant change in
smoking behavior ofthe control group. (Brigham, Gross, Stitzer, & Felch, 1994)
Other studies discovered in the literature not only evaluate change in smoking behavior,
but also evaluate change in work behavior as a result ofan implementation ofa smoking
policy. One such study occurred when employees (smokers and nonsmokers) ofa
Scottish University evaluated their change in work habits, a majority ofthe workers did
not change their work habits. The work habits specifically measured in this study were
the amount oftime employees spent in the immediate work area before their shift actually
started, working beyond their scheduled time, lunch breaks, and the number ofhours
worked per week. However, ofthe minority that reported a change in work behaviors,
there were significant differences between smokers and nonsmokers. Eighteen percent of
smokers reported that they spend less time in their working area since the introduction of
the ban whereas only 0.2% ofnonsmokers reported doing the same. In terms ofspending
less time before the official start ofthe shift, 19.7% ofthe employees who smoke and
0.1 % ofthe employees who don't smoke reported this behavior change. In terms of
18
staying late, 16.7% smokers and 0.3% nonsmokers reported staying late. (Parry, Platt, &
Thomson, 1999)
These concepts could be considered the most problematic and may cause dissension
among smokers and nonsmokers. The notion that nonsmokers are less productive than
smokers is usually the precursor to the dissension. Although the dangers ofETS and
second hand smoke have been well documented, management must be sensitive not only
to the needs ofthe nonsmokers, but to the needs ofsmokers as well.
Summary
Cigarette smoking is a habit that not only affects the person doing it. It
also has effects psychologically and physiologically on others as well as the smoker.
This chapter provides a theoretical framework examining smoking behavior and how it
relates to the workplace. The Health BeliefModel is the theoretical framework used to
associate these aspects to this habit.
Chapter III outlines the sample, instrumentation, methodologies for
collection and analysis ofthe data to determine the relationship between smoking
behavior as it relates to the habit in general, and the presence or absence ofworkplace
smoking policies.
Chapter IV presents the results ofthe analysis as it relates to significant
literature presented here as well as the hypotheses.
Chapter V presents a discussion ofthe findings as well as the
recommendations for further research.
19
CHAPTER III
METHODS
Chapter III provides an outline ofthe sample research methodologies and statistical
treatment ofthe data obtained. The data were used to detect differences in attitudes and
behaviors oftobacco use in people in a work environment. Therefore, the subjects for
this study were acquired by contacting the local hospitals and getting permission to
distribute the questionnaire to employees at the hospital.
Research Design
The design consisted ofsurvey research using both descriptive and inferential statistics.
Two hospitals, both part ofthe same corporation, agreed to participate in the study. One
hospital was located in Mahoning County and the other hospital was located in Trumbull
County. They were categorized into three categories smokers, non-smokers and former
smokers by self-report. Descriptive statistics were used to understand the prevalence of
smoking behavior in this population augmented by various demographical characteristics.
The inferential statistics were used to develop two scales. One scale is a combination
measure ofknowledge and attitudes. The other scale is a measure ofan employee's
tolerance to workplace smoking.
20
Subjects.
The sample consisted of206 healthcare workers in a hospital environment. Although all
employees ofthe hospital were allowed to participate, the majority ofthe respondents
were nurses. There were both male and female respondents ranging from 22 to 82 years
ofage.
Instrument
The questionnaire used for this study was developed specifically for this research.
Realizing there were some differences in the demographics ofthe subjects, the
questionnaire was developed respecting cultural sensitivity. The questions were
developed to prevent bias against anyone based upon one's health choices. The 38-item
questionnaire was divided into three parts. The first part contained demographic,
occupational and personal smoking behavior questions. The second part was nineteen
Likert scale questions that asked knowledge and attitudes about smoking in general. The
responses for the second part were Strongly Agree, Agree, Disagree, and Strongly
Disagree. These questions were used to develop the Total Smoking Inventory Scale. The
third part was six Likert scale questions that asked the subject's opinion about workplace
smoking policy. The responses for these questions were also Strongly Agree, Agree,
Disagree, and Strongly Disagree. These questions were used to develop the Workplace
Permissiveness Scale. A copy ofthe questionnaire is located in Appendix A.
21
Collection ofData
Prior to collection ofdata, the researcher obtained an approval letter from the
Youngstown State University Human Subjects Committee contained in Appendix B.
The researcher also submitted a proposal to the hospital system Institutional Review
Board and was granted permission to administer the survey to both hospitals. The
approval letter from the Institutional Review Board ofthe employer is in Appendix C.
The researcher then agreed to meet with a member ofthe hospital on the dates to
administer the survey. The researcher distributed the questionnaires to employees during
the lunch hours i.e. 11 :00 a.m. to 2:00 p.m. One hospital was done one day and the other
was done the following day. The respondents were given the questionnaire on their way
in the hospital cafeteria and asked to return them on their way out. The letter ofconsent
that preceded the questionnaire is in Appendix D. In the letter ofconsent, it indicated to
participants that taking part in the study was voluntary. All potential subjects were
assured absolute anonymity. At the completion ofeach questionnaire, the letter of
consent was placed in a separate envelope from the survey. Once the survey was
collected, they were placed in an envelope and sealed until the researcher was ready to
enter the data. All data were entered into the SPSS Data sheet by the researcher. To
make sure the data was entered correctly, the researcher checked the data sheet for errors.
Statistical Treatment ofData
All analyses ofthe data were done with SPSS (Statistical Package for the Social
Sciences) for Windows 8.0(1998). The data were tested with both descriptive and
inferential statistical procedures.
22
The data from the first section ofthe questionnaire were used for most ofthe
descriptive statistics. Frequency distributions and percentages for answers to the first
thirteen items ofthe questionnaire were performed to assess for trends and for
comparison in later tests.
