Health Psychology
2009, Vol. 28, No. 6, 753–761
© 2009 American Psychological Association
0278-6133/09/$12.00 DOI: 10.1037/a0016943
Social–Cognitive Beliefs, Alcohol, and Tobacco Use: A Prospective
Community Study of Change Following a Ban on Smoking
in Public Places
Sheina Orbell, Patrick Lidierth, Caroline J. Henderson, Nicolas Geeraert, Claudia Uller,
Ayse K. Uskul, and Maria Kyriakaki
University of Essex, United Kingdom
Objective: To examine social– cognitive change associated with behavior change after the introduction of a
smoke-free public places policy. Design: Adults (N ⫽ 583) who use public houses licensed to sell alcohol
(pubs) completed questionnaires assessing alcohol and tobacco consumption and social– cognitive beliefs 2
months prior to the introduction of the smoking ban in England on July 1, 2007. Longitudinal follow-up (N ⫽
272) was 3 months after the introduction of the ban. Main outcome measures: Social– cognitive beliefs, daily
cigarette consumption, and weekly alcohol consumption. Results: Smokers consumed considerably more
alcohol than did nonsmokers at both time points. However, a significant interaction of Smoking Status ⫻ Time
showed that while smokers had consumed fewer units of alcohol after the ban, nonsmokers showed an
increase over the same period. There was a significant reduction in number of cigarettes consumed after the
ban. Subjective norms concerning not smoking, and perceived severity of smoking-related illness increased
across time. Negative outcomes associated with not smoking were reduced among former smokers and
increased across time among smokers. Regression analyses showed that changes in subjective norm and
negative outcome expectancies accounted for significant variance in change in smoking across time.
Conclusion: Results suggest that the smoking ban may have positive health benefits that are supported by
social– cognitive change.
Keywords: alcohol, tobacco, protection motivation theory, theory of planned behavior, health action
process approach
2004). A number of studies have reported positive behavioral
effects of public health efforts to reduce smoking-related disease.
For example, a comprehensive tobacco control program including
price increases, media campaigns, and smoking restrictions begun
in 1989 in Californiawas associated with a steady increase in
smoking cessation and a decrease in number of cigarettes consumed by smokers between 1992 and 2002 (Al-Delaimy et al.,
2007; Messer et al., 2007).
There is, however, a paucity of research adopting a withinparticipants design that might contribute to an understanding of the
process by which social policy interventions might impact upon
personal behavior. While it might be speculated that policy change
impacts upon behavior directly, for example, by making it too
expensive or too difficult to find opportunities to smoke, such
behavior change might be described as dissonant if it is not also
associated with attitudinal change. However, public policy might
also impact upon personal beliefs and attitudes that motivate and
facilitate desired behavior change. The present 5-month prospective investigation sought to investigate social– cognitive processes
associated with behavior change during the introduction of a
smoke-free public places policy in England.
A custom loathsome to the eye, harmful to the brain, dangerous to the
lungs, and in the black stinking fume thereof, nearest resembling the
horrible Stygian smoke of the pit that is bottomless.—James I of
England, Counterblaste to Tobacco (1604)
The introduction of a total ban on smoking in all public places
in England in July 2007 follows similar government action in other
countries (e.g., California, in the United States, in 1994; Republic
of Ireland, 2004; Spain, 2006; and France, 2007) and in Scotland
(2006) and Wales (2007) within the United Kingdom. There is
considerable evidence that cigarette smoking will cause the premature death of approximately half of those who continue to
smoke (e.g., U.S. Department of Health & Human Services, 2004)
and that incidence of smoking-related diseases varies with consumption level and smoking duration. There is also evidence that
smokers who quit will avoid a proportion of excess mortality risk
associated with continuing to smoke (Doll, Peto, Boreham, et al.,
Sheina Orbell, Patrick Lidierth, Caroline J. Henderson, Nicolas
Geeraert, Claudia Uller, Ayse K. Uskul, and Maria Kyriakaki, Department
of Psychology, University of Essex, Colchester, Essex, United Kingdom.
We thank Christopher Barry for his role-play as “stunt customer” during
preparation for the study and Roger Grace for his assistance with the online
questionnaire.
Correspondence concerning this article should be addressed to Sheina
Orbell, Department of Psychology, University of Essex, Wivenhoe Park,
Colchester, Essex, United Kingdom CO4 6SQ. E-mail: [email protected]
Social–Cognitive Variables Associated With Health
Behavior Change
A number of important social– cognitive theories specify variables associated with development of motivation for health behav753
754
ORBELL ET AL.
ior change. One important distinction that might be drawn between
different theoretical accounts is the degree of emphasis upon
perceived illness threat associated with a current behavior (Orbell,
2007). Protection-motivation theory (PMT; Rogers, 1983), for
example, conceptualizes the development of protection motivation
(usually operationalized in terms of behavioral intention) as driven
by two processes: threat appraisal and coping appraisal. Threat
appraisal comprises perceived severity, perceived susceptibility,
and fear associated with illness. High threat appraisal is proposed
to lead to protection motivation when an individual perceives an
adaptive coping response to reduce threat. Coping appraisal comprises expectancies that changing a current behavior will lead to
positive benefits that reduce threat (response efficacy) and expectancies that behavior change will lead to negative outcomes (response costs), together with self-efficacy to enact behavior change.
Motivation for behavior change is proposed to be highest when
high threat appraisal is combined with high response efficacy, low
response costs, and high self-efficacy.
The role of threat appraisal is less explicit in other models that
focus upon the evaluation of behavior change and infer threat
reduction from behavioral outcome expectancies. Social–
cognitive theory (SCT; Bandura, 1986, 1997) proposes that outcome expectancies regarding physical, social, and self-evaluative
outcomes of behavior change combine with self-efficacy expectancies to predict goal intentions. Similarly, the theory of planned
behavior (TPB; Ajzen, 1991) proposes that behavior change is
motivated by a positive attitude toward behavior, combined with
high levels of perceived behavioral control (similar to Bandura’s
notion of self-efficacy) and subjective norm. The TPB is distinctive in including the concept of social influence. Subjective norm
refers to the individual’s perception that others who are important
to him or her would want him or her to change behavior and
captures the perception of social support for change. While the
concept of self-efficacy is critical to each of the models, recent
developments have suggested that attention should also be given to
the role of specific efficacies associated with the process of
change. The health action process approach (HAPA; Schwarzer,
2008; Schwarzer & Renner, 2000) proposes that perceived susceptibility, outcome expectancies, and self-efficacy each contribute to
the development of intention to enact behavioral change. The
HAPA further proposes that specific efficacies associated with
maintaining change should also be considered1 (Luczynska &
Schwarzer, 2003; see also Marlatt, Baer, & Quigley, 1995). Situational temptations are an important contributor to relapse among
smokers (Shiffman et al., 1996). Temptation self-efficacy (also
called coping or maintenance self-efficacy) refers to perceived
personal efficacy to overcome temptations to relapse, and recovery
self-efficacy refers to perceived personal efficacy to reinstate a
nonsmoking goal after a relapse has occurred.
