Journal of Psychosomatic Research 62 (2007) 631 – 639
Emotional reactions following exposure to idealized bodies predict
unhealthy body change attitudes and behaviors in women and men
Sara Cahill, Alexander J. Mussap4
School of Psychology, Deakin University, Melbourne, Australia
Received 28 November 2005
Abstract
Objective: We explored the extent to which changes in
emotional states following exposure to images of idealized bodies
predict unhealthy body change attitudes and behaviors in women
and men, and whether particular psychological traits mediate
these effects. Method: One hundred thirty-three women and 93
men were assessed for unhealthy attitudes and behaviors related
to body weight and muscles using the Eating Disorder Inventory-
2 (EDI-2), the Obligatory Exercise Questionnaire, and the
strategies to increase muscles subscale of the Body Change
Inventory. Psychological traits assessed included body dissat-
isfaction (EDI-2), internalization of the thin/athletic ideal (Socio-
cultural Attitudes Towards Appearance Questionnaire-3), body
comparison (Body Comparison Scale), self-esteem (Rosenberg
Self-Esteem Inventory), depression (Beck Depression Inventory-
II), and identity confusion (Self-Concept Clarity Scale). Partic-
ipants were then exposed to photographs of thin female models
Keywords: Body image; Disordered eating; Exercise; Muscle development; Gender; Media; Unhealthy body change
Introduction
According to the tripartite influence model, sociocultural
pressure from peers, family, and the media to conform to
unrealistic physical ideals can contribute to negative body
image and unhealthy body change attitudes and behaviors
[1]; the thin ideal is implicated in drive for thinness and
disordered eating in women [2,3], and the athletic/muscular
ideal is implicated in drive for muscularity, compulsive
overexercising, overeating, and use of chemical supplemen-
tation in men [4–6]. Consistent with a role for the media are
4 Corresponding author. School of Psychology, Deakin University,
221 Burwood Highway, Melbourne 3125, Australia.
E-mail address: mussap@deakin.edu.au (A.J. Mussap).
0022-3999/07/$ – see front matter D 2007 Elsevier Inc. All rights reserved.
doi:10.1016/j.jpsychores.2006.11.001
and muscular male models, and visual analogue scales were used
to measure changes in postexposure state body dissatisfaction,
anger, anxiety, and depression. Results: Postexposure increases in
state anger, anxiety, depression, and body dissatisfaction corre-
lated with drive for thinness and disordered eating symptomatol-
ogy in women, while postexposure increases in state body
dissatisfaction correlated with muscle development in men.
Analyses revealed that internalization and body comparison
mediated these relationships, with trait body dissatisfaction, trait
depression, self-esteem, and self-concept/identity confusion serv-
ing as mediators for women only. Conclusion: These results are
indicative of gender differences in: (a) reactions to idealized
bodies; (b) psychological traits that predispose individuals to
experience these reactions; and (c) types of body change behavior
that are associated with these reactions.
D 2007 Elsevier Inc. All rights reserved.
the results of correlational studies suggesting that a history
of exposure to the media predicts level of body dissat-
isfaction, weight concerns, and disordered eating sympto-
matology in women, and body dissatisfaction, muscle
concern, and symptoms of unhealthy muscle development
in men [7–11], and experimental studies suggesting that
even short-term exposure to images of idealized bodies can
negatively impact on body satisfaction and mood in both
women and men [12–19].
To account for individual differences in unhealthy body
change attitudes and behaviors (especially given that
idealized bodies are almost ubiquitous in Western soci-
eties) [20], it has been suggested that certain psycholo-
gical traits predispose individuals to react emotionally to
images of idealized bodies, and that unhealthy body
change reflects an accumulated impact of these emotional
632 S. Cahill, A.J. Mussap / Journal of Psychosomatic Research 62 (2007) 631 – 639
reactions over a lifetime [1,21]. In the present study, an
experimental design was employed to identify which
psychological states are most reactive to short-term
exposure to idealized bodies, the extent to which these
reactions are similar for females and males, and whether
certain psychological traits mediate the relationship
between state reactivity and long-term symptoms of un-
healthy body change. Identifying relevant psychological
factors would not only enhance our understanding of the
sociocultural origins of unhealthy body change but also
facilitate the identification and early intervention of
individuals who are most vulnerable to sociocultural
pressures on body image.
