Content uploaded by Tanya R Berry
Author content
All content in this area was uploaded by Tanya R Berry
Content may be subject to copyright.
Am J Health Behav.
™™
™™
™ 2005;29(2):117-126
117
The Effects of Exercise Advertising on
Self-efficacy and Decisional Balance
Tanya R. Berry, PhD; Bruce L. Howe, PhD
Tanya R. Berry, Assistant Professor, Depart-
ment of Kinesiology and Physical Education,
Wilfrid Laurier University Waterloo, Ontario,
Canada. Bruce L. Howe, Professor, School of
Physical Education, University of Victoria, Victoria,
BC, Canada.
Address correspondence to Dr Berry, Depart-
ment of Kinesiology and Physical Education,
Wilfrid Laurier University Waterloo, Ontario, N2L
3C5, Canada. E-mail: tberry@wlu.ca
Objectives: To investigate the
effects of exercise advertising on
self-efficacy and decisional bal-
ance for changing exercise be-
havior. Methods: One hundred sev-
enty-four university students (fe-
males = 108; males = 66) watched
a video that contained health,
appearance, or control advertis-
ing and completed stage of change,
exercise self-efficacy, and deci-
sional balance questionnaires.
Results: There was a significant
condition by gender interaction
indicating that men in the appear-
ance condition had lower self-
efficacy than did females. Health
promotion advertising did not
have any effect. Conclusions: This
study provides evidence that the
message in appearance-based ad-
vertising that men should be mus-
cular may be detrimental.
Key words: advertising, self-effi-
cacy, decisional balance, exercise
Am J Health Behav. 2005;29(2):117-126
T
here is considerable evidence that
exercise is associated with the pre-
vention of chronic diseases such as
cardiovascular disease, obesity, type II
diabetes, colon cancer, osteoporosis, and
several psychological disorders.
1,2
Despite
this, a large proportion of North Ameri-
cans do not participate in sufficient physi-
cal activity to gain the health benefits.
2,3
One way to approach the issue of seden-
tary behavior is to promote exercise
through the media. Exercise advertising
already exists in various forms, from pub-
licly funded health promotion advertising
to advertising that uses appearance en-
hancement to sell a product. However,
little is known about the effects of such
advertising on consumers. Those who
promote exercise for health do not seem
to have much of an impact as the majority
of people still do not engage in adequate
amounts of physical activity,
4
and although
people are buying fitness products (an $8
billion US-a-year industry according to
Dishman
4
), this is not translating to in-
creased exercise participation.
Researchers have identified both
health and appearance as frequently cited
reasons for exercise.
5,6
Because of the
potential importance of appearance as a
motivator, Dishman
4
emphasized that the
focus of social marketers needs to expand
beyond the health promotion aspect to
include other possible motivators, such
as appearance and weight loss, that are
prevalent in commercial exercise adver-
tising. However, before this step is taken,
it is necessary to understand how such
messages can influence viewers. It is
possible that advertising that emphasizes
appearance may negatively influence an
individual’s self-concept,
7
making con-
sumers feel inadequate in terms of
weight, fitness level, or age
8
and subse-
quently influencing exercise participa-
tion.
As yet, little research has examined
the effects of exercise advertising on con-
Effects of Exercise
118
sumers’ attitudes and beliefs toward ex-
ercise. Of existing research, almost all
has evaluated the effects of public health
promotion campaigns with the general
finding that such advertising results in
small shifts in knowledge but no increases
in actual exercise behavior.
9,10
Very few
studies have looked at appearance-based
exercise messages. One study found that
health promotion advertising had a sig-
nificant positive effect on self-presenta-
tional beliefs for those who were already
active, whereas appearance-based adver-
tising had negative effects on
nonexercisers.
11
Other researchers found
a positive effect on exercise attitudes for
people who were low self-monitors who
read health-based persuasive exercise
messages.
12
One model that may be applied to such
research is the transtheoretical model
(TTM) of behavior change. This model
emphasizes the importance of increas-
ing knowledge of the benefits of exercise
and making people aware of the risks of
being inactive,
13
messages that are often
contained in exercise advertising. Slater
14
recommended the TTM as a theoretical
framework for examining the relation-
ship between communication interven-
tions and behavior change. The TTM iden-
tifies specific stages that individuals go
through when trying to change their pat-
terns of behavior and the processes of
change people use in the various stages.
15
Researchers have identified other fac-
tors within the model that serve as me-
diators between interventions and be-
havior including decisional balance and
self-efficacy. In the exercise domain, the
TTM has proven to be a useful model
when examining exercise behavior in
such diverse populations as high school
students, university students, employed
adults,
16
and seniors.