While meta-analytic and narrative reviews demonstrate support
for each of the models in relation to health behavior, relatively few
studies have examined their utility in the context of smoking
cessation (Conner & Norman, 2005), and very few studies have
included more than one model (Orbell, Hagger, Brown, & Tidy,
2006; Weinstein, 1993). Inclusion of variables specified by each of
the models PMT, SCT, TPB, and HAPA in the present research
permits simultaneous evaluation of the role of threat appraisal
associated with smoking-related disease, outcome expectancies
associated with smoking cessation, social influence, and specific
efficacies associated with smoking cessation within a single study.
The Present Study
The present investigation focuses on adults living in one medium sized town in the east of England who make regular use of
pubs for the consumption of alcohol and where, prior to the ban,
they also used to smoke cigarettes. Although the smoking ban
applied to all public places, by focusing on a convenience sample
of pub-goers we sought to obtain a sample of people likely to
engage in harmful levels of smoking and/or alcohol consumption.
Moreover, since many publicans had mounted substantial opposition during the years leading up to the ban, this sample might be
expected to be aware of, and affected by the introduction of the
ban. We examine the influence of the ban on change in social–
cognitive and behavioral indices over a 5-month period.
Our first research question concerns changes in alcohol and
cigarette consumption following the introduction of the smoking
ban. We hypothesized that cigarette consumption would decline
following the ban and that, since smoking would no longer be
possible in pubs, that pub visits and alcohol consumption among
smokers would also demonstrate decline over time. Our second
question concerned the ability of variables specified by PMT,
SCT, TPB, HAPA and a combined model to explain intentions not
to smoke in future, and predict smoking behavior change over time
among current and former smokers. The most stringent test of the
effects of the ban on social– cognitive and behavioral change is
provided by an analysis in which change in social– cognitive
variables over time is related to change in behavior over time. If
the ban has resulted in behavior change that is not merely the result
of environmental restriction, we hypothesized that we would observe motivational changes, such as increased threat perception,
positive outcome expectancy and subjective norms associated with
not smoking. Since the ban might also have a facilitating effect by
removing some temptations to smoke from the environment, we
also hypothesized an increase in personal efficacy. Our final aim
was to examine whether change in social– cognitive variables is
capable of explaining significant variance in change in smoking
over time.
Method
Participants and Procedure
Recruitment took place during May 2007 (t1). Participants were
recruited by two methods. The primary method was via pubs. After
obtaining the permission of the landlord in each establishment
(n ⫽ 29; 1 landlord refused permission), customers were approached in person by one of the researchers and invited to take
part in a study about the smoking ban by completing a question1
Recent operationalizations of HAPA have additionally suggested inclusion of a measure of planning. The present study included a question:
“At this point in time have you committed yourself to when you will give
up smoking?” A total of 9 t1 current smokers answered yes to this question,
usually specifying the date of the ban’s introduction. However, regression
of t2 smoking on t1 smoking and planning resulted in a nonsignificant
value of beta (.09, ns), Fchange(1, 133) ⫽ 1.62, ns.
IMPACT OF SMOKING BAN IN ENGLAND
naire now and another after the ban came into force. Those who
agreed were then provided with a questionnaire and a pen and were
invited to complete the questionnaire then and there. If requested,
participants were given a license paid envelope in which to return
the completed questionnaire at a later date. The researchers remained in the bar and provided literacy assistance if required. In
addition, an e-mail message was sent to all staff at the University
of Essex inviting people “who used public houses” to request a
questionnaire, which was then posted to them, or to complete an
online version of the questionnaire. All participants were requested
to provide a name and an address at which they could be contacted
in the autumn after the ban. By means of incentive, three £100 cash
lottery prizes were offered to those who completed both questionnaires. The University of Essex ethics committee provided ethical
approval.
Approximately 50% of all questionnaires distributed were subsequently returned completed, a total sample of 583. Of these,
68.8% were completed in the pub, 27.3% were received in the
post, and 3.9% were completed online. At follow-up, in October
2007 (t2), participants were contacted by post according to the
name and address provided at t1 or by e-mail in the case of those
who had completed the questionnaire online. If no reply was
received within 2 weeks, a reminder letter and a second copy of the
questionnaire were sent. As might be expected in a study in which
detailed personal lifestyle information is requested, 23 participants
returned completed questionnaires but omitted their name or address, and 40 t2 questionnaires were returned undelivered. Thus,
520 out of 583 were contacted at t2, and 272 of those returned a
completed questionnaire, a response rate of 52%.
Questionnaire Measures
The questionnaire took the form of an A5 paper size booklet. All
behavioral and social– cognitive measures were identical at both
time points of measurement. People who had never smoked completed sections relating to demographic information (age, gender,
occupational status, and relationship status) and alcohol use only.
Smokers and ex-smokers completed all measures. All social–
cognitive measures were operationalized with respect to “me not
smoking in the future” and utilized 6-point scales unless specified
otherwise.
Alcohol use. Alcohol consumption was derived from a series
of items concerning frequency and quantity of alcohol consumed
in the pub and at home. Participants were asked, “How many times
a week/month do you typically visit the pub (drink alcohol at
home)?” They were then asked, “During a typical visit to the pub
(typical ‘drinking session’ at home), what do you usually drink?”
A box was provided in which participants were asked to specify up
to three types of alcoholic drink, and the size and number of each
drink. This information was transformed into a number of alcohol
units, (mL ⫻ ABV)/1000, consumed per session and multiplied by
frequency to compute measures of alcohol units per week consumed in the pub, and alcohol units per week consumed at home.2
Tobacco use. Participants were asked first to respond yes or no
to the question, “Have you ever smoked cigarettes or tobacco
regularly?” To assess quit status, we asked participants, “If you
have successfully quit smoking, how long has it been since you last
smoked?” Current smokers who reported smoking daily for at least
a year then indicated, “How many manufactured/tipped cigarettes
755
do you currently smoke in total on a typical day?” and “How many
grams of ‘roll your own’ tobacco do you currently smoke in total
on a typical day?” The same two questions were put regarding
smoking in the pub: “How many manufactured cigarettes/g of ‘roll
your own’ tobacco do you currently smoke in total on a typical
visit to a public house?”3 Participants were also asked to indicate
whether they currently smoked in their place of residence (no,
never; yes, but only outside, e.g., porch or garden; or yes, I smoke
indoors and outside).