The psychological factors in the present study were chosen
on the basis of sociocultural models and the results of
previous research. In the dual-pathway model [22], socio-
cultural pressures on appearance (e.g., via the media),
coupled with internalization of the physical ideal (e.g., the
thin ideal), lead to body dissatisfaction, which in turn
contributes both to body change behaviors (dieting in the
case of dissatisfaction with weight) and to negative affect
associated with failure to achieve and maintain the internal-
ized ideal. Together, dieting and negative affect promote
unhealthy body change attitudes and behaviors characteristic
of disordered eating. If the dual-pathway model is correct,
women with elevated trait body dissatisfaction, who internal-
ize the thin ideal, value and frequently engage in body
comparisons, and experience depressed mood, would be
vulnerable to experiencing negative emotional reactions to
thin models. Since individuals with an unclear body image
would be most susceptible to sociocultural pressures, it is also
expected that elevated trait identity confusion and diminished
trait self-esteem would be associated with negative emotional
reactions to idealized bodies. Research has confirmed each of
these relationships in terms of vulnerability to experiencing
decreases in state body satisfaction and increases in anger,
anxiety, and depression in women following exposure to thin
female models [12,21,23–25].
On the basis of these results on females, together with
emerging evidence of similar negative emotional reactions to
idealized muscular bodies by males [11,17,26], it was
hypothesized that viewing idealized same-sex bodies would
result in increased state body dissatisfaction, depression,
anxiety, and anger in both women and men. If these emo-
tional reactions are of clinical relevance, then they should
also correlate with unhealthy body change attitudes and
behaviors; in the present study, we evaluated drive for
thinness, symptoms of bulimia, obligatory exercise, and
participation in strategies to increase muscles. Most impor-
tantly, if the psychological traits identified above actually
predispose individuals to experiencing negative emotional
reactions, then any relationship obtained between change in
emotional state and unhealthy body change should be
mediated by trait body dissatisfaction, internalization of the
thin/muscular ideal, body comparison, self-esteem, depres-
sion, and identity confusion.
Method
Participants
Participants were a self-selected sample of 93 males (age:
mean=22.34, S.D.=4.21) and 133 females (age: mean=22.42,
S.D.=4.59), predominantly Australian students of Deakin
University. An invitation to participate was announced during
lectures across a range of faculties and classes. No reward for
participation was offered. There is evidence to suggest that
university exposes both females and males to peer pres-
sure on appearance that can impact on their body image, self-
esteem, and adoption of unhealthy body change attitudes
and behaviors [27].
Materials and procedure
Two versions of a questionnaire were produced: one for
women (with photographs of thin female models) and one
for men (with photographs of muscular male models). Both
versions contained the same scales, completed in the order
presented below. Participants were given only a broad
overview of the study; hypotheses were not discussed.
Body mass index (BMI)
BMI (for women: mean=22.3, S.D.=3.8; for men:
mean=23.5, S.D.=2.9) was calculated from self-reported
weight and height [BMI=weight (kg)/height (m)2]. Apply-
ing National Institutes of Health guidelines, 7.5% of females
and 3.2% of males were bunderweightQ (BMIb18.5); 80.5%
of females and 68.8% of males were of bnormalQ weight
(BMI=18.5–24.9); 7.5% of females and 23.7% of males
were boverweightQ (BMI=25.0–29.9); and 4.5% of females
and 4.3% of males were bobeseQ (BMIz30.0).
Trait measures
Body dissatisfaction
The body dissatisfaction subscale (DISS; for women:
a=.86, mean=3.0, S.D.=0.8; for men: a=.76, mean=2.6,
S.D.=0.6) of the Eating Disorder Inventory-2 (EDI-2) [28]
was used. Participants responded to statements such as
bI think my hips are too big.Q Three additional items were
included from the Body Change Inventory (BCI) [5] to assess
male-relevant concerns, such as bI think that the size/width of
my shoulders is just right.Q All EDI-2 and BCI responses were
scored between 1 (never) and 5 (always).