17
The stages of change identified by the
TTM include precontemplation, contem-
plation, preparation, action, and mainte-
nance. With exercise behavior, individu-
als who are in the precontemplation stage
are unaware that lack of exercise is prob-
lematic and have no intention of chang-
ing their behavior. In the contemplation
stage, individuals are aware that there is
a problem and are thinking of addressing
it, but have yet to commit to action. Dur-
ing the preparation phase, people start to
make small changes such as going for an
occasional walk but have not yet started
regular exercise. In the action stage,
individuals make a significant commit-
ment of time and resources and start
regular exercise. If individuals change
their behavior for up to a 6-month period
of time, they are considered to be in the
action stage. If they adhere to the behav-
ior changes for longer than 6 months,
they are considered to be in the mainte-
nance stage.
Decisional balance refers to the poten-
tial benefits (pros) and costs (cons) of
behavior change and can include approval
and instrumental gain by self and others
or disapproval and losses by self and oth-
ers.
18
In earlier stages, when no action to
change behavior has occurred, possible
costs of changing behavior outweigh pos-
sible benefits. Prochaska
13
reported strong
support for the use of pros and cons of
behavior change across 12 health behav-
iors including exercise. Marshall and
Biddle,
19
in a meta-analysis of the TTM
and exercise, conclude that the biggest
increase in evaluation of the benefits of
exercise occurs during the contempla-
tion phase; as stage increases, there are
small increases in perceived benefits,
but large decreases in the costs of change.
Self-efficacy has been well researched
in the exercise domain and has been
found to be a strong contributor to exer-
cise behavior. Exercise self-efficacy is
the belief that an individual holds in his
or her ability to achieve a given exercise
goal.
20
It has been concluded that exercise
adherence is most affected by self-regula-
tory efficacy; that is, adherence is af-
fected by the strength of the belief that
one can make oneself to exercise regu-
larly in the face of numerous obstacles.
20
Self-efficacy is developed by various means
including vicarious learning through ob-
servation of others. For example, televi-
sion advertising has been identified as a
source of observational learning that can
model exercise behavior.
14,20
However, al-
though the media have been identified as
an important source of observational
learning in smoking,
21
the influence of
exercise advertising on self-efficacy needs
to be researched. Such research is par-
ticularly necessary as Slater
14
has identi-
fied that although the modeling of behav-
iors in health promotion campaigns can
be positive, a challenge to health promot-
ers is the potentially negative norms that
exist in entertainment media.
Some research has been conducted to
Berry & Howe
Am J Health Behav.
™™
™™
™ 2005;29(2):117-126
119
examine the influence of health promo-
tion advertising on self-efficacy and deci-
sional balance. In a review of existing
literature, it was concluded that self-effi-
cacy has support as a mediator between
exercise interventions and behavior, but
that support for decisional balance as a
mediator has been mixed.
22
For example,
it has been found that using stage-
matched interventions that included tele-
vision advertising and comic strips in-
creased self-efficacy in targeted individu-
als. However, perceived costs and ben-
efits of physical activity did not change for
those participants in precontemplation
and contemplation stages.
23
This research
looked exclusively at health promotion
advertising and needs to be extended to
include appearance-based advertising.
The possibility that both health-based
and appearance-based advertising can
serve as a source of observational learn-
ing and influence self-efficacy beliefs and
decisional balance needs to be examined.
The purpose of this study was to exam-
ine the effects of health-based and ap-
pearance-based exercise advertising on
self-efficacy and decisional balance in an
undergraduate sample. Hypotheses re-
garding the effects of health-based adver-
tising were based on the findings of previ-
ous research that indicated that self-
efficacy can be enhanced by health pro-
motion advertising.
22,23
It was hypothesized
that self-efficacy would increase for par-
ticipants in the health condition. Be-
cause the purpose of health promotion
advertising is to increase awareness of
the benefits of physical activity, it was
hypothesized that the benefits (pros) of
changing exercise behavior would in-
crease for participants in the health con-
dition. Further, based on Slater’s
14
con-
tention that the norms present in the
entertainment media are a potentially
negative influence on health behavior; it
was hypothesized that the costs (cons) of
changing behavior would increase for
participants in the appearance condition.
METHOD
Participants and Procedures
Participants were 178 university stu-
dents (females = 108; males = 70) from
introductory psychology classes (total en-
rollment in the classes was 1865) who
participated for class credit. One week
prior to the study, the participation oppor-
tunity was posted on an online experi-
ment-scheduling system. Students were
able to choose from a number of experi-
ments posted by several researchers and
register in a time slot that was conve-
nient for them; when the available times
were full, the experiment was closed for
registration. The present experiment was
called “TV Watching,” and the description
on the website stated that participants
would be asked to “watch a 20-minute
video and answer some questionnaires.”
They were not told that the experiment
had specifically to do with exercise. Ethi-
cal approval was obtained from the uni-
versity, and all participants completed an
informed consent form and were given
the option to leave if they so chose.