Social Cognitive Measures
Five items were used to assess intention not to smoke (e.g., “I
will try to not smoke in the future,” “I plan to not smoke in the
future,” and “I intend to not smoke in the future”; strongly
disagree–strongly agree; ␣t1 ⫽ .96, ␣t2 ⫽ .97). Three threat
appraisals were included. Perceived severity of smoking-related
disease (PMT) was operationalized by 4 items referring to psychosocial severity (e.g., “developing a smoking related disease
would put my financial security at risk/affect important relationships in my life/stop me living my life the way I want to”; strongly
disagree–strongly agree; ␣t1 ⫽ .83, ␣t2 ⫽ .84). Perceived susceptibility (PMT and HAPA) comprised 4 items (e.g., “the chances of
me developing a smoking related disease/dying young because of
smoking/becoming disabled (unable to walk long distances) because of smoking are high”; strongly disagree–strongly agree;
␣t1 ⫽ .91, ␣t2 ⫽ .96). Four items assessed fear (PMT) of smokingrelated disease (“The thought of me developing a smoking related
disease makes me feel”: anxious–not at all anxious; afraid–not at
all afraid; scared–not at all scared; worried–not at all worried;
␣t1 ⫽ .88, ␣t2 ⫽ .90). Three of the models included outcome
expectancies associated with adopting the intended behavior
(PMT, SCT, and HAPA). Positive outcome expectancies (referred
to as response benefits in PMT) comprised 9 items referring to
positive physical, social, and self-evaluative outcomes of not
smoking derived from a pilot study (e.g., “Not smoking means I
have more stamina and energy/I have reduced my chances of
serious illness such as heart disease/I am more attractive to other
people/I save a lot of money/I feel satisfied with myself”; very
unlikely–very likely; ␣t1 ⫽ .86, ␣t2 ⫽ .88). Negative outcome
expectancies (referred to as response costs in PMT) comprised 5
negative outcomes of not smoking (“Not smoking means I don’t
know what to do with my hands/my smoking friends don’t like me
any more/I put on weight/I feel more stressed/I am irritable toward
other people”; very unlikely–very likely; ␣t1 ⫽ .70, ␣t2 ⫽ .71).
2
Beers and ciders are sold in pint measures in England, and spirits and
wines are sold in milliliter measures. Named brands of “strong lagers” were
coded as 5.0 ABV so that 1 pint ⫽ 2.8 units. Lager or beer was coded as
3.5 ABV so that 1 pint ⫽ 2.3 units. Wine was assumed to be 12.0 ABV and
a wine glass 175 mL unless specified otherwise, so that 1 glass ⫽ 2.1 units
and 1 bottle ⫽ 9 units. Spirits were assumed to be 40.0 ABV and a
measure 25 mL so that 1 “single” ⫽ 1 unit. If the pub frequented by a
participant sold spirits in 35 mL measures, 1 “single” was 1.4 units.
3
In order to transform “roll your own” tobacco into a measure of
number of cigarettes smoked per day for subsequent analyses, a subset of
participants (n ⫽ 100) were asked to specify number of grams of tobacco
and number of “roll your own” cigarettes smoked during a typical pub visit.
Division of these measures indicated a median (that was equivalent to the
mode) of 0.5 g per cigarette.
756
ORBELL ET AL.
Each of the models PMT, SCT, HAPA, and TPB specifies a
measure of self-efficacy to perform an intended behavior (referred
to as perceived behavioral control in TPB). General self-efficacy
to not smoke in the future comprised 7 items (e.g., “I am confident
that I will not smoke if I don’t want to,” “Not smoking is under my
control,” “It is difficult for me not to smoke,” “Not smoking is
something I can do”; strongly disagree–strongly agree; ␣t1 ⫽ .78,
␣t2 ⫽ .80). The HAPA additionally specifies efficacy judgments
associated with resisting temptations to smoke during the attempt
to quit and efficacy judgments associated with recovery, or getting
back on track if resistance fails during a quit attempt. Temptation
self-efficacy was operationalized by 9 items representing typical
temptation situations (Shiffman et al., 1996). Each item began with
the stem “I am confident that I can manage not to smoke if I don’t
want to. . .(e.g., “even if I am drinking alcohol,” “when someone
else smokes near to me,” “when someone offers me a cigarette,”
“when I am tired,” “when I am angry about something”; strongly
disagree–strongly agree; ␣t1 ⫽ .97, ␣t2 ⫽ .97). Recovery selfefficacy comprised 4 items concerning relapse incidents (e.g., “If
you had a ‘drag’ on someone else’s cigarette when you had
stopped smoking, how confident are you that you could tell yourself it was a ‘one-off’ and stop again?” “If you smoked in a
stressful situation when you had stopped smoking, how confident
are you that you could tell yourself it was a ‘one-off’ and stop the
next day?”; extremely unconfident– extremely confident; ␣t1 ⫽
.97, ␣t2 ⫽ .97).
The direct form of measurement (Ajzen & Fishbein, 1980) was
used to operationalize attitude and subjective norm from the TPB.
Attitude comprised bipolar scales with the stem “For me to not
smoke in the future would be” (harmful— beneficial, importantunimportant, pleasant– unpleasant, good– bad, necessary–
unnecessary, desirable– undesirable; ␣t1 ⫽ .86, ␣t2 ⫽ .87). Subjective norm was operationalized with respect to two different
referent groups. Four items assessed subjective norm associated
with the “generalized other” (“Most people who are important to
me would encourage me not to smoke/approve of me not smoking/
support me in not smoking/want me not to smoke”; strongly
disagree–strongly agree; ␣t1 ⫽ .94, ␣t2 ⫽ .96). Participants were
also asked to complete the same 4 items with respect to “most of
my friends that I drink with in the pub.” These 4 items comprised
our measure of subjective norm (friends), ␣t1 ⫽ .91, ␣t2 ⫽ .94.
Analytic Strategy
Our first question concerned change in alcohol and smoking
across time. We used a mixed-model analysis of variance
(ANOVA) to examine these hypotheses. Second, we used a series
of linear regression analyses to test the ability of variables specified by each of the theoretical models and a combined model to
predict intentions and t2 smoking behavior.4 Because we wished to
explain intention to change and actual change in smoking behavior, we included t1 number of cigarettes smoked at the first step of
each of these analyses. Controlling for past behavior is also important in order to guard against difficulties of interpretation
arising from covariability in behavior at t1 and social– cognitive
beliefs (see Weinstein & Nicolich, 1993, for a discussion of this
problem in relation to risk perception). There were 0 –5% missing
values across social– cognitive variables, and missing values were
imputed using the expectation-maximization method in SPSS.
Finally to guard against potential multicollinearity, we computed
bivariate correlations between all study variables. None of the
correlations between predictors exceeded r ⫽ .62, suggesting that
multicollinearity was unlikely to be a problem for the present data
(Tabachnik & Fidell, 1989).