Internalization
The TV–magazine and athlete subscales of the Socio-
cultural Attitudes Towards Appearance Questionnaire-3
[29] (INTERN; for women: a=.84, mean=2.9, S.D.=0.8;
for men: a=.90, mean=2.6, S.D.=0.9) were used to assess
agreement with appearance-related information. Responses
were scored between 1 (never) and 5 (always).
 S. Cahill, A.J. Mussap / Journal of Psychosomatic Research 62 (2007) 631 – 639
Comparison
The muscular and weight subscales of the Body
Comparison Scale [30] (COMP; for women: a=.84,
mean=2.6, S.D.=0.7; for men: a=.88, mean=2.3, S.D.=0.8)
were used to assess the frequency of comparing body
regions relevant to weight and musculature. Responses were
scored between 1 (never) and 5 (always).
Self-esteem
The 10-item Rosenberg Self-Esteem
P Inventory [31]
(SELFEST; for women:
P a=.90, =36.5, S.D.=6.6;
for men: a=.91, =39.7, S.D.=6.5) was used, with
responses scored between 0 (strongly disagree) and 3
(strongly agree).
Beck Depression Inventory (BDI)
P The 21-item BDI-II [32] (BDI; for P women: a=.92,
=11.6, S.D.=9.7; for men: a=.92, =8.8, S.D.=8.0)
was used. The sample did not display a distribution of
depression substantially different from that of US university
samples [32]; on the basis of cutoff score guidelines for the
BDI-II, approximately 15% of the female sample and 11%
of the male sample were within the moderate to severe range
for depression, with most (67% and 73% for females and
males, respectively) being within the minimal risk range.
Identity confusion
The 12-item Self-Concept P Clarity Scale [33] (IDEN-
CON; P for women: a=.89, =35.4, S.D.=8.2; for men:
a=.89, =33.1, S.D.=8.6) was used, with responses scored
between 0 (strongly disagree) and 4 (strong agree).
Body change attitudes and behaviors relevant to
disordered eating
Bulimia
The bulimia subscale (BUL; for women: a=.78,
mean=2.1, S.D.=0.6; for men: a=.72, mean=1.7, S.D.=0.5)
of the EDI-2 was used to assess the level of agreement with
statements such as bI eat moderately in front of others and
stuff myself when they are gone.Q Alternating binging and
purging distinguishes bulimia nervosa from nonspecific
disordered eating [28].1
Drive for thinness
The drive for thinness subscale (DFT; for women: a=.89,
mean=2.9, S.D.=0.8; for men: a=.82, mean=1.9, S.D.=0.7)
of the EDI-2 was used to assess the level of agreement with
statements such as bI am preoccupied with the desire to be
1
Although the EDI-2 is normally scored on a 6-point scale, for
consistency, all scales that measured unhealthy body change attitudes and
behaviors were converted to 5-point scales scored between 1 (always) and 5
(never).
633
thinner.Q Drive for thinness is a fundamental characteristic of
disordered eating and constitutes a primary diagnostic
criterion for diagnosis with an eating disorder [28].
Body change attitudes and behaviors relevant to
body/muscle development
Obligatory exercise
The Obligatory Exercise Questionnaire [34] (EXER;
for women: a=.75, mean=2.5, S.D.=0.8; for men: a=.82,
mean=2.4, S.D.=0.7) was used to assess the emotional
aspects of exercising, exercise frequency and intensity, and
preoccupation with exercise using statements such as
bI exercise even when I am advised against it (by a doctor,
friend, etc).Q
Strategies to increase muscles
The revised strategies to increase muscles subscale
(MUSC; for women: a=.75, mean=2.1, S.D.=0.7; for men:
a=.78, mean=2.6, S.D.=0.8) of the BCI [35] assessed the
level of engagement in activities designed to increase
muscle size using statements such as bI take food supple-
ments (such as vitamins, protein drinks, or diet pills) to
increase my muscles.Q
State visual analogue scales (VAS)
A series of VAS was used to assess psychological
states. Four questions of the following form were
presented: bRight now, how ________ do you feel?Q
The words/phrases inserted in each case were bsatisfied
with you body,Q bdepressed,Q banxious,Q or bangry.Q
Immediately below each question was a horizontal line
100 mm in length anchored with the phrase not at all on
the left and very satisfied, very depressed, very anxious,
or very angry on the right. Participants indicated their
response to each VAS item by placing a mark on the line
with a pen. The level of agreement with each VAS item
was scored by measuring the length of the left half of the
bisected line.