A modified Solomon design was used so
that the possibility of pretest sensitiza-
tion could be assessed. The modified
Solomon design involved random assign-
ment to 1 of 6 groups. Although all groups
watched a video, 2 of the videos were for
the treatment conditions and as such
contained either health or appearance
advertising, and 1 of the videos was for a
control condition and contained no exer-
cise advertising. The experimental de-
sign is outlined in Table 1, including
sample sizes used in analyses after dele-
tion of outliers and missing data.
The data from 4 participants were not
used because either there were missing
data (N=1), or participants in the pretest
groups did not show up for the video and
posttest session (N=3), leaving 174 in the
final data set. One hundred ten partici-
pants (females = 69; males = 41) com-
pleted pretest questionnaires 1 week be-
fore watching 1 of the 3 videos and then
completed posttest questionnaires; 64
participants (females = 35; males = 29)
completed posttest questionnaires only.
After completion of the posttests a full
debriefing was provided. Participants were
randomly assigned to one of health, ap-
pearance, or control treatment conditions
and participated in groups of up to 5 par-
ticipants.
Materials: Questionnaires
Measures were obtained from the Can-
cer Research Prevention Centre (CRPC)
24
website and included a stages-of-change
algorithm, a self-efficacy questionnaire,
and a decisional balance questionnaire.
Stages-of-change algorithm. This
short-form stage questionnaire, devel-
oped from work by Marcus and Simkin,
25
Effects of Exercise
120
asks participants to mark 1 of 5 state-
ments that reflect 1 of the 5 stages of
behavior change. Researchers have con-
firmed its use in assessing stage of be-
havior change,
26
but conclude that it is
more useful for assessing moderate to
strenuous exercise, rather than mild ex-
ercise.
17
Self-efficacy. Self-efficacy was deter-
mined by responses to an 18-item ques-
tionnaire that consisted of a 5-point Likert
scale ranging from not at all confident (1)
to very confident (5). This 18-item scale
has 6 factors that have been validated by
the CRPC. The 6 factors with correspond-
ing alpha, as reported by the CRPC, are
negative affect (.852), excuse making
(.829), must exercise alone (.869), incon-
venient to exercise (.773), resistance from
others (.853), and bad weather (.837).
Decisional balance. Decisional bal-
ance was determined by responses to a
10-item questionnaire regarding benefits
and costs of behavior change. Each area
consisted of 5 items rated on a Likert
scale ranging from extremely important
(1) to not important (5). A low score on the
pros subscale indicates that the benefits
of behavior change are important whereas
a low score on the cons subscale indicates
that the costs of behavior change are
important. The questionnaire has been
shown to be valid with North American
adults.
24,27
Videos
The videos used in this study were
developed for previous research.
11
There
was 1 video for each of the 2 treatment
conditions of exercise for health promo-
tion and exercise for appearance, and 1
video for a control condition. Each video
contained an 18-minute television pro-
gram that portrayed aspects of Japanese
culture but made no reference to appear-
ance or exercise. In addition to the pro-
gram, each video contained 6 identical
neutral advertisements that made no
reference to exercise or appearance and
3 advertisements that were selected spe-
cifically for the 2 treatment and control
conditions. Each advertisement was 20 to
30 seconds long. The neutral and appear-
ance-based advertisements were selected
from commercial television. The appear-
ance advertisements promoted a private
fitness club, an exercise video targeting
the thighs, and a nutritional supplement
advertisement that featured an attrac-
tive female model in a fitness club who
talked about “looking like me” through
exercise with the help of the supplement.
Both men and women were featured in 2
of these advertisements, and all models
were young, attractive white people. The
exercise promotion advertisements were
obtained on request from Participation
Canada,
28
the Hillary Commission of New
Zealand,
29
and the Dairy Nutrition Coun-
cil of Alberta,
30
They featured people of all
ages, different ethnic groups, and both
genders and focused on disease reduction
and improved health through exercise.
The advertisements were shown in groups
of 3 during each of 3 breaks of the pro-
gram. Two of the advertisements during
each break were neutral, and 1 was a
treatment advertisement. They were
Table 1
An Overview of the Modified Solomon Design and Sample Size
Group Pretest Treatment Posttest
Group 1 O1 (n = 41) X1 Health Video O2 (n = 41)
Group 2 O3 (n = 36) X2 Appearance Video O4 (n = 36)
Group 3 O5 (n = 33) - Control Video O6 (n = 33)
Group 4 - X1 Health Video O7 (n = 21)
Group 5 - X2 Appearance Video O8 (n = 20)
Group 6 - - Control Video O9 (n = 19)
Berry & Howe
Am J Health Behav.