The Likert scale items 14-32 in the survey were standardized into z-scores. Some ofthe
items were reverse-scored. All ofthe items in this section were then tested for reliability
using the reliability procedure in SPSS. This procedure determined the reliability ofeach
item individually and then those items that were similar in terms ofthe correlation
coefficient were selected. Those items that were reliable (r 2: .31) were combined to
form an index ofsmoking knowledge and beliefs. Those items that were not reliable
were discarded from further analysis.
A bivariate linear regression analysis was performed using the smoking category
variable as the independent variable and the total smoking inventory score as the
dependent variable. The results ofthe tests along with the corresponding ANOVAs and
coefficients were used to assess Hypothesis 1.
The last portion ofthe questionnaire described items related to respondents' opinions on
workplace smoking policies. These items were summed to form a total score for
permissiveness for smoking in the workplace. All ofthe items were standardized
including the total score. A bivariate linear regression analysis was performed using
smoking as the independent variable and the work permissiveness score as the dependent
variable. The results ofthe tests along with the corresponding ANOVAs and correlation
coefficients were calculated to test Hypothesis 2.
23
Summary
This chapter described the research methodologies used to determine the relationship of
smoking as it relates to beliefs and knowledge about the consequences ofthe habit. It
also provides a foundation for the relationship ofsmoking behavior as it relates to the
formulation ofa workplace smoking policy.
Chapter IV presents the results ofthe study using the methodologies outlined in this
chapter.
Chapter V further explains the findings in discussion, implications, and
recommendations for further research.
24
CHAPTER IV
RESULTS OF THE ANALYSIS
Information from the smoking in the workplace questionnaire was used
to determine smoking behavior's relationship toward the knowledge and attitudes about
smoking. From the results, two scale scores were produced to determine the
predictiveness and relative significance ofperformance on these scales based upon
smoking behavior. One scale reviewed a combination ofattitudes and knowledge about
smoking cigarettes. The other scale reviewed opinions on smoking restrictions in a
healthcare setting.
Each participant completed the questionnaire that was divided into three parts. The
first part ofthe questionnaire identified the participant's demographical information
along with participant's smoking behavior. The second part ofthe questionnaire
consisted ofLikert Scale questions about knowledge, attitudes, and behavior as it related
to smoking using the various constructs ofthe Health BeliefModel. This part
comprised the questions that would be used in the development ofthe Total Smoking
Inventory Scale. The third part ofquestionnaire consisted ofLikert scale questions that
asked opinions about workplace smoking policy and were used to construct the
Workplace Permissiveness Scale.
Profile ofthe subjects
Three-hundred seventy-five questionnaires were distributed to employees of
two hospitals that were part ofthe same corporation. One was located in Youngstown,
Ohio in Mahoning County and the other was located in Warren, Ohio in Trumbull
County. Two hundred six returned the surveys for a response rate of55%.
25
The age ofthe participants ranged from 22 to 82 years ofage. When grouped in age
categories similar to CDC intervals, the greatest numbers ofparticipants were aged 22-44
years at 55% ofthe entire sample followed by 45-64 years accounting for 39% ofthe
sample. The mean age for the entire group was 43.9 years old. All ofthe subjects who
reported that they currently smoke were between 25 and 64 years ofage. The 18-24 year
old, and the 65 and older intervals reported no current smoking behavior (Table 1).
Table 1
Number and Percentage of Healthcare Workers Smoking Behavior by Age
(N=206)*
VARIABLE
Age 22-44
CIGARETTE
CIGAR OR PIPE
SMOKELESS
NEVER SMOKER
FORMER SMOKER
TOTAL
Age 45-64
CIGARETTE
CIGAR OR PIPE
SMOKELESS
NEVER SMOKER
FORMER SMOKER
TOTAL
Age 65 and over
CIGARETTE
CIGAR OR PIPE
SMOKELESS
NEVER SMOKER
FORMER SMOKER
TOTAL
* There were 2 missing
values
N
17
4
2
68
23
114
15o
1
47
17
80
oo
o
6
4
10
26
Percent by age
14.91%
3.51%
1.75%
59.65%
20.18%
100%
18.75%o
1.25%
58.75%
21.25%
100%
0%
0%
0%
60%
40%
100%
Nearly 80% ofthe respondents were female. Over 18% ofthe females smoked either
cigarettes, cigars or a pipe. Only 3% ofthe women reported smoking used cigars or
pipes. None ofthe women reported smokeless tobacco use. Fourteen percent ofmen
reported using cigarettes, cigars, and pipes. However, the frequency ofmen smoking
cigarettes was equal to the number ofmen smoking cigars or pipes. Factoring in
smokeless tobacco use, 21 % ofmen used tobacco products (Table 2).
Table 2
Number and Percentage of Healthcare Workers Smoking Behavior by Gender
(N=206)*
Percent by
VARIABLE N gender
Male
CIGARETTE 3 7.00%
CIGAR OR PIPE 3 7.00%
SMOKELESS 3 7.00%
NEVER SMOKER 24 55.80%
FORMER SMOKER 10 23.30%
TOTAL 43 100%
Female
CIGARETTE 29 17.90%
CIGAR OR PIPE 1 0.60%
SMOKELESS 0 0.00%
NEVER SMOKER 98 60.50%
FORMER SMOKER 34 21.00%
TOTAL 162 100%
*There was 1
missing value
27
The majority ofthe respondents had completed some post secondary education (29.6%
bachelor's degree, 23.8% beyond a bachelor's degree). In terms ofsmoking behavior,
18% ofthe respondents were current smokers (16% cigarette and 2% cigar or pipe), 2%
smokeless tobacco users, 20% were former smokers and 60% were never smokers. Sixty 
nine percent ofthe cigarette smokers had the educational level less than a completed
bachelor's or trade school degree. However, all ofthe respondents who reported using
other tobacco products such as cigars, pipes, and smokeless were at least trade school
graduates or higher. Over 50% ofthis population had some postgraduate education.