Results
Sample and Attrition Analyses
The characteristics of the samples obtained at t1 and t2 are
displayed in Table 1. Attrition analyses, comparing people who
completed both questionnaires with those who completed the first
questionnaire only, showed that the final sample were older, more
likely to be married, less likely to be smokers, and visited the pub
less often. There were no significant differences in gender composition, occupational class, or children living at home. We consider the implications of these findings in the Discussion section.
Change in Alcohol Use
We analyzed change in number of units alcohol consumed per
week before and after the ban (see Figure 1). We conducted a 2
(group: smokers vs. non/ex-smokers) ⫻2 (location of alcohol
consumption: pub vs. home) ⫻2 (time: t1 vs. t2) mixed-model
ANOVA, with repeated measures on the latter two factors. All
three main effects were significant. First, smokers consumed more
units (M ⫽ 23.49, SD ⫽ 1.62) than did nonsmokers or ex-smokers
(M ⫽ 13.39, SD ⫽ .95), F(1, 270) ⫽ 29.06, p ⬍ .001. Second,
more alcohol units were consumed at the pub (M ⫽ 21.66, SD ⫽
1.54) than were at home (M ⫽ 15.22, SD ⫽ 1.18), F(1, 270) ⫽
10.36, p ⬍ .001. Finally, consumption of alcohol units was higher
at t1 (M ⫽ 19.89, SD ⫽ 1.11) than at t2 (M ⫽ 16.99, SD ⫽ 1.02),
F(1, 270) ⫽ 8.02, p ⬍ .01. Furthermore, the analysis revealed a
two-way interaction between group and location, F(1, 270) ⫽
5.59, p ⬍ .05. However, these effects were qualified by a threeway interaction between group, location, and time, F(1, 270) ⫽
5.23, p ⬍ .05. To investigate the nature of this interaction, we
conducted two separate 2 (group) ⫻ 2 (time) mixed ANOVAs on
alcohol consumption at the pub and at home.
The ANOVA on alcohol consumption at the pub (Figure 1, left
panel) revealed a main effect of group, F(1, 270) ⫽, p ⬍ 23.10, but
this time the effect of time was not significant, F(1, 270) ⫽ 1.83,
p ⫽ .17, ␩2p ⫽ .08. More important, the interaction between group
and location was significant, F(1, 270) ⫽ 5.98, p ⬍ .05, ␩2p ⫽ .02.
Simple main effects revealed that smokers at t2 consumed fewer
alcohol units at the pub than they did at t1 ( p ⬍ .05). In contrast,
nonsmokers and ex-smokers at t2 consumed slightly more alcohol
units at the pub in comparison with t1, although this result was not
statistically reliable ( p ⫽ .28). The ANOVA on alcohol consumption at home (Figure 1, right panel) revealed both a main effect of
group F(1, 270) ⫽ 11.12, p ⬍ .001, ␩2p ⫽ .04, and a main effect
of time F(1, 270) ⫽ 5.22, p ⬍ .05, ␩2p ⫽ .02, but the interaction of
4
Because it is important to maintain an adequate participant-to-variable
ratio in multiple linear regression, preliminary analysis of change in social–
cognitive variables across time was conducted so that only variables
showing change between t1 and t2 were included in the final analysis,
relating change in cognition to change in behavior over time.
IMPACT OF SMOKING BAN IN ENGLAND
757
Table 1
Sample Characteristics and Attrition Analyses (Comparison of t2 Sample With Those Who
Completed t1 Only)
Variable
t1 sample (N ⫽ 583)
t2 sample (N ⫽ 272)
Attrition analysis
Mean age (SD)
Age range
% men
% married/living together
% no children at home
Occupational class
% manual labor
% administrative
% professional
% not employeda
% never smoked
Mean visits to pub per month
36.11 (13.06)
18–77
57.3
48.3
80.6
39.78 (13.66)
18–77
54.0
53.1
82.0
t(580) ⫽ ⫺6.58ⴱⴱ
26.9
39.7
17.6
15.8
39
8.75
24.4
42.3
16.9
16.1
50
7.82
a
ⴱ
␹2(1) ⫽ 28.61ⴱⴱ
t(580) ⫽ .26ⴱ
Not employed includes unemployed, students, retired, and homemakers.
p ⬍ .05. ⴱⴱ p ⬍ .01.
group and location was not significant (F ⬍ 1). In summary,
smokers showed a decrease in alcohol consumption at the pub, but
this was not compensated with an increase in consumption at
home.
Findings regarding alcohol consumption suggest a decline in
frequency of pub visits among smokers and an increase among
nonsmokers and ex-smokers. A 2 (group: smokers vs. non/exsmokers) ⫻2 (time: t1 vs. t2) mixed-model ANOVA, with repeated measures on the second factor, revealed a significant main
effect of group, F(1, 270) ⫽ 14.05, p ⬍ .01, a nonsignificant main
effect of time, F(1, 270) ⫽ .22, ns, and the anticipated two-way
interaction between group and time, F(1, 270) ⫽ 8.29, p ⬍ .01,
␩2p ⫽ .03. Nonsmokers and ex-smokers increased their frequency
of pub visits from a mean of 6.52 times per month (SD ⫽ 7.68) at
t1 to 7.41 (SD ⫽ 6.84) at t2, F(1, 201) ⫽ 5.85, p ⬍ .05, while
smokers reduced their frequency of pub visits from 11.30 (SD ⫽
9.08) times per month at t1 to 10.07 (SD ⫽ 7.87) at t2, F(1, 69) ⫽
3.48, p ⫽ .07.
Change in Tobacco Use
One hundred thirty-six people who completed the questionnaire
at both time points had smoked cigarettes for at least a year during
35
t1
30
t2
25
Alcohol units
ns
␹2(1) ⫽ 13.18ⴱⴱ
ns
ns
20
15
10
5
their lifetime. Of these, 70 (52%) reported smoking daily at t1,
while 14 (11%) reported that they had given up smoking within the
past 6 months, and 52 (38%) reported having given up smoking
more than 6 months previously. At t2, 6 ex-smokers had relapsed,
and 5 people who were smoking at t1 had quit smoking, so that a
net total of 13 people who had been smoking within 6 months of
the introduction of the ban (n ⫽ 84) had quit smoking at t2
follow-up (15.5%). For t1 smokers and ex-smokers considered
together (n ⫽ 136), mean number of cigarettes smoked per day
reduced from 8.31 (SD ⫽ 11.27) to 6.94 (SD ⫽ 9.10), F(1, 134) ⫽
5.89, p ⬍ .05, ␩2p ⫽ .04; mean change (t2⫺t1) ranged from ⫺40
to ⫹20 cigarettes per day. Considering only those who were
current smokers at t1 (n ⫽ 70), mean number of cigarettes smoked
per day was reduced from 16.14 (SD ⫽ 10.97) to 12.75 (SD ⫽
9.05), t(69) ⫽ 3.44, p ⬍ .01; mean change (t2⫺t1) ranged from
⫺40 to ⫹10 cigarettes per day. Thus the introduction of the ban
was associated with a net decrease in daily cigarette consumption.