VAS items were presented twice to participants:
immediately preceding exposure to photographs of female
and male models (see below), and immediately following
exposure to these photographs. VAS items have been
shown to be sensitive enough to register differences
before and after brief interventions, and to have good
convergent validity with various body dissatisfaction and
mood scales [36].
Photographs of thin female models and muscular
male models
After the first set (preexposure) of VAS measures had
been taken, six photographs of female and male models,
634 S. Cahill, A.J. Mussap / Journal of Psychosomatic Research 62 (2007) 631 – 639

Table 1
Means and paired-samples t tests of preexposure versus postexposure VAS scores
Women [mean (S.D.)] Men [mean (S.D.)]
State VAS Preexposure Postexposure t Preexposure Postexposure t
Anger 16.20 (21.60) 18.85 (22.90) 2.184 14.94 (21.05) 16.60 (20.56) 1.49
Anxiety 29.49 (27.54) 27.77 (25.40) 1.28 26.67 (27.49) 25.37 (26.62) 0.99
Depression 27.19 (40.73) 27.76 (27.79) 0.21 19.56 (26.48) 21.52 (25.89) 1.49
Satisfaction 46.77 (26.23) 42.34 (27.97) 3.2744 58.59 (25.53) 54.82 (27.35) 2.454
4 Significant at Pb.05.
44 Significant at Pb.01.

sourced from the media, were presented two to a page
(participants were exposed only to models of the same sex
as themselves). Female models were dressed in bikinis,
while male models were topless and dressed in shorts. All
female models were visibly thin (the thinnest models we
could source), and all male models were visibly lean and
muscular bfitnessQ models. The photographs were sourced
from the Internet, and participants’ own ratings of these
photographs were used to confirm the idealized thin/
muscular status of the bodies depicted (see Results section).
These ratings also ensured that participants attended and
engaged in an aesthetic evaluation of each image. Specif-
ically, each photograph was accompanied by two questions:
(a) bHow well does she/he compare to your ideal body?Q,
which was scored between 1 (she is much too thin/his
muscles are much too small) and 5 (she is much too fat/his
muscles are much to large), and (b) bHow much would you
like to look like her/him?Q, which was scored between 1 (not

Table 2
Measures of unhealthy body change regressed on each preexposure VAS score in Step 1 and on each postexposure VAS score in Step 2 (data for women are
shown)
DV VAS Step R R2 DR 2 B b t r sr 2
BUL Anger (1) Preexposure .10 .01 .02 0.00 0.09 0.61 .10 0.00
(2) Postexposure .17 .03 0.01 0.23 1.60 .16 0.02
Anxiety (1) Preexposure .2344 .05 .054 0.00 0.09 0.59 .2344 0.00
(2) Postexposure .324 .10 0.01 0.39 2.574 .3144 0.04
Depression (1) Preexposure .3044 .09 .044 0.00 0.14 1.28 .3044 0.01
(2) Postexposure .364 .13 0.01 0.25 2.404 .3444 0.04
Satisfaction (1) Preexposure .3944 .15 .034 0.00 0.15 1.02 .3944 0.01
(2) Postexposure .434 .18 0.01 0.29 2.044 .4244 0.03
DFT Anger (1) Preexposure .01 .00 .0944 0.02 0.39 2.764 .01 0.05
(2) Postexposure .2944 .09 0.02 0.49 3.4944 .184 0.08
Anxiety (1) Preexposure .02 .00 .0944 0.01 0.43 2.884 .02 0.06
(2) Postexposure .3144 .09 0.02 0.55 3.6444 .194 0.09
Depression (1) Preexposure .2444 .06 .1044 0.00 0.03 0.26 .2444 0.00
(2) Postexposure .4044 .16 0.01 0.42 4.0144 .4044 0.10