(Table 3).
28
Table 3
Number and Percentage of Healthcare Workers Smoking Behavior by Education
(N=206)*
VARIABLE N Percent
GED
CIGARETTE 0 0.00%
CIGAR OR PIPE 0 0.00%
SMOKELESS 0 0.00%
NEVER SMOKER 1 100.00%
FORMER SMOKER 0 0.00%
TOTAL 1 100%
High School Graduate
CIGARETTE 12 38.70%
CIGAR OR PIPE 0 0.00%
SMOKELESS 0 0.00%
NEVER SMOKER 16 51.60%
FORMER SMOKER 3 9.70%
TOTAL 31 100%
Some college or trade school
CIGARETTE 10 21.70%
CIGAR OR PIPE 0 0.00%
SMOKELESS 0 0.00%
NEVER SMOKER 25 54.30%
FORMER SMOKER 11 23.90%
TOTAL 46 100%
Trade school graduate
CIGARETTE 2 11.80%
CIGAR OR PIPE 0 0.00%
SMOKELESS 1 5.90%
NEVER SMOKER 11 64.70%
FORMER SMOKER 3 17.60%
TOTAL 17 100%
Bachelor's degree
CIGARETTE 4 6.60%
CIGAR OR PIPE 2 3.30%
SMOKELESS 0 0.00%
NEVER SMOKER 41 67.20%
FORMER SMOKER 14 23.00%
TOTAL 61 100%
Post graduate degree
CIGARETTE 4 8.20%
CIGAR OR PIPE 2 4.10%
SMOKELESS 2 4.10%
NEVER SMOKER 28 57.10%
FORMER SMOKER 13 26.50%
TOTAL 49 100%
* 1 missing value
29
In tenns ofracial composition, 89% ofthe sample was Caucasian. Ofthis group,
18.5% were current smokers and 22% were fonner smokers. Twenty-nine percent of
African-Americans who responded were smokers as opposed to 14% who were fonner
smokers. None ofthe Native Americans or Asian Americans reported that they were
currently smoking. However, 66% Asian Americans were fonner smokers (Table 4).
30
Table 4
Number and Percentage of Healthcare Workers Smoking Behavior by Race
(N=206)*
VARAIBLE N Percent
Native-American
CIGARETTE 0 0.00%
CIGAR OR PIPE 0 0.00%
SMOKELESS 0 0.00%
NEVER SMOKER 8 100.00%
FORMER SMOKER 0 0.00%
TOTAL 8 100%
Asian-American
CIGARETTE 0 0.00%
CIGAR OR PIPE 0 0.00%
SMOKELESS 0 0.00%
NEVER SMOKER 2 66,67%
FORMER SMOKER 1 33.33%
TOTAL 3 100%
African-American
C!G.ARETTE 2 28.60%
CIGAR OR PIPE 0 0.00%
SMOKELESS 0 0.00%
NEVER SMOKER 4 57.10%
FOR~.~ER S~1f~OKER 14.30~/o
TOTAL 7 100%
U'.t""- __ni_ A ",",r,,?i__n
I lt~tJ0lln ....-1\11 n;;,1 ''-'011
CIGARETTE 0 0.00%
CIGAR OR PIPE " " nnol
SMOKELESS 1 100.00%
....\/in C""""I/rn 1'\ " "nOJI'\IC V Cl"I. ';'IVIVI".Cl"I. v V.VV'IO
FORMER SMOKER 0 0.00%
"T""""'''T''A I 1'\ .A n",nlIVIf'\L U IUU70
Caucasian
CiGARETTE 30 16.30%
CIGAR OR PIPE 4 2.20%
SiviOKELESS 2 ?i.'iOO/O
NEVER SMOKER 107 58.20%
FORMER SMOKER 41 22.30%
TOTAL 184 100%
Other
CIGARETTE 0 0.00%
CIGAR OR PIPE 0 0.00%
SMOKELESS 0 0.00%
NEVER SMOKER 2 100.00%
FORMER SMOKER 0 0.00%
TOTAL 2 100%
* 2 missing values
31
Before a test for significance was run, some manipulation ofthe data were necessary.
Some ofthe items for the Total Smoking Inventory were reverse-scaled so that a higher
value reflected a pro-smoking behavior, knowledge, or attitude. (i.e. 1=antismoking,
4=pro-smoking)
After reverse scaling was completed, an item analysis using the reliability procedure
was implemented. The item analysis was conducted on 19 items. Based on the results,
nine ofthe items were found not to be reliable because the corrected item correlations
individually placed back in with the 10 chosen items to double check the current
correlation range. None ofthe nine eliminated items could be assimilated back in with
the 10 selected. (Table 5)
Coefficient alpha for the Total Smoking Inventory Scale was .75. Because the sample
was one ofconvenience and used both for the item analysis and the computation ofthe
coefficient alpha, the reliability is likely to be an overestimate ofthe general population
Coefficient Alpha. A total score was computed by summing the total ofthe values ofthe
10 selected items. The total score was then standardized by the use ofz-scores. This
would make it easier to conduct further analyses.