™™
™™
™ 2005;29(2):117-126
121
yoked across the 3 conditions so that the
presence and order of neutral and target
advertisements appearing in the health,
appearance, and control conditions were
matched.
Analyses
Because a modified Solomon design
was used in this experiment, analyses
followed the recommendations of Braver
and Braver
31
for the statistical treatment
of a Solomon design. As such, analyses
followed the following steps:
Step 1: A factorial ANOVA to determine
if there is an interaction between treat-
ment condition and pretest (yes or no).
Step 2: If the interaction is not signifi-
cant, then a main effects ANOVA test is
performed on the experimental condi-
tions to determine if there is a treatment
effect.
Step 3: If this test is not significant,
then an ANCOVA is performed on the data
from the participants who completed both
pretests and posttests, using pretest
scores as a covariate. As Braver and
Braver
31
state, if the test done in step 2 is
“significant, there is unqualified evidence
of the treatment effect. This test clearly
is not the most powerful available, how-
ever, and if not significant it should not be
considered conclusive evidence against
a treatment effect” (p 151). Braver and
Braver therefore suggest an ANCOVA as a
further powerful test of treatment effect,
an approach also suggested by Huck and
Sandler.
32
Step 4: If the interaction between treat-
ment condition and pretest (yes or no) is
significant, a series of tests is performed
to determine if the pretest resulted in
sensitization to the treatment.
These procedures were followed so that
in the first step, a 3 (health, appearance,
or control condition) by 2 (gender) by 2
(pretest or no pretest) ANOVA was per-
formed for each of the dependent vari-
ables: self-efficacy, pros of changing be-
havior, and cons of changing behavior.
Similarly, a 3 (condition) by 4 (stage of
change) by 2 (pretest) ANOVA was per-
formed for each of the dependent vari-
ables. Following analyses depended on
the results of step 1.
RESULTS
Data were screened for outliers, and it
was found that there was 1 participant
with a self-efficacy score greater than 3
standard deviations from the mean, 3
participants with cons scores greater
than 3 standard deviations from the mean,
and 1 participant with a pro score greater
than 3 standard deviations from the mean.
These participants were deleted from the
data set.
At posttest, 5 participants were in the
precontemplation stage, 9 in the contem-
plation stage, 40 in the preparation stage,
39 in the action stage, and 77 in the
maintenance stage. Because of the small
numbers of participants in earlier stages,
participants in precontemplation and con-
templation stages were aggregated into 1
stage. Plotnikoff et al
27
used a similar
rationale for aggregating action and main-
tenance as stages of “action” when their
sample size was low for the action stage.
Stage of behavior change was very stable
for participants who completed both pre-
test and posttest questionnaires. Only 2
participants indicated a shift in stage of
change (both moved from maintenance to
action).
For participants who completed the pre-
test, there were no significant differences
at pretest in self-efficacy for treatment
condition, F(2, 92) = 2.60, P>.05, condition
by gender, F(1, 92) = .67, P>.05, or condi-
tion by stage, F(4, 92) = .34, P>.05. Simi-
larly, there were no significant differ-
ences at pretest in cons of behavior change
for condition, (2, 92) = .43, P>.05, condi-
tion by gender, F(2, 92) = 2.08, P>.05, or
condition by stage F(4, 92) = .78, P>.05.
Nor were there any differences at pretest
for pros of behavior change for treatment
condition, F(2, 92) = .90, P>.05, condition
by gender, F(1, 92) = .56, P>.05, or condi-
tion by stage, F(4, 92) = .25, P>.05.
Self-efficacy Analyses
Condition by gender by pretest. Step
1: There was no significant interaction of
condition by pretest (yes or no), F(2, 161) =
0.63, P>.05, or condition by gender by
pretest, F(2, 161) = 0.489, P>.05; there-
fore, main effects tests of condition and
condition by gender were performed, and
step 4 was not required.
Step 2: No significant effects were found
for treatment condition, F(2, 167) = 1.84,
P>.05, or treatment condition by gender,
F(2, 167) = 1.89, P>.05.
Step 3: The ANCOVA with pretest score
as a covariate showed a significant inter-
action of treatment condition by gender,
F(2, 105) = 4.49, P<.05, η
2
=.08. Follow-up
Effects of Exercise
122
univariate tests with a Bonferonni ad-
justment for multiple comparisons showed
that males in the appearance condition
had significantly lower self-efficacy than
did either males in the control condition
or females in the appearance and health
condition. Figure 1 shows the estimated
marginal means (adjusted for the
covariate).