Satisfaction (1) Preexposure .5544 .30 .054 0.01 0.22 1.71 .5544 0.01
(2) Postexposure .5944 .35 0.01 0.40 3.094 .5844 0.05
MUSC Anger (1) Preexposure .184 .03 .01 0.00 0.02 0.12 .184 0.00
(2) Postexposure .21 .05 0.01 0.20 1.38 .214 0.01
Anxiety (1) Preexposure .174 .03 .00 0.00 0.10 0.64 .174 0.00
(2) Postexposure .17 .03 0.00 0.09 0.55 .174 0.00
Depression (1) Preexposure .07 .01 .02 0.00 0.05 0.46 .07 0.00
(2) Postexposure .17 .03 0.01 0.20 1.77 .174 0.02
Satisfaction (1) Preexposure .16 .03 .00 0.00 0.13 0.82 .16 0.00
(2) Postexposure .16 .03 0.00 0.04 0.24 .15 0.00
EXER Anger (1) Preexposure .00 .00 .01 0.00 0.12 0.81 .00 0.00
(2) Postexposure .09 .01 0.01 0.15 1.02 .05 0.01
Anxiety (1) Preexposure .08 .01 .00 0.00 0.03 0.21 .08 0.00
(2) Postexposure .09 .01 0.00 0.06 0.35 .08 0.00
Depression (1) Preexposure .07 .00 .01 0.00 0.01 0.12 .07 0.00
(2) Postexposure .12 .02 0.00 0.13 1.12 .12 0.01
Satisfaction (1) Preexposure .214 .05 .00 0.00 0.15 1.00 .214 0.01
(2) Postexposure .22 .05 0.00 0.77 0.45 .204 0.00
DV, dependent variable.
4 Significant at P b.05.
44 Significant at P b.01.
 S. Cahill, A.J. Mussap / Journal of Psychosomatic Research 62 (2007) 631 – 639 635

Table 3
Measures of unhealthy body change regressed on each preexposure VAS score in Step 1 and on each postexposure VAS score in Step 2 (data for men
are shown)
DV VAS Step R R2 DR 2 B b t r sr2
BUL Anger (1) Preexposure .12 .02 .04 0.00 0.21 1.01 .12 0.01
(2) Postexposure .23 .05 0.00 0.38 1.86 .204 0.04
Anxiety (1) Preexposure .19 .03 .00 0.00 0.11 0.49 .184 0.00
(2) Postexposure .19 .04 0.00 0.08 0.35 .184 0.00
Depression (1) Preexposure .15 .02 .00 0.00 0.14 0.62 .15 0.00
(2) Postexposure .15 .02 0.00 0.02 0.07 .14 0.00
Satisfaction (1) Preexposure .15 .02 .01 0.00 0.33 1.71 .15 0.03
(2) Postexposure .19 .04 0.00 0.21 1.10 .07 0.01
DFT Anger (1) Preexposure .12 .02 .00 0.00 0.14 0.65 .12 0.00
(2) Postexposure .12 .02 0.00 0.02 0.08 .10 0.00
Anxiety (1) Preexposure .03 .00 .00 0.00 0.12 0.51 .03 0.00
(2) Postexposure .05 .00 0.00 0.10 0.42 .01 0.00
Depression (1) Preexposure .01 .00 .00 0.00 0.08 0.37 .01 0.00
(2) Postexposure .04 .00 0.00 0.08 0.36 .01 0.00
Satisfaction (1) Preexposure .18 .03 .01 0.00 0.00 0.01 .184 0.00
(2) Postexposure .21 .04 0.00 0.20 1.10 .214 0.01
MUSC Anger (1) Preexposure .18 .03 .01 0.00 0.02 0.09 .184 0.00
(2) Postexposure .21 .05 0.01 0.23 1.12 .214 0.01
Anxiety (1) Preexposure .09 .01 .02 0.01 0.18 0.77 .09 0.01
(2) Postexposure .16 .03 0.01 0.30 1.31 .14 0.02
Depression (1) Preexposure .09 .01 .00 0.01 0.21 0.93 .09 0.01
(2) Postexposure .11 .01 0.00 0.14 0.61 .05 0.00
Satisfaction (1) Preexposure .04 .00 .064 0.01 0.33 1.74 .04 0.03
(2) Postexposure .244 .06 0.01 0.44 2.314 .16 0.06
EXER Anger (1) Preexposure .20 .04 .02 0.00 0.01 0.07 .204 0.00
(2) Postexposure .24 .06 0.01 0.25 1.21 .2444 0.01
Anxiety (1) Preexposure .12 .01 .02 0.00 0.15 0.64 .12 0.00
(2) Postexposure .18 .03 0.01 0.30 1.31 .174 0.02
Depression (1) Preexposure .15 .02 .00 0.00 0.07 0.30 .15 0.00
(2) Postexposure .15 .03 0.00 0.09 0.41 .15 0.00
Satisfaction (1) Preexposure .09 .01 .03 0.01 0.18 0.94 .09 0.01
(2) Postexposure .20 .04 0.01 0.32 1.67 .174 0.03
DV, dependent variable.