32
Table 5
Reliability Analysis of Total Smoking Inventory Items
VARIABLE
Nicotine is addictive
Tobacco leads to disease
Tobacco costs 100 billion per year
Smoking should not be regulated
Advertising influences tobacco use
Promotions influence tobacco use
Secondhand smoke effects are exaggerated
Smoking laws should be enforced
Cigarette taxes are unfair
It's okay if my friends smoke
Mean
1.2606
1.1543
1.5479
2.3032
2.1064
2.2766
1. 6755
2.2500
1. 8989
2.3511
Std Dev
.5385
.4171
.9881
1.0283
.8461
.8389
.7712
.8629
.9223
1.3098
Cases
188.0
188.0
188.0
188.0
188.0
188.0
188.0
188.0
188.0
188.0
Statistics for
SCALE
Mean
18.8245
Variance
23.8032
Std Dev
4.8789
N of
Variables
10
Item-total Statistics
VARIABLE
Scale
Mean
if Item
Deleted
Scale
Variance
if Item
Deleted
Corrected
Item 
Total
Correlation
Alpha
if Item
Deleted
Nicotine is addictive
Tobacco leads to disease
Tobacco costs 100 billion per year
Smoking should not be regulated
Advertising influences tobacco use
Promotions influence tobacco use
Secondhand smoke effects are exag
Smoking laws should be enforced
Cigarette taxes are unfair
It's okay if my friends smoke
Reliability Coefficients
17 . 5638
17.6702
17.2766
16.5213
16.7181
16.5479
17.1489
16.5745
16.9255
16.4734
21. 9264
21.8479
19.9552
18.7963
20.3212
20.2063
19.5606
19.5934
18.6468
17.3950
.3147
.4569
.3253
.4430
.3626
.3836
.5348
.4536
.5405
.4295
.7385
.7311
.7390
.7201
.7315
.7286
.7093
.7186
.7043
.7300
N of Cases
Alpha = .7459
188.0 N of Items
33
10
The rest ofthe chapter was organized by the hypotheses that guided this study.
Hypothesis 1: There are significant differences among smokers, nonsmokers, and former
smokers in terms ofhealth beliefs about smoking.
This hypothesis evaluates the significance ofthe differences in smoking beliefs,
behaviors, and knowledge about cigarette use among three groups; smokers, never
smokers, and former smokers.
A linear regression equation was calculated using actual smoking behavior as
independent variable and the total inventory score as the dependent variable. The
following equation was derived from analysis:
Y predicted inventory score = -.69X smoking behavior + 3.692
The equation indicates that smoking behavior predicts higher scores on the total
smoking inventory scale. The slope weight is negative because the value assigned to
smokers is lower than nonsmokers and former smokers. The regression equation was
then standardized so the slope weight could be interpreted more easily (Figure 1).
Predicted Zinventory score = -.54 smoking behavior
The correlation between the total inventory score and smoking behavior was r2 = .29,
t(203) = 92, p:::: .001.
34
Figure 1. The Relationship Between Smoking and Total Smoking Inventory
4
0
3
c
~ 00
c 0+0'
C 2 c 0Q)
c>
0c
c
C>c
:.i20
E
Cf)
m 0+0'
0I--
N
-1
0 0
0
0 0
-2 0 0
0 1 2 3 4 5 6
SMOKING
On the smoking line 1= cigarette smoker, 2= cigar or pipe smoker, 3= smokeless tobacco,
4= never smoker, and 5= former smoker
In order to determine significance in the differences in Total Smoking Inventory Scale
scores among the various groups categorized by smoking behavior, a one-way Analysis
ofYariance (ANOYA) was conducted to evaluate the relationship between smoking
behavior and the differences in the scores on the Total Smoking Inventory scale. The
independent variable, smoking, included five levels, cigarette smokers, cigar or pipe
smokers, smokeless tobacco users, never smokers, and former smokers. The dependent
variable was the score on the Total Smoking Inventory scale. (Figure 1)
35
The ANOVA was significant, F (4,200) = 36.96, p < .001. The strength ofthe
relationship between smoking and scores on the scale, as assessed by eta-squared, were
moderately strong as the smoking factor accounted for 43% ofthe variance ofthe
dependent variable.
Table 6
ANOVA Hypothesis 1
Dependent Variable: Zscore(Total Smoking Inventory)
Type III df Mean F Sig. Eta
Source Sum of Square Squared
Squares
Corrected 87.108 4 21.777 36.964 .000 .425
Model
Intercept 5.009 1 5.009 8.502 .004 .041
SMOKING 87.108 4 21.777 36.964 .000 .425
Error 117.829 200 .589
Total 204.937 205
Corrected 204.937 204
Total
a R Squared =.425 (Adjusted R Squared =.414)
Follow-up tests were conducted to evaluate pairwise differences among the means.
Because the variance among the five groups ranged from .49 to 1.54, post-hoc
comparisons assuming variances to be homogenous were conducted. The post-hoc test
used was the Tukey HSD. There were significant differences in the means ofcigarette
smokers with all other categories except cigar and pipe smokers. Differences in the
means among smokeless tobacco users, never smokers, and former smokers were not
significant. (Table 7)
36
Table 7
Tukey HSD Post hoc Comparisons
Dependent Variable: Zscore(Total Smoking Inventory)
Mean Difference (I-J) Std. Error Sig.
CIGARETTE CIGAR OR PIPE -.1216978 .407 .998
SMOKELESS 1.9962710 .463 .000
NEVER SMOKER 1.7431452 .152 .000
FORMER 1.3782713 .178 .000
SMOKER
CIGAR OR PIPE CIGARETTE .1216978 .407 .998
SMOKELESS 2.1179688 .586 .003
NEVER SMOKER 1.8648430 .390 .000
FORMER 1.4999691 .401 .002
SMOKER
SMOKELESS CIGARETTE -1.9962710 .463 .000
CIGAR OR PIPE -2.1179688 .586 .003
NEVER SMOKER -.2531258 .449 .980
FORMER -.6179997 .458 .660
SMOKER
NEVER SMOKER CIGARETTE -1.7431452 .152 .000
CIGAR OR PIPE -1.8648430 .390 .000
SMOKELESS .2531258 .449 .980
FORMER -.3648739 .135 .053
SMOKER
FORMER SMOKER CIGARETTE -1.3782713 .178 .000
CIGAR OR PIPE -1.4999691 .401 .002
SMOKELESS .6179997 .458 .660
NEVER SMOKER .3648739 .135 .053
Based on observed means.