We also examined change in smoking during a typical visit to
the pub. It should be recognized that although smoking was no
longer permitted inside pubs at t2, smokers might continue to
smoke, either by sitting outdoors or going out for a cigarette from
time to time. Considering t1 smokers and ex-smokers together, we
found that the mean number of cigarettes smoked on a typical visit
declined from 4.64 (SD ⫽ 6.36) to 3.05 (SD ⫽ 4.36), F(1, 135) ⫽
13.93, p ⬍ .01, ␩2p ⫽ .09. Considering only current smokers at t1,
we found that the mean number of cigarettes smoked during a pub
visit was reduced from 9.01 (SD ⫽ 6.26) to 5.69 (SD ⫽ 4.68), F(1,
69) ⫽ 18.62, p ⬍ .01, ␩2p ⫽ .21.
Interestingly, at t1, only 58.6% of current smokers reported that
they smoked indoors at home, while 31.4% smoked outside their
homes, and 10% reported never smoking at home. These proportions did not change over time.
Regression of Intention Not to Smoke
0
Smokers
Non-smokers
Alcohol consumption in pubs
Smokers
Non-smokers
Alcohol consumption at home
Figure 1. Change in units of alcohol consumed per week in the pub and
at home over time among smokers (n ⫽ 70) and non/ex-smokers (n ⫽ 202).
Bivariate correlations between study variables are displayed in
Table 2. A summary of the regression of intention is presented in
Table 3. At the first step, t1 number of cigarettes smoked per week
explained 18% of variance in intention not to smoke. The inclusion
of variables specified by each of the models contributed significant
ORBELL ET AL.
758
Table 2
Bivariate Correlations Between t1 Social–Cognitive Measures and Cigarette Smoking (N ⫽ 136)
Variable
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
ⴱ
1
t1 cigarettes per day
t2 cigarettes per day
Intention
Attitude
Subjective norm
(general)
Subjective norm
(friends in pub)
Severity
Susceptibility
Fear
Negative outcome
expectancy
Positive outcome
expectancy
General self-efficacy
Temptation self-efficacy
Recovery self-efficacy
p ⬍ .05.
ⴱⴱ
2
3
ⴱⴱ
4
ⴱⴱ
5
6
7
ⴱⴱ
—
ⴱⴱ
8
ⴱⴱ
9
ⴱⴱ
10
ⴱⴱ
11
ⴱ
12
13
14
⫺.42
⫺.22
⫺.23
⫺.32
⫺.24
.31
⫺.30
.19
⫺.41
⫺.50
⫺.46
⫺.47ⴱⴱ
⫺.46ⴱⴱ ⫺.21ⴱ ⫺.22ⴱⴱ ⫺.38ⴱⴱ ⫺.27ⴱⴱ
.29ⴱⴱ ⫺.37ⴱⴱ
.30ⴱⴱ ⫺.49ⴱⴱ ⫺.18ⴱ ⫺.41ⴱⴱ ⫺.42ⴱⴱ
—
.31ⴱⴱ
.44ⴱⴱ
.41ⴱⴱ
.44ⴱⴱ ⫺.10
.33ⴱⴱ ⫺.10
.58ⴱⴱ
.39ⴱⴱ
.45ⴱⴱ
.37ⴱⴱ
—
.40ⴱⴱ
.30ⴱⴱ
.35ⴱⴱ
.01
.30ⴱⴱ ⫺.25ⴱⴱ
.34ⴱⴱ
.30ⴱⴱ
.34ⴱⴱ
.19ⴱ
— .81
—
ⴱⴱ
.62ⴱⴱ
.44ⴱⴱ ⫺.10
.21ⴱ
—
.30ⴱⴱ ⫺.10
—
.08
—
.43ⴱⴱ ⫺.10
.31ⴱⴱ ⫺.10
.07
.16
—
⫺.10
⫺.10
—
ⴱⴱ
ⴱⴱ
ⴱⴱ
.47ⴱⴱ
.26ⴱⴱ
.30ⴱⴱ
.28ⴱ
.59ⴱⴱ
.32ⴱⴱ
.21ⴱ
.27ⴱ
.39ⴱⴱ
.24ⴱⴱ
.33ⴱⴱ
.14
⫺.10
⫺.33ⴱⴱ ⫺.27ⴱⴱ ⫺.34ⴱⴱ
.58ⴱⴱ
.31ⴱⴱ
.21ⴱ
.12
⫺.10
—
⫺.37ⴱⴱ ⫺.38ⴱⴱ ⫺.27ⴱⴱ
.43ⴱⴱ
—
.37ⴱⴱ
.61ⴱⴱ
—
.41ⴱⴱ
.45ⴱⴱ
.36ⴱⴱ
—
p ⬍ .01.
additional explained variance ranging from 16% (TPB), to 21%
(SCT), to 24% (HAPA), to 26% (PMT). The last column of Table
3 shows the test of a combined model. One threat appraisal
variable, perceived severity, and positive outcome expectancies, or
response efficacy, obtained significant beta values, and the model
explained 29% of variance in intention not to smoke after accounting for t1 smoking. Thus, at t1, intention to not smoke in future
was associated with perceiving serious consequences associated
with smoking and perceiving that these consequences might be
alleviated by not smoking.
Regression of t2 Smoking
Table 4 summarizes the regression of t2 smoking on t1 smoking
and social– cognitive variables. T1 smoking accounted for 66% of
t2 smoking at the first step, and each of the social– cognitive
Table 3
Regression of t1 Intention to “Not Smoke” After the Ban
Variable entered
a
t1 cigarettes per day
Attitude
Subjective norm (general)
Subjective norm (friends in pub)
Severity
Susceptibility
Fear
Negative outcome expectancyb
Positive outcome expectancyc
General self-efficacy
Temptation self-efficacy
Recovery self-efficacy
dfs
F(model)
R2
PMT ␤
ⴱ
⫺.16
SCT ␤
ⴱ
⫺.18
.23ⴱⴱ
⫺.03
⫺.07
.00
.43ⴱⴱ
.08
.00
.47ⴱⴱ
.10
7, 136
14.06ⴱⴱ
.44
4, 136
2.89ⴱⴱ
.39
HAPA ␤
⫺.13
TPB ␤
ⴱⴱ
⫺.24
.08
.25ⴱ
.10
.02
.05
.43ⴱⴱ
⫺.01
.24ⴱⴱ
.08
7, 136
13.48ⴱⴱ
.42
.15
5, 136
13.25ⴱⴱ
.34
Combined ␤
⫺.12
.01
.12
.00
.16ⴱ
⫺.01
⫺.02
.04
.34ⴱⴱ
.00
.17ⴱⴱⴱ
.07
12, 135
8.94ⴱⴱ
.47
Note. PMT ⫽ protection-motivation theory; SCT ⫽ social– cognitive theory; HAPA ⫽ health action process
approach; TPB ⫽ theory of planned behavior.