Self-efficacy subscale. A similar se-
ries of analyses were also performed by
self-efficacy subscale. Steps 1 and 2 were
not significant for all subscales. However,
for the step 3 ANCOVA, a significant con-
dition by gender interaction was found for
exercising alone, F(2, 105) = 6.43, P<.005,
η
2
=.11. Follow-up univariate tests with a
Bonferonni adjustment for multiple com-
parisons showed that males in the ap-
pearance condition had significantly lower
self-efficacy than did females in the ap-
pearance and health conditions. A simi-
lar series of analyses showed a signifi-
cant condition by gender interaction when
there was resistance from others, F(2,
105) = 4.03, P<.05, η
2
=.07. Follow-up
univariate tests with a Bonferonni ad-
justment for multiple comparisons showed
that males in the appearance condition
had significantly lower self-efficacy than
did females in the appearance condition.
Condition by stage by pretest.
Step 1: There was no significant inter-
action of condition by pretest (yes or no),
F(2, 146) = 0.85, P>.05, or condition by
stage by pretest, F(6, 146) = 0.29, P>.05.
Step 2: No significant results were
found for the main effects tests of condi-
tion, or condition by stage, F(2, 158) =
0.87, P>.05. There was a significant dif-
ference on self-efficacy scores by stage of
behavior change, F(3, 158) = 6.26, P<.001,
η
2
=.10. Posthoc tests with a Bonferroni
adjustment for multiple comparisons
showed that participants in the mainte-
nance stage had significantly higher self-
efficacy than did participants in the prepa-
ration or contemplation stages. Partici-
pants in the action stage also had signifi-
cantly higher self-efficacy than did par-
ticipants in the contemplation stage.
Step 3: No significant results were
found for the ANCOVA with pretest scores
as covariate, F(6, 97) = 0.48, P>.05.
No significant results were found for
any of the self-efficacy subscales.
Cons Analyses
Steps 1: No significant interactions
were found for condition by gender by
pretest, F(2, 158) = 0.12, P>.05 or condi-
tion by stage by pretest, F(6, 146) = 1.70,
P>.05.
Steps 2: Subsequent analyses showed
no significant differences for condition, F
(2, 158) = 1.27, P>.05, condition by gender,
F(2, 158) = 2.26, P>.05, or condition by
stage, F(6, 158) = 0.73, P>.05.
Steps 3: No significant results were
found for the condition by gender ANCOVA
with pretest scores as covariate, F(2, 103)
= .411, P>.05. Similarly, there were no
significant results for condition by stage
of behavior change, F(6, 97) = .20, P>.05.
Pros Analyses
Steps 1: No significant interactions
were found for condition by gender by
pretest, F(2, 158) = .19, P>.05, or condition
by stage by pretest, F(6, 146) = .49, P>.05.
Steps 2: Subsequent analyses showed
no significant differences for condition,
F(2, 164) = .05, P>.05, condition by gender,
F(2, 164) = .05, P>.05, or condition by
stage, F(6, 158) = .61, P>.05. There was a
significant difference on pros scores by
stage of behavior change, F(3, 160) = 2.72,
P<.05, η
2
=.05. Posthoc tests with a
Bonferroni adjustment for multiple com-
parisons showed that participants in the
60
58
56
54
52
50
48
46
44
Figure 1
Estimated Marginal Means
for Self-efficacy
Health Appearance Control
55.32
53.26
55.02
46.57
52.93
57.15
T r e a t m e n t C o n di t i o n
Female
Male
Berry & Howe
Am J Health Behav.
™™
™™
™ 2005;29(2):117-126
123
maintenance and action stages rated the
pros of changing exercise behavior as
significantly more important than did
participants in the contemplation stage.
Steps 3: No significant results were
found for the condition by gender ANCOVA
with pretest scores as covariate, F(2, 103)
= .01, P>.05. Similarly, there were no
significant results for condition by stage
of behavior change, F(6, 97) = .54, P>.05.
Estimated marginal means for the con-
dition by gender ANCOVAs for each de-
pendent variable and the self-efficacy
subscales are shown in Table 2.
DISCUSSION
The results of this study add to the
literature that examines the effects of
exercise advertising on consumers. Al-
though the hypotheses were not supported
in that there were no overall main effects
of condition for any of the dependent vari-
ables, there were significant interactions.
Results showed that males in the appear-
ance condition were negatively affected
by the advertising. Men who viewed ap-
pearance-based exercise advertising had
a significant negative change in their
overall self-efficacy and, more specifi-
cally, their self-efficacy for exercising
alone or for exercising when there is
resistance from others. This study pro-
vides evidence that appearance-based
advertising that focuses on images of
very attractive exercisers can have nega-
tive effects on self-efficacy for men in a
similar way that images of attractive,
thin women can have deleterious effects
on women’s body image.
These results support other research-
ers who have found that men who were
briefly exposed to advertising that fea-
tured muscular men showed a greater
discrepancy between their ideal body size
and their actual body size.