* The mean difference is significant at the .05 level.
37
Hypothesis 2: There are significant differences between smokers, nonsmokers, and
former smokers in terms ofattitudes toward workplace smoking policies
This hypothesis evaluates differences among smokers, former smokers, and never
smokers and their attitudes toward workplace smoking policies. A linear bivariate
regression was conducted using smoking behavior as the independent variable and the Z
score ofthe total score ofthe workplace permissiveness scale questions. The regression
equation derived from the sample was the following:
Y predicted smoking permissiveness = .335X smoking behavior - 1.232
The results suggested that people who do not smoke will support more restrictive
workplace smoking policies and conversely those who do smoke will support those
policies that will allow for less restriction. The slope is positive in this equation because
a higher score on this score indicates support for more restrictions and the value assigned
to smokers is lower than never smokers and exsmokers. (Figure 2) To better understand
how well the slope weight predicts a workplace permissiveness scale score, the above
equation was standardized to the following:
Predicted Zworkplace permissiveness = .42 Smoking behavior
The correlation between smoking and the workplace permissiveness score was r2 =.18, t
(203) = 6.7, p<.OOl. In order to determine significance in differences in Work
Permissiveness scores among the various groups categorized by smoking behavior, a one 
way Analysis ofYariance (ANOYA) was conducted to evaluate the relationship between
smoking behavior and the differences in the scores on the Work Permissiveness scale.
The independent variable, smoking, included five levels, cigarette smokers, cigar or pipe
38
smokers, smokeless tobacco users, never smokers, and former smokers. The dependent
variable was the score on the Work Permissiveness scale.
Figure 2. The Relationship Between Smoking and Workplace Permissiveness
10
8 0
enen
(J)c
6?(J)
.~ en
0
.!QE
4?
I-
(J)a..
(J) 2?
()ro
a.
~ 0 r;! -I-
0 ?0 O? 0 0 0
S B ~ 08 0 BB B 0
N - B 0 0 0
0
-2? 0
0 0
0 0-4
0 1 2 3 4 5 6
SMOKING
On the smoking line 1= cigarette smoker, 2= cigar or pipe smoker, 3= smokeless tobacco,
4= never smoker, and 5= former smoker
The ANOVA was significant, F (4,200) = 9.79, p = .019. The strength ofthe
relationship between smoking and scores on the scale, as assessed by eta-squared, were
39
moderately strong as the smoking factor accounted for 19% ofthe variance ofthe
dependent variable. (Table 8)
Source
Corrected
Model
Table 8
ANOVA Hypothesis 2
Dependent Variable: Zscore(Workplace Permissiveness)
Type III df Mean F Sig. Eta
Sum of Square Squared
Squares
39.155 4 9.789 11.866 .000 .192
Intercept
SMOKING
4.606
39.155
1
4
4.606
9.789
5.584
11.866
.019
.000
.027
.192
Error 164.990 200 .825
Total 204.149 205
Corrected 204.145 204
Total
a R Squared =.192 (Adjusted R Squared =.176)
Follow-up tests were conducted to evaluate pairwise differences among the means.
Because the variance among the five groups ranged from .03 to 2.25 and Levene's Test
for Equality ofError Variances was not significant, p= .255, post-hoc comparisons
assuming variances to be nonhomogenous were conducted. Since Tukey's HSD test is
more appropriate when the variances are homogenous, the post-hoc test used with this
hypothesis was the Dunnett's C test. There were significant differences in the means of
cigarette smokers with former smokers and never smokers. The differences were not
significant when cigarette smokers were compared to cigar/pipe smokers and smokeless
tobacco users. Similar to cigarette smokers, differences in the means ofcigar/pipe
smokers were significant when compared to never smokers and former smokers.
Smokeless tobacco users had no significant differences among any other category. Never
smokers and former smokers also showed no significant difference in means.
40
Table 9
Dunnett's C Post hoc Comparisons
Dependent Variable: Zscore(Workplace Permissiveness)
Mean Difference (I-J) Std. Error Sig
CIGARETTE CIGAR OR PIPE -.1502096 .482
SMOKELESS -.4429258 .548
NEVER SMOKER -1.1147335 .180 *
FORMER SMOKER -1.2384416 .211 *
CIGAR OR PIPE CIGARETTE .1502096 .482
SMOKELESS -.2927162 .694
NEVER SMOKER -.9645239 .462 *
FORMER SMOKER -1.0882320 .474 *
SMOKELESS CIGARETTE .4429258 .548
CIGAR OR PIPE .2927162 .694
NEVER SMOKER -.6718077 .531
FORMER SMOKER -.7955158 .542
NEVER SMOKER CIGARETTE 1.1147335 .180 *
CIGAR OR PIPE .9645239 .462 *
SMOKELESS .6718077 .531
FORMER SMOKER -.1237081 .160
FORMER SMOKER CIGARETTE 1.2384416 .211 *
CIGAR OR PIPE 1.0882320 .474 *
SMOKELESS .7955158 .542
NEVER SMOKER .1237081 .160
Based on observed means.
* p < .05
Summary
This chapter presented the results ofthe analysis. Both hypotheses were confirmed as a
result ofthe analysis. Also cigarette smoking is a moderately strong predictor for a
higher score on the Total Smoking Inventory Scale and a lower score on the Workplace
Permissiveness Scale.
Chapter V will present some conclusions and recommendations for further research
beyond the scope ofthis study.
41
CHAPTER V
SUMMARY, CONCLUSIONS AND RECOMMENDATIONS
Summary
Despite the documented harmful effects ofsmoking, not only on the smoker but
anyone in the immediate area, smoking remains a major barrier to a healthy lifestyle for
many. Close to 50 million adults in the United States smoke (CDC, 1998). In the
workplace, smoking is increasingly being banned. Banning smoking in the workplace
seems to be an effective tool in reducing the frequency with which one smokes.