a
The t1 cigarettes accounted for 18% variance in intention “not to smoke” before the addition of social–
cognitive variables to the models, F(1, 134) ⫽ 29.26, p ⬍ .01. The addition of social– cognitive variables
contributed significant additional variance to each of the models: protection motivation theory, Fchange(6, 128) ⫽
9.64, p ⬍ .01; social– cognitive theory, Fchange(3, 131) ⫽ 15.04, p ⬍ .01; health action process account, Fchange(6,
128) ⫽ 9.09, p ⬍ .01; theory of planned behavior, Fchange(4, 130) ⫽ 7.77, p ⬍ .01; and the combined model,
Fchange(11, 123) ⫽ 6.00, p ⬍ .01. b Referred to as response costs by protection motivation theory. c Referred
to as response efficacy by protection motivation theory.
ⴱ
p ⬍ .05. ⴱⴱ p ⬍ .01. ⴱⴱⴱ p ⫽ .063.
IMPACT OF SMOKING BAN IN ENGLAND
759
Table 4
Regression of t2 Smoking Behavior
PMT
Variable entered
t1 cigarettes per daya
Attitude
Subjective norm (general)
Subjective norm (friends in pub)
Severity
Susceptibility
Fear
Negative outcome expectancyb
Positive outcome expectancyc
General self-efficacy
Temptation self-efficacy
Recovery self-efficacy
Intention to “not smoke”
dfs
F(model)
R2 (model)
SCT
HAPA
TPB
␤
␤
␤
␤
␤
␤
.69ⴱⴱ
.68ⴱⴱ
.70ⴱⴱ
.70ⴱⴱ
.72ⴱⴱ
.71ⴱⴱ
.05
.05
ⴱⴱ
ⴱⴱ
⫺.03
.05
⫺.05
.14ⴱ
⫺.15ⴱ
.00
⫺.01
.05
⫺.06
.14ⴱ
⫺.13
.01
.14
⫺.18ⴱⴱ
⫺.12
.14
⫺.16ⴱⴱ
⫺.01
—
7, 128
45.11ⴱⴱ
.71
⫺.05
8, 127
39.41ⴱⴱ
.71
—
4, 131
79.42ⴱⴱ
.71
⫺.15
5, 130
63.50ⴱⴱ
.71
ⴱⴱ
.16
⫺.21ⴱⴱ
⫺.04
.07
.06
—
7, 128
45.62ⴱⴱ
.71
.16ⴱⴱ
⫺.16ⴱⴱ
⫺.04
.09
.06
⫺.07
8, 127
4.20ⴱⴱ
.72
Combined
␤
␤
␤
␤
.73ⴱⴱ
.00
.07
⫺.16ⴱ
.70ⴱⴱ
.00
.10
⫺.14ⴱ
⫺.10
⫺.08
—
5, 130
55.78ⴱⴱ
.68
⫺.12ⴱⴱⴱ
6, 129
48.16ⴱⴱ
.69
.70ⴱⴱ
.04
.09
⫺.07
⫺.07
.06
⫺.03
.16ⴱⴱ
⫺.18ⴱ
⫺.03
.06
.04
—
12, 123
26.69ⴱⴱ
.72
.70ⴱⴱ
.04
.10
⫺.07
⫺.05
.06
⫺.03
.16ⴱⴱ
⫺.15ⴱⴱⴱⴱ
⫺.03
.07
.05
⫺.08
13, 122
24.84ⴱⴱ
.73
Note. PMT ⫽ protection-motivation theory; SCT ⫽ social– cognitive theory; HAPA ⫽ health action process approach; TPB ⫽ theory of planned
behavior. Figures in the top half of the table are betas. Figures in the lower portion of the table are as indicated in the left-hand column of the table.
a
The t1 cigarettes accounted for 66% variance in t2 cigarettes per day before the addition of social– cognitive variables to the models, F(1, 134) ⫽ 256.97,
p ⬍ .01. The addition of social– cognitive variables contributed significant additional variance to each of the models: protection motivation theory, Fchange(6,
128) ⫽ 4.02, p ⬍ .01; social– cognitive theory, Fchange(3, 131) ⫽ 7.59, p ⬍ .01; health action process account, Fchange(6, 128) ⫽ 4.22, p ⬍ .01; theory of
planned behavior, Fchange(4, 130) ⫽ 2.54, p ⬍ .05; and the combined model, Fchange(11, 123) ⫽ 2.63, p ⬍ .01. b Referred to as response costs by
protection motivation theory. c Referred to as response efficacy by protection motivation theory.
ⴱ
p ⬍ .05. ⴱⴱ p ⬍ .01. ⴱⴱⴱ p ⫽ .051. ⴱⴱⴱⴱ p ⫽ .056.
models contributed from 2% (TPB) to 5% (PMT, SCT, and
HAPA) additional explained variance. In the combined model,
shown in the far right-hand column of Table 4, negative outcome
expectancy, or the perceived costs associated with trying not to
smoke, was associated with an increase in smoking over time, and
positive outcome expectancy, or response efficacy, was associated
with a decrease in smoking over time. The inclusion of intention
not to smoke did not obtain a significant beta value, suggesting that
the effect of social– cognitive variables on smoking at t2 was not
mediated by intention.
It is also of interest to note that in the test of the theory of
planned behavior, the subjective norm concerning “my friends in
the pub” was inversely associated with smoking at t2, suggesting
that the social influence of pub-going friends who did not encourage smoking contributed to a decline in smoking over time. In the
combined model, the subjective norm (friends) did not obtain a
significant independent beta value, possibly because positive outcome expectancies included social evaluative outcomes.