33
The present
study extended existing advertising re-
search to include exercise advertising,
and it is possible that it is the exercise
Table 2
ANCOVA Results and Estimated Marginal Means for
Condition by Gender for All Primary Dependent
Variables and Self-efficacy Subscales
Health Appearance Control
Female Male Female Male Female Male
Variable (n=26) (n=15) (n=24) (n=12) (n=21) (n=12) F
Self-efficacy 55.32 53.26 55.02 46.57 52.93 57.15 4.49*
S-E
a
alone 10.69 10.46 11.16 8.06 9.67 11.30 6.43**
S-E resistance 10.14 9.76 10.53 7.79 9.99 10.84 4.03*
S-E excuses 7.80 6.95 7.32 7.89 7.67 7.30 1.25
S-E inconvenient 8.66 8.02 8.66 7.94 8.76 7.98 0.01
S-E weather 9.25 9.39 8.94 8.38 8.73 10.70 2.44
S-E affect 8.82 7.86 9.10 7.98 8.36 8.99 1.66
Pros 10.02 10.86 9.27 9.90 10.06 11.40 1.10
Cons 22.12 20.59 21.95 20.70 21.99 21.55 0.41
* P<.05
** P<.005
a self-efficacy
Effects of Exercise
124
component of the advertisements used,
rather than just images of attractive
people, that made the male participants
particularly susceptible to appearance-
based exercise messages. It has been
found that men who read fitness maga-
zines had greater body dissatisfaction and
self-objectification than did men who did
not read such magazines and that men
read significantly more fitness magazines
than women did, whereas women read
more beauty magazines.
34
Further, Berry
and Lauzon
35
found that men’s fitness
magazines invariably focused on a very
muscular look, and other researchers
have found that 85% of sampled college-
aged men wanted to be more muscular.
36
The results of the present study showed
that besides body dissatisfaction, expo-
sure to appearance-based advertising may
result in men’s having less confidence in
their ability to exercise.
That there were no significant inter-
actions for any of the dependent variables
by stage of behavior change does not sup-
port the findings of Berry and Howe,
11
who
found that health-based exercise adver-
tising had a positive effect for those al-
ready active whereas appearance-based
advertising had a negative effect on
nonexercisers. These results are also
contrary to the findings of Bauman et al,
37
who reported that a mass media cam-
paign aimed at increasing exercise re-
sulted in positive changes in exercise
self-efficacy. However, these authors used
only 3 questions to measure self-efficacy
and did not specify what aspects of exer-
cise self-efficacy were examined. Fur-
ther research is needed to clarify whether
exercise advertising has different effects
for people at different stages of exercise
behavior.
Prochaska
13
recommended that public
health policies be dedicated to increasing
the benefits of healthy behavior change,
while decreasing the costs. However,
these recommendations may be prema-
ture as only mixed support for decisional
balance as a mediator between physical
activity interventions and behavior has
been found.
22
Further, the present study
found that neither health-based nor ap-
pearance-based exercise advertising af-
fected decisional balance for changing
exercise behavior. However, Rosen
38
found
that weakly argued messages (eg, exer-
cise will make you less bored) resulted in
more thoughts opposing exercise than did
messages that were strongly argued (eg,
exercise will result in more energy and
will make you more comfortable with your
body).
The health messages in the present
study were that exercise can “add years to
your life,”
29
that “it’s all about prevention
from diseases like diabetes, cancer, and
heart disease,”
28
and that you “won’t feel
as tired from a hectic day of living.”
30
These arguments could be described as
“strong-argument” messages and so may
not have resulted in many thoughts (ei-
ther positive or negative) regarding exer-
cise. It should be noted that participants
in Rosen’s
38
study read the messages and
so may have had more opportunity for
processing the information than did par-
ticipants in the present study, who saw
televised advertisements. Rosen con-
cluded that “one approach when targeting
people with low readiness for change is
using messages that require little pro-
cessing, such as imagery or paired asso-
ciations. Images of attractive people with
gym membership are examples of such
an approach” (p 178). In the present study
“images of attractive people with gym
memberships” had no effect on changing
attitudes.
Limitations in the present study in-
clude a convenience sample of students
enrolled in introductory psychology
classes. Results of a meta-analysis showed
that studies that had passive recruit-
ment, as in the present study, had fewer
precontemplators than did studies that
were randomly sampled.
19
Another limi-
tation is that the advertisements used in
this study included images of both attrac-
tive women and men, and so conclusions
are necessarily limited as to whether it
was the images of men, women, or both
that resulted in a decrease in self-effi-
cacy for men. Further, it is possible that
participants had previously been exposed
to the neutral or appearance advertise-
ments through regular television watch-
ing. It is less likely that participants
would have previously seen the health
advertisements because the promotion
campaigns were either from different
geographical areas than the one in which
the study took place, or the promotion
campaign had finished more than 2 years
prior to the study’s starting. A final limi-
tation is that there was no measure of
advertisement recall. It may be that dif-
ferent types of advertising are recalled
differently or processed differently. Fu-
Berry & Howe
Am J Health Behav.