However, there are many workplaces that tolerate smoking. Thirty-eight percent ofthe
adult smoking population can smoke. For every person that smokes in the workplace,
two coworkers that do not smoke will be exposed to ETS (Gerlach, Shopland, Hartman,
Gibson & Pechacek, 1997).
The purpose ofthe present study was to measure various aspects ofsmoking behavior
among a working population in a local healthcare setting. Relationships between
smoking behaviors, age, education and race were examined. The relationship between
smoking behavior, knowledge, and attitudes about smoking was evaluated for linear
relationships and differences in attitudes and knowledge based upon the smoking
behavior ofthe sample. Two hundred and six respondents completed the questionnaire.
The respondents were employees oftwo hospitals located in Northeast Ohio.
42
The theoretical basis for the study was the Health BeliefModel. This model
explained willingness ofa person to make a health behavior change based upon various
perceptions ofthe individual. These perceptions basically weighed the benefits of
changing the behavior; assess the risks in maintaining the negative health behavior, and
evaluating the sacrifices and the difficulty in making the behavior change.
These perceptions were heavily based on the beliefs and the knowledge that person has
about the behavior previous to making the choice whether or not to undertake the positive
behavior. In terms ofsmoking behavior, based upon the linear relationship found in
Hypothesis 1, most ofthose people who continued to smoke did so because the mindset
was such that the barriers to quitting did not outweigh the perceptions ofsusceptibility
and severity ofand to illness.
Banning smoking in the workplace, although thrust upon smokers, is a cue to behavior
change. It changes the perception oftheir ability to deal with the sacrifices they must
make in order to keep theirjob. Although there was a linear relationship and significant
differences among smokers, nonsmokers and former smokers, it was not as strong in
Hypothesis 2 as in Hypothesis 1. Thus the cue to action with the advent ofa workplace
smoking ban has had an effect on smokers.
43
Conclusions
There was a linear relationship between smoking behavior and the score on the Total
Smoking Inventory scale (which consisted ofknowledge and attitudes about smoking
behavior). Hypothesis 1 was confirmed, as there were significant differences among the
smokers, never smokers, and former smokers score on this scale. Thus indicating there
still are some differences in knowledge and attitudes. Smokers tended to have less
knowledge about the effects ofsmoking and the attitudes ofsmokers tended to favor a
lowered perception ofsusceptibility and severity.
There was a linear relationship between smoking behavior and the score on the Work
Permissiveness scale (which asked questions about smoking in the workplace).
Hypothesis 2 was confirmed, as there were significant differences among the smokers,
never smokers, and the former smokers on this scale. This result would indicate that
nonsmokers (both former smokers and never smokers) favored more stringent workplace
smoking policies than smokers.
The participants in the study displayed differences in their beliefs and knowledge of
smoking based upon their own behavior. This is consistent with the literature. Most of
the subjects thought that smoking was not a positive health behavior, but the severity of
the behavior was the construct where the difference in knowledge and attitude prevailed.
The participants in this study also showed differences in their beliefs about workplace
smoking policy based upon their own behavior. The differences in this evaluation were
not as great as the differences in knowledge and attitudes about smoking behavior in
general. However, there was more acceptance ofpolicies against smoking in the
44
workplace by smokers. Ifemployers would offer a smoking cessation program in
conjunction with a ban on smoking, this may lead more smokers to quitting.
Smoking behavior in relationship with other demographic factors such as age, gender,
race, and education remained fairly consistent with the literature. A majority ofthe
smokers were aged 25-44. Contradictory to the literature, the smoking rates among men
were less than women. However ifyou added the use ofsmokeless tobacco, rates ofuse
in men would surpass women. African-Americans had higher smoking rates than
Caucasians. This is not consistent with the literature and probably due to the sample
being a sample ofconvenience. Those subjects whose highest educational attainment
was a high school diploma were more likely to be smokers than those who had either a
college or trade school degree.
Recommendations
Studies similar to the one-presented need to be conducted in Mahoning and Trumbull
Counties. Particularly those worksites that have little or no ban on smoking need to be
evaluated. Due to the underepresentation ofmen and various minority races including
African-Americans, Native Americans, and Asian Americans, more studies with a better
representation ofthese groups should be conducted.
There was a strong relationship between education and smoking behavior. However,
an interesting trend that should be investigated is that men who have the highest
education levels (i.e. graduate degrees) preferred other tobacco products such as
smokeless tobacco, cigars, or tobacco pipes. The health effects such as cancers ofmouth
and larynx are different than the effects ofcigarette smoking (Hahn & Payne 1998). So
prevalence rates ofthese cancers and other conditions related to use ofalternative
45
tobacco products should be examined for people, especially men with an education
beyond a bachelor's degree.
This study was intended to explore differences in knowledge, beliefs about smoking in
general, and opinions about smoking in the workplace among people who smoke and
people who do not smoke. Ifyou are able to gain a better understanding ofthe
differences in beliefs and knowledge ofsmoking behavior among smokers and
nonsmokers, it provides a foundation for a dialogue to address changes that affect both
parties as individuals making healthy choices and employees ofhospitals.
46
BIBLIOGRAPHY
American Lung Association (1999). Lung Disease Data, 1998-99. New York:
American Lung Association Pub.
Bandura, A. (1977). Social Learning Theory. New Jersey: Prentice Hall, Inc.
Brenner, H., Born, J., Novak, P., & Wanek, V. (1997). Smoking behavior and
attitude toward smoking regulations and passive smoking in the workplace: A study
among 974 employees in the German metal industry. Preventive Medicine 26(1), 138 
143.
Brigham, J., Gross, J., Stitzer, M., & Felch, L. (1994). Workplace smoking helps
employees reduce cigarette use. The Brown University Digest ofAddiction Theory and
Application 13(9), 4-6.