Relationship of Change in Social–Cognitive Variables to
Change in Smoking
While the regression of t2 behavior suggests that behavior change
may be associated with negative and positive expectancies, a more
stringent test might be afforded by investigating the ability of change
in social– cognitive variables to account for variance in change in
behavior. A preliminary mixed 2 (t1 smoker vs. ex-smoker) ⫻2 (t1 vs.
t2) ANOVA, in which social– cognitive variables assessed at two time
points were treated as within- participants variables, showed significant change across time for three variables. A significant main effect
of time, F(1, 134) ⫽ 6.43, p ⬍ .05, ␩2p ⫽ .05, showed that perceived
severity of illness associated with smoking increased for both smokers
and ex-smokers between the two time points (smokers: t1 M ⫽ 4.19,
SD ⫽ 1.36, t2 M ⫽ 4.47, SD ⫽ 1.25; ex-smokers: t1 M ⫽ 5.04, SD ⫽
1.03, t2 M ⫽ 5.22, SD ⫽ .97). This effect was not moderated by
smoking status. A significant interaction of Smoking Status ⫻Time
was obtained for subjective norm (people in general), F(1, 134) ⫽
4.62, p ⬍ .05, ␩2p ⫽ .03, such that t1 smokers showed an increase in
subjective norm for not smoking over time (t1 M ⫽ 4.62, SD ⫽ 1.57,
t2 M ⫽ 5.00, SD ⫽ 1.10), t(69) ⫽ ⫺2.05, p ⬍ .05, and ex-smokers
subjective norm did not change over time (t1 M ⫽ 5.57, SD ⫽ 1.01,
t2 M ⫽ 5.46, SD ⫽ 1.11), t(65) ⫽ .85, ns. A significant interaction of
t1 Smoking Status ⫻ Time was also obtained for negative outcome
expectancies associated with not smoking (response costs), F(1,
134) ⫽ 6.18, p ⬍ .05, ␩2p ⫽ .04, such that while t1 ex-smokers
experienced a decrease in perceived costs (t1 M ⫽ 2.53, SD ⫽ .98, t2
M ⫽ 2.37, SD ⫽ .91), t(65) ⫽ 1.87, p ⫽ .07, t1 smokers experienced
an increase in perceived costs over time (t1 M ⫽ 3.08, SD ⫽ 1.13, t2
M ⫽ 3.35, SD ⫽ .97), t(69) ⫽ ⫺1.82, p ⫽ .073.
In order to assess the contribution of changes in cognition to
changes in smoking over time, a three-step linear multiple
regression of t2 smoking was conducted in which t1 smoking
was entered at the first step; t1 severity, subjective norm, and
negative outcome expectancies were entered at the second step;
followed by t2 severity, subjective norm, and negative outcome
expectancies at the third step (see Table 5). Inspection of the
right-hand column of Table 5 shows that change in negative
outcome expectancies (response costs) over time was positively
associated with change in smoking across time, while change in
subjective norm across time was inversely associated with
change in smoking.
ORBELL ET AL.
760
Table 5
Regression of t2 Smoking on t1 Smoking and t1 and t2 Social–
Cognitive Variables
Variable entered
Step 1
t1 cigarettes per day
Step 2
t1 severity
t1 negative outcome expectancy
t1 subjective norm (general)
Step 3
t2 severity
t2 negative outcome expectancy
t2 subjective norm (general)
Model R2
Model F
Model df
F change
Step df
␤
␤
␤
.81ⴱⴱ
.77ⴱⴱ
.72ⴱⴱ
.66
256.97ⴱⴱ
1, 134
⫺.08
.15ⴱⴱ
.01
⫺.05
.09
.08
.69
71.11ⴱⴱ
4, 131
3.79ⴱ
3, 131
⫺.06
.13ⴱ
⫺.11ⴱ
.71
43.85ⴱⴱ
7, 128
3.06ⴱ
3, 128
Note. Figures in the top half of the table are betas. Figures in the lower
portion of the table are as indicated in the left-hand column of the table.
ⴱ
p ⬍ .05. ⴱⴱ p ⬍ .01.
Discussion
The consideration that alcohol use increased among non/exsmokers suggests that the decline we observed in smokers cannot
be attributed merely to seasonality effects but is more likely
attributable directly to the impact of the smoke-free policy. It may
have been anticipated that a reduction in pub drinking would be
compensated for by home consumption. Home alcohol consumption declined between May and October in both smokers and
non/ex-smokers, a finding that may perhaps be attributed to a
seasonality effect. It is possible that warmer weather or major TV
sporting events determine levels of home alcohol consumption in
early summer that decline into autumn, irrespective of smoking
status. Smokers may not have increased home alcohol consumption after the ban because the home environment does not provide
a meaningful alternative to visiting a pub to smoke, socialize, and
drink alcohol. Consistent with this possibility, 41% of regular
smokers in the present study reported that they do not smoke inside
their homes, and this proportion did not increase after the ban.
A total of 15.5% of people who were smoking regularly within
6 months before the July ban reported that they had quit by
October follow-up. This relatively high quit rate should be interpreted with some caution because many former smokers relapse
within a year. Indeed, only half of those who quit during the 6
months leading up to the ban were smoke-free at follow-up. We
also observed a significant decline in number of cigarettes smoked
during pub visits and in the total number of cigarettes smoked per
day among smokers, from 16.14 to 12.75. In summary, results
suggest that the introduction of the smoke-free policy has had
positive effects in reducing levels of harmful smoking and drinking in smokers. While it should be acknowledged that we cannot
rule out the possibility that these changes might have occurred in
the absence of the ban, the present findings are consistent with
those from other sources reporting a decline in smoking prevalence
in England during 2007–2008 that is greater than that observed in
any previous year (West, 2008).
Our main study questions related to social– cognitive processes
associated with change in smoking. After controlling for t1 smoking, higher perceived severity and positive outcome expectancy
were independently associated with intention not to smoke in
future. More important, in the prediction of t2 smoking, both
positive and negative outcome expectancies were associated with
behavior change, while threat appraisal variables did not explain
significant independent variance. Negative outcome expectancies
(response costs) associated with trying not to smoke at t1 were
positively associated with t2 smoking. Anticipation that not smoking increases stress, irritability, or fidgeting predicted increases in
smoking across time, perhaps explaining relapse between the two
time points. Yong and Borland (2008) reported that people who
report smoking to functionally reduce stress are less likely to quit
and more likely to relapse. Clearly, preparation in terms of developing means of coping with urges to smoke is important in any
quit attempt.
The finding from the analysis of TPB, the perception that
“drinking” friends would discourage smoking, explained significant variance in t2 smoking is important and underlines the role of
the social context of pub smoking. Such friends may influence
reductions in smoking by supporting abstinence and limiting the
number of times a smoker might leave the social group and go
outside to smoke. Gerrard, Gibbons, Lane, and Stock (2005), for
example, have shown that social comparison with people who are
successfully not smoking has been associated with successful
cessation.
We set out to determine whether change in smoking behavior
was associated with social– cognitive change over time. Media
campaigns leading up to the introduction of the ban emphasized
the health consequences of smoking, and it is interesting to note
that both former and current smokers showed increased perceived
severity over time. Decline in number of cigarettes smoked over
time was most clearly associated with an increase in subjective
norm (general) among smokers. Subjective norm refers to a sense
of disapproval about one’s own smoking and is more important
than are perceptions of how many other people smoke (descriptive
norm) in the development of smoking-related intentions (Van den
Putte, Yzer, & Brunsting, 2005). The finding that subjective norm
showed reliable change after the introduction of the ban and that
this change explained variance in behavior change provides important evidence that behavior change was motivationally supported. Negative outcome expectancies also increased across time
among smokers and were positively associated with change in
smoking. This might be explained in terms of the consequences of
periods of abstinence since the introduction of the ban. Smokers
and those who tried to quit but relapsed might have increased their
experience, during the study period, of negative outcomes such as
irritability.