™™
™™
™ 2005;29(2):117-126
125
ture research should address these limi-
tations.
CONCLUSION
It has been claimed that “boys learn to
view their bodies as a tool to master the
environment whereas girls learn that
their bodies should be used to attract
others.”
39
This may be true, but the re-
sults of this study highlight that men are
susceptible to mass media messages and
that perceived pressure to be strong and
muscular may be detrimental. Bandura
20
has identified that cultural modeling of
gender-role stereotypes is an influence
on self-efficacy and that people who watch
more television hold stronger gender-role
stereotypes than do people who watch less
television. Further, television images
have been identified as a major influence
on mental well-being and self-esteem.
40
The results of the present study showed
that appearance-based images of exer-
cisers can indeed have a negative influ-
ence on exercise self-efficacy in men.
These results may have strong implica-
tions because self-efficacy has been con-
sistently shown to increase across stages
of behavior change;
25
therefore, advertis-
ing that decreases self-efficacy may re-
sult in less actual exercise behavior.
These results have implications for
health promoters who are trying to in-
crease exercise participation. Although
Dishman
4
recommended that health and
body-image approaches be melded, the
results of this study would indicate that
this suggestion is premature. Health-
based exercise advertising did not have
any effect in the present study, whereas
appearance-based advertising had a nega-
tive influence. Slater
14
identified that a
challenge to health promoters is the nega-
tive modeling of health behaviors that is
evident in much entertainment and ad-
vertising media. This research provides
evidence for one way in which health may
be negatively influenced by entertain-
ment media images. With many North
Americans not engaging in adequate
amounts of physical activity, the re-
sults of this study highlight an impor-
tant health issue that needs to be fur-
ther examined.
REFERENCES
1.Blair SN, Brodney S. Effects of physical inac-
tivity and obesity on morbidity and mortality:
current evidence and research issues. Med Sci
Sports Exerc. 1999;31(11 Suppl):S646-S662.
2.U.S. Centers for Disease Control and Preven-
tion, 1996. Physical Activity and Health: A
Report of the Surgeon General (on-line). Avail-
able at: http://www.cdc.gov/nccdphp/sgr/
adults.htm. Accessed November 28, 2002.
3.Canadian Fitness and Lifestyle Research In-
stitute, 2001. 2001 Physical Activity Monitor
(on-line). Available at: http:.//www.cflri.ca/
cflri/pa/surveys/2001survey/
2001survey.html. Accessed April 3, 2003.
4.Dishman RK. The problem of exercise adher-
ence: fighting sloth in nations with market
economies. Quest. 2001;53(3):279-294.
5.Davis C, Fox J, Brewer H, et al. Motivations to
exercise as a function of personality charac-
teristics, age, and gender. Personality and
Individual Differences. 1995;19(2):165-174.
6.Gill K, Overdorf V. Incentives for exercise in
younger and older women. Journal of Sport
Behavior. 1994;17(2):87-97.
7.Lavine H, Sweeney D, Wagner SH. Depicting
women as sex objects in television advertis-
ing: effects on body dissatisfaction. Personal-
ity and Social Psychology Bulletin.
1999;25(8):1049-1058.
8.Brooks CM. Sport/exercise identity theory
and participation marketing: theory formation
and theoretical justification. Sport Marketing
Quarterly. 1998;7(1):38-47.
9.Marcus BH, Owen N, Forsyth LH, et al. Physi-
cal activity interventions using mass media,
print media, and information technology. Am
J Prev Med. 1998;15(4):362-378.
10.Cavill N. National campaigns to promote physi-
cal activity: can they make a difference?
International Journal of Obesity.1998;22(Suppl
2):S48-S51.
11.Berry TR, Howe BL. Effects of health-based
and appearance-based exercise advertising
on exercise attitudes, social physique anxi-
ety, and self-presentation in an exercise set-
ting. Social Behavior and Personality.
2004;32(1):1-12.
12.Rhodes RE, Courneya KS. Effects of a health-
based versus appearance based persuasive
message on attitudes towards exercise: test-
ing the moderating role of self-monitoring.
Journal of Social Behavior and Personality.
2001;15:321-330.
13.Prochaska JO. Strong and weak principles
for progressing from precontemplation to ac-
tion on the basis of twelve problem behaviors.
Health Psychol. 1994;13(1):47-51.
14.Slater MD. Integrating application of media
effects, persuasion and behavior change theo-
ries to communication campaigns: a stages-of-
change framework. Health Comm.
1999;11(4):335-354.
15.Prochaska JO, Norcross JC. Systems of Psy-
chotherapy. A Transtheoretical Analysis (3
rd
Ed.). Pacific Grove, CA: Brooks/Cole,
1994:453-492.