Brownson, R (1992). Demographic and social differences in beliefs about the
health effects ofsmoking. American Journal ofPublic Health 82(1), 99-103.
Chapman, S., Borland, R, Scollo, M., Browson, R, Dominello, A., & Woodward,
S. (1999). The impact ofsmoke-free workplaces on declining cigarette consumption in
Australia and the United States. American Journal ofPublic Health. 89, 1018-1023.
Chapman, S., Wong, W., & Smith, W. (1993). Self-exempting beliefs about
smoking and health: Differences between smokers and exsmokers. American Journal of
Public Health 83(2), 215-219.
Centers for Disease Control and Prevention (1998) Tobacco use among U.S.
racial/ ethnic minority groups. A report ofthe Surgeon General. Washington D.C.
DHHS Publication.
47
Centers for Disease Control and Prevention (1993) Cigarette smoking-attributable
mortality and years ofpotential life lost. Morbidity and Mortality Weekly Report 43,
469-472.
U.S. Department ofHealth and Human Services, (1990). Reducing the health
consequences ofsmoking: 25 years ofprogress. A report ofthe Surgeon General.
Washington D.C., DHHS Publication No. 90-8416.
Environmental Protection Agency, (1992). Respiratory health effects ofpassive
smoking: lung cancer and other disorders. Washington DC: EPA
Gerlach, K., Shopland, D., Hartman, A., Gibson, J., & Pechacek, T. (1997).
Smoking policies in the United States. Tobacco Control 6, 199-206.
Henningfield, J. & Keenan, R. (1993). The anatomy ofnicotine addiction. The
Journal ofHealth Behavior, Education & Promotion 17(2), 12-19.
Hahn, D. & Payne, W. (1998). Focus on Health. (pp. 202-225). St. Louis:
Mosby-Year Book, Inc.
Mikanowicz, C., & Altman, N. (1995). Developing smoking policies in the
workplace. Journal ofHealth Education 26(3), 183-185.
Mikanowicz, C., Fitzgerald, D., Leslie, M., & Altman, N. (1999). Medium-Sized
business employees speak out about smoking. Journal ofCommunity Health 24(6),
439-449
Paavola, M., Vartiainen, E., & Puska, P. (1996). Predicting adult smoking: the
influence of smoking during adolescence and smoking among friends and family.
Health Education Research 11(3), 309-315.
48
Parry, 0., Platt, S., & Thomson, C. (1999). The perceived impact ofa workplace
smoking ban on the routines ofsmokers. Health Education Journal 58, 157-164.
Royce, J., Hymnowitz, N., Corbett, K., Hartwell, T., & Orlandi, M. (1993).
Smoking factors among African-Americans and Whites. American Journal ofPublic
Health 83(2), 220-226.
Shopland, D., Hartmann, A., Gibson, J., Mueller, M., Kessler, L., & Lynn, W.
(1996). Cigarette smoking among U.S. adults by state and region. Journal ofThe
National CancerInstitute 88(23), 1748-1758.
Statistical Package for the Social Sciences. (1998). (Version 8.0 for Windows)
[computer software]. New Jersey: Prentice-Hall, Inc.
Strecher, V., & Rosenstock, 1. (1997). The health be1iefmodel. In Glanz, K.,
Lewis, F., & Rimer, B. Health behavior and education: Research and practice. (pp.41 
59). San Francisco: Jossey-Bass.
Strecher, V. (1994). Psychosocial aspects ofchanges in smoking cessation
behavior. Patient Education and Counseling 7(3), 249-262.
Styles, G., & Capewell, S. (1998). No smoking at work: The effect ofdifferent
types ofworkplace smoking restrictions on smokers' attitudes, behavior and cessation
intentions. Health Education Journal 57, 385-395.
Wolfson, M., Forster, J., Claxton, A., & Murray, D. (1997). Adolescent smokers
provision oftobacco to other adolescents. American Journal of Public Health 87(4),
649-651.
49
Appendix A: Smoking in the Workplace Questionnaire
Smoking in the Workplace Questionnaire
This questionnaire allows you to give your opinion about smoking and smoking in
the workplace. Your honest responses are anonymous and will not affect yourjob. Please
circle the response that best represents your opinion. It should take approximately 15
minutes to complete. Your assistance is appreciated.
1. What is your age?
2. What is your gender?
a. Female
3. Are you a parent?
a. Yes
b. Male
b. No
4. What is the highest level ofeducation you have attained?
a. did not graduate high school e. Trade School graduate
b. OED f. College graduate
c. High School graduate g. Post-baccalaureate or other
d Some college or trade school professional certification
5. What is your race?
a. Native Alaskan/American c. Black
b. Asian/Pacific Islander d. Hispanic
6. What is your working category?
a. Laborer c. Clerical
b. Management d. Technical/Skilled
e. White
f. Other
e. Contract
f. Other please specify _
e. don't smoke
f. other
7. How many hours do you work per week?
a. Less than 10 c. 21-30 hours e. 40+ hours
b. 10-20 hours d. 31-40 hours
8. How would you identify your current smoking status?
a. cigarette smoker c. smokeless user (chew dip or snuff) e. ex-smoker
b. cigar or pipe d. never smoked
9.Ifyou smoke, how old were you when you had your first tobacco product
(cigarette, cigar, pipe, or smokeless tobacco)?
a. less than 12 c. 17-20
b. 12-16 d. don't know
10. Ifyou smoke, how much do you smoke?
a. 2 or more packs per day
b. between land 2 packs per day
c. less than a pack per day
51
d. less than a pack per week
e. don't smoke
11. Ifyou smoke, have you tried to quit smoking and been unable?
a. Yes b. No
12. How often do you try to quit smoking each year?
a. Don't smoke c. 2-3 times
b. Once a year d. 4 or more
13. Ifyou do not smoke do you ask people to stop smoking when you are in the
area?
a. Yes b. No
Please circle the appropriate number to show how strongly you agree or disagree with
the following statements.