It is also of interest to consider those variables that did not
demonstrate change across time. While perceived severity of
smoking-related disease increased, there was no evidence of
change in acceptance of personal susceptibility to smoking-related
disease. We observed a reduction in negative outcome expectancy
among ex-smokers over time, suggesting that the smoke-free policy might have reduced some of the irritations formerly experienced by ex-smokers “forced” to socialize among smokers. However, it might also have been expected that the removal of some of
IMPACT OF SMOKING BAN IN ENGLAND
the temptations to smoke from the pub environment in particular
might have led to an increase in self-efficacy.
A number of aspects of the present study should be taken into
consideration before generalizing the findings. It should be acknowledged that the present data were not derived from a representative sample of all smokers in England but from a sample of
those who use public houses for the consumption of alcohol. The
consideration that 61% of all those recruited to the study were
current or former smokers is perhaps indicative of the interdependence of these lifestyle behaviors. Attrition analyses also showed
that participants in both waves of the study were older, more likely
to be married and to visit the pub less often. It is possible that the
present sample may have been more amenable to behavioral and
cognitive change. Finally, all behavioral measures were selfreported. Because we observed within-participant increases and decreases in alcohol consumption and smoking across time, it seems
unlikely that the findings can be attributed entirely to socially
desirable responding. Moreover, studies in which a subsample
have undergone cotinine testing to validate self-reports show a
similar decline in smoking in the past year (West, 2008).
Notwithstanding these limitations, the present study provides
important evidence that a smoke-free public policy has not only
positive behavioral effects that are likely to benefit health but also
the fact that these changes are accompanied by social– cognitive
changes that may help to sustain and enhance such benefits.
Publicity campaigns that reinforce perceived severity, positive
psycho-social outcomes of quitting, and subjective norm for not
smoking, and that endeavor to confront and modify beliefs about
negative outcomes and enhance self-efficacy skills, might further
enhance behavior change in this population and in similar ones.
References
Ajzen, I. (1991). The theory of planned behavior. Organizational Behavior
and Human Decision Processes, 50, 179 –211.
Ajzen, I., & Fishbein, M. (1980). Understanding attitudes and predicting
social behavior. Englewood Cliffs, NJ: Prentice Hall.
Al-Delaimy, W. K., Pierce, J. P., Messer, K., White, M. M., Trinidad,
D. R., & Gilpin, E. (2007). The California Tobacco Control Program’s
effect on adult smokers: 2. Daily cigarette consumption levels. Tobacco
Control, 16, 91–95.
Bandura, A. (1986). Social foundations of thought and action: A social
cognitive theory. Englewood Cliffs, NJ: Prentice Hall.
Bandura, A. (1997). Self-efficacy: The exercise of control. New York:
Freeman.
Conner, M., & Norman, P. (2005). Predicting health behaviour (2nd ed.).
Maidenhead, England: Open University Press.
Doll, R., Peto, R., Boreham, J., & Sutherland, I. (2004). Mortality in
relation to smoking: 50 years’observations on male British doctors.
British Medical Journal, 328, 1519 –1524.
Gerrard, M., Gibbons, F., Lane, D., & Stock, M. (2005). Smoking cessa-
761
tion: Social comparison level predicts success for adult smokers. Health
Psychology, 24, 623– 629.
Luczynska, A., & Schwarzer, R. (2003). Planning and self-efficacy in the
adoption and maintenance of breast self-examination: A longitudinal
study on self-regulatory cognitions. Psychology and Health, 18, 93–108.
Marlatt, G. A., Baer, J. S., & Quigley, L. A. (1995). Self-efficacy and
addictive behavior. In A. Bandura (Ed.), Self-efficacy in changing societies (pp. 289 –315). New York: Cambridge University Press.
Messer, K., Pierce, J. P., Zhu, S.-H., Hartman, A. M., Al-Dealaimy, W. K.,
Trinidad, D. R., et al. (2007). The California Tobacco Control Program’s
effect on adult smokers: 1. Smoking cessation. Tobacco Control, 16,
85–90.
Orbell, S. (2007). Motivational models and volitional processes in the
promotion of health behaviors. In D. C. Park & L. L. Liu (Eds.), Medical
adherence and aging (pp. 169 –200). Washington, DC: American Psychological Association.
Orbell, S., Hagger, M., Brown, V., & Tidy, J. (2006). Comparing two
theories of health behavior: A prospective study of noncompletion of
treatment following cervical cancer screening. Health Psychology, 25,
604 – 615.
Rogers, R. W. (1983). Cognitive and physiological processes in fear
appeals and attitude change: A revised theory of protection motivation.
In J. J. Cacioppo & R. E. Petty (Eds.), Social psychophysiology: A
source book (pp. 153–176). New York: Guilford Press.
Schwarzer, R. (2008). Modeling health behaviour change: The health
action process approach (HAPA). Retrieved July 15, 2008 from http://
userpage.fu-berlin.de/⬃health/hapa.htm
Schwarzer, R., & Renner, B. (2000). Social– cognitive predictors of health
behavior: Action self-efficacy and coping self-efficacy. Health Psychology, 19, 487– 495.
Shiffman, S., Guys, M., Richards, T. J., Paty, J. A., Hickcox, M., & Kassel,
J. D. (1996). Temptations to smoke after quitting: A comparison of
lapsers and maintainers. Health Psychology, 15, 455– 461.
Tabachnik, B. G., & Fidell, L. S. (1989). Using multivariate statistics. New
York: Harper Row.
U.S. Department of Health and Human Services. (2004). The health
consequences of smoking: A report of the Surgeon General, 2004.
Atlanta, GA: Author.
Van den Putte, B., Yzer, M. C., & Brunsting, S. (2005). Social influences
on smoking cessation: A comparison of the effect of six social influence
variables. Preventive Medicine, 41, 186 –193.
Weinstein, N. D. (1993). Testing four competing theories of healthprotective behavior. Health Psychology, 12, 324 –333.
Weinstein, N. D., & Nicolich, M. (1993). Correct and incorrect interpretations of correlations between risk perceptions and risk behaviors.
Health Psychology, 12, 235–245.
West, R. (2008). Key performance indicators on smoking cessation in
England: Findings from the Smoking Toolkit Study. Retrieved July 15,
2008 from http://www.smokinginengland.info
Yong, H., & Borland, R. (2008). Functional beliefs about smoking and
quitting activity among adult smokers in four countries: Findings from
the international tobacco control four-country survey. Health Psychology, 27(Suppl. 1), S216 –S223.