16.Rodgers WM, Courneya KS, Bayduza AL. Ex-
amination of the transtheoretical model and
Effects of Exercise
126
exercise in 3 populations. Am J Health Behav.
2001;25(1):33-41.
17.Schumann A, Nigg CR, Rossi JS, et al. Con-
struct validity of the stages of change of
exercise adoption for different intensities of
physical activity in 4 samples of differing age
groups. Am J Health Promot. 2002;16(5):280-
287.
18.Biddle SJH, Nigg CR. Theories of exercise
behavior. International Journal of Sport Psychol-
ogy. 2000;31(2):290-304.
19.Marshall SJ, Biddle SJH. The transtheoretical
model of behavior change: a meta-analysis of
applications to physical activity and exercise.
Ann Behav Med. 2001;23(4):229-246.
20.Bandura A. Self-efficacy: The Exercise of
Control. New York: W. H. Freeman, 1997:1-
604.
21.Wakefield M, Flay B, Nichter M, et al. Role of
the media in influencing trajectories of youth
smoking. Addiction. 2003;98(Suppl 1):79-103.
22.Lewis BA, Marcus BH, Pate RR, et al. Psycho-
social mediators of physical activity behavior
among adults and children. Am J Prev Med.
2002;23(Suppl 2):26-35.
23.Renger R, Steinfelt V, Lazarus S. Assessing
the effectiveness of a community-based media
campaign targeting physical inactivity. Fam
Community Health. 2002;25(3):18-30.
24.Cancer Prevention Research Center (on-line).
Available at: http://www.uri.edu/research/
cprc/measures.htm#Exercise. Accessed June
12, 2002.
25.Marcus BH, Simkin LR. The stages of exercise
behavior. J Sports Med Phys Fitness. 1993;33:83-
88.
26.Reed GR, Velicer WF, Prochaska JO, et al.
What makes a good staging algorithm: ex-
amples from regular exercise. Am J Health
Promot. 1997;12(1):57-66.
27.Plotnikoff RC, Blanchard C, Hotz SB, et al.
Validation of the decisional balance scales in
the exercise domain from the Transtheoreti-
cal model: a longitudinal test. Measurement in
Physical Education and Exercise Science.
2001;5(4):191-206.
28.Participation (Producer). Physical Activity: the
best investment you’ll ever make [Advertise-
ment]. 2001 Available: Participaction, 6-2400
St. W. Suite 726, Mississauga, Ontario, L5K
2R8.
29.Hillary Commission (Producer). Push Play
Campaign. [Advertisement]. 2001 Available:
Sport & Recreation New Zealand, Level 4, 78
Victoria Street, PO Box 2251,Wellington
New Zealand.
30.Dairy Nutrition Council of Alberta (Producer).
The nutrition file [Advertisement]. 2001 Avail-
able: Dairy Nutrition Council of Alberta.
31.Braver MCW, Braver SL. Statistical treatment
of the Solomon four-group design: a meta-
analytic approach. Psychol Bull.
1988;104(1):150-154.
32.Huck SW, Sandler HM. A note on the Solomon
4-group design: appropriate statistical analy-
ses. Journal of Experimental Education.
1973;42(2):54-55.
33.Leit RA, Pope JJ, Gray HG. The media’s
representation of the ideal male body: a case
for muscle dysmorphia? Int J Eat Disord. 2002;
31(3); 334-338.
34.Morry MM, Staska SL. Magazine exposure:
internalization, self-objectification, eating at-
titudes, and body satisfaction in male and
female university students. Canadian Journal
of Behavioral Science. 2001;33(4):269-279.
35.Berry T, Lauzon L. A content analysis of
fitness magazines. Avante. 2003;9(1):25-33.
36.Vartanian LR, Giant CL, Passino RM. Ally
McBeal vs. Arnold Schwarzenegger: Compar-
ing mass media, interpersonal feedback and
gender as predictors of satisfaction with body
thinness and muscularity. Social Behavior and
Personality. 2001;29(7):711-724.
37.Bauman AE, Bellew B, Owen N. Impact of an
Australian mass media campaign targeting
physical activity in 1998. Am J Prev Med.
2001;21(1):41-47.
38.Rosen CR. Integrating stage and continuum
models to explain processing of exercise mes-
sages and exercise initiation among seden-
tary college students. Health Psychol.
2000;19(2):172-180.
39.Groesz LM, Levine MP, Murnen SK. The effect
of experimental presentation of thin media
images on body satisfaction: a meta-analytic
review. Int J Eat Disord. 2002;31(1):1-16.
40.Signorielli N. Mass Media Images and Impact
on Health: A Sourcebook. Westport, CT: Green-
wood Press, 1993:1-28.