Strongly Strongly
Agree Agree Disagree Disagree
14. Nicotine is a strong addictive substance 1 2 3 4
15. Tobacco use increases the risk ofmany diseases
such as lung cancer and heart disease. 1 2 3 4
16. Smoking helps reduce stress or anxiety 1 2 3 4
17. Tobacco use costs the American Public
$100 billion per year in healthcare dollars. 1 2 3 4
18. Laws regarding smoking behavior should be 1 2 3 4
strictly enforced
19. Itdoesn't bother me ifmy friends smoke. 1 2 3 4
20. Nicotine in cigarettes makes it difficult for people to quit 1 2 3 4
21 Smoking is a personal matter and should not be
regulated for adults 1 2 3 4
22. I don't mind being in a room with others that smoke 1 2 3 4
23.Tobacco use is higher among people who are poor 1 2 3 4
52
24. Tobacco advertisements influences a person's use of
tobacco products 1 2 3 4
25. Exercise and sports help get all the tar 1 2 3 4
26. Smoking is not as harmful as you hear 1 2 3 4
27.The tobacco companies should be responsible
who are addicted to tobacco. 1 2 3 4
28. The tobacco companies' promotional activities such as
free clothing influence people's choice oftobacco products 1 2 3 4
29. Smoking helps control your weight. 1 2 3 4
30. The effects ofsecond hand smoke are exaggerated 1 2 3 4
31. People who break smoking laws should be
fined ( in the same manner we enforce speeding) 1 2 3 4
32. The tax placed on smokers when they
buy cigarettes is unfair 1 2 3 4
33. Smoking should be allowed in conference rooms at work 1 2 3 4
34. Smoking should be allowed in rest rooms at work 1 2 3 4
35. Smoking should be allowed in lunchroom at work 1 2 3 4
36. Smoking should be allowed in company vehicles 1 2 3 4
37. Smoking should be allowed in hallways at work 1 2 3 4
38. Smoking should be allowed outside the
building during lunch hours at lunchtime 1 2 3 4
53
Appendix B: Human Subjects Committee Approval
Youngstown State University lOne University Plaza IYoungstown, Ohio 44555-0001
July 2, 1999
Dr. Carolyn Mikanowicz, Associate Professor, for
Nicholas V. Cascarelli
Department of Health Professions
CAMPUS
RE: Human Subjects Research Protocol #65-99
Dear Dr. Mikanowiczand Mr. Cascarelli:
The Human Subjects Research Committee has reviewed your protocol, HSRC#65-99, "Differences
Among Smokers, Former Smokers, and Non-smokers: A Work Site Investigation," and determined
that it is exempt from review based on DHHS Category 4 subject to the following condition:
(1) The researcher should provide the Committee with written consent from each of the
sites he will distribute the questionnaire.
Any changes in your research activity should be promptly reported to the Human Subjects Research
Committee and may not be initiated without HSRC approval except where necessary to eliminate
hazard to human subjects. Any unanticipated problems involving risks to subjects should also be
promptly reported to the Human Subjects Research Committee. Best wishes in the conduct ofyour
study.
Sincerelyy
Eric Lewandowski
Administrative Co-chair
Human Subjects Research Committee
cc:ECL
c: Mr. Joseph Mistovich, Chair
Department of Health Professions
55
Appendix C: Letter ofPennission from Employer
~ %~ ~ HUMILITY OF MARY~~ Health Partners
December 15, 1999
Nicholas Cascarelli
Department ofHealth Professions
Youngstown State University
Youngstown, Ohio 44555
RE: IRB Approval 99-038
Dear Mr. Cascarelli,
At the Institutional Review Board meeting held on December 15, 1999, your project
entitled, "Differences AmongSmokers, Former Smokers andNon Smokers, A
Worksite Investigation" was reviewed. The IRB committee has concurred with the
Chairman's decision ofexpedited approval ofthe protocol and the informed consent
document for one year, expiring on December 15, 2000. The approval number for the
protocol is 99-038 and should be used in all future correspondence.
According to federal guidelines, all human research projects are approved for one year. If
the project lasts for more than one year, an annual progress report must be submitted to
the IRB with a request for reapproval.
Ifyou have any questions, please do not hesitate to contact me at (330) 480-3341.
Sincerely,
co '(jJt:::LL_____._
Chatrchai Watanakunakom, M.D.
Chairperson
Institutional Review Board
CW/mdc
St. Elizabeth Health Center
1044 Belmont Avenue / Youngstown, Ohio 44501 / (330)746-7211
57
MEMBER OF CATHOLIC HEALTHCARE PARTNERS
Appendix D: Letter ofConsent
Letter OfConsent
Dear participant,
Weare conducting a study to examine beliefs about smoking and smoking in the
workplace. In this study you will be asked to answer a 38-item questionnaire. Your
participation should take about 15 minutes. There are no risks to you as all the responses
you make will be strictly anonymous. So when the results are recorded, no one will be
able to identify you.
Your participation in this study is totally voluntary and you may withdraw at
anytime with no negative consequences. Ifyou wish to withdraw, simply hand in the
questionnaire.
Ifyou have any questions regarding the study, you may feel free to contact:
Carolyn Mikanowicz Ph.D.
Professor, Dept. ofHealth Professions
Youngstown State University
Youngstown, Ohio 44555
(330) 742-3658
or Nicholas Cascarelli
Dept. ofHealth Professions
Youngstown State University
Youngstown, Ohio 44555
(330) 742-3327
(330) 372-6000 x146
I understand the study described above. I am at least 18 years old and agree to
participate.
Signature
59
Date