4 Significant at Pb.05.
44 Significant at Pb.01.

at all) and 5 (extremely). Because the study was question-
naire-based, the timing of each component of the study and
the intervals between these components (preexposure items,
photographs, and postexposure items) were entirely under
the control of participants.
Results
Data screening
Female and male data were screened separately for
missing values (which were replaced with mean values
for normally distributed variables and median values
for skewed variables). Scales were created from averaged
responses to internally consistent items (item–total
correlationN.3 and Cronbach’s aN.70; see Method sec-
tion). Univariate outliers (NF3 S.D. from the mean)
were identified and replaced with the value corre-
sponding to 3 S.D. All scales were screened for viola-
tions of normality and were transformed when skewness
(skewness statistic/standard error of skewness) was N4.
No multivariate outliers were identified.
Aesthetic ratings of images of idealized bodies
On the basis of the ratings given, women perceived
female models to be between ba little too thinQ and bjust
rightQ (women: mean=2.2, S.D.=0.5), while men perceived
male models to possess muscles that are between ba little
too largeQ and bjust rightQ (men: mean=3.6, S.D.=0.4). On
average, both female and male participants rated their desire
to appear like the models as bmoderateQ (women: mean=2.8,
S.D.=0.9; men: mean=2.6, S.D.=0.7; where 3=moderately).
Emotional reactions and unhealthy body change attitudes
and behaviors
A series of paired-samples t tests tested for postexposure
changes in VAS score. Table 1 shows that both women and
men reported significantly less state body satisfaction, and
women reported significantly more state anger. These results
636 S. Cahill, A.J. Mussap / Journal of Psychosomatic Research 62 (2007) 631 – 639

confirm that exposure to the models induced nontrivial
psychological changes in the two groups.
The hypothesis that individual differences in these VAS
changes would be related to preexisting symptoms of
unhealthy body change was tested in a series of two-step
hierarchical regression analyses. Each measure of unhealthy
body change was regressed on each preexposure state VAS
measure in Step 1 and on each postexposure state VAS
measure in Step 2. Significant additional variance predicted
at Step 2 (i.e., significant DR 2 due to the inclusion of
postexposure state measure) would support the presence of a
relationship between change in that psychological state and
unhealthy body change. (The use of hierarchical analyses to
measure change in a variable has important advantages over
simply using difference scores [21,36].)
The results of hierarchical analyses are presented in
Tables 2 and 3. For women, increased state anxiety and
depression, and decreased state body satisfaction following
exposure to the same-sex models were significantly related
to bulimia (DR 2=.05, .04, and .03, Pb.05, for state anxiety,
depression, and body satisfaction, respectively) and drive
for thinness (DR 2=.09, Pb.01; DR 2=.10, Pb.05; and
DR 2=.05, Pb.05 for state anxiety, depression, and body
satisfaction, respectively). In addition, change in state anger
was also significantly related to drive for thinness (DR 2=.09,
Pb.05). Behaviors associated with body and muscle
development (strategies to increase muscles, obligatory
exercise) were unrelated to change in any of the state
VAS scores. For men, decreased body satisfaction was
significantly related to level of engagement in strategies to
increase muscles (DR 2=.06, Pb.05), and none of the state
measures was related to unhealthy weight loss.
Mediation analyses
A variant of the method of Baron and Kenny [37] was
then employed to test which of the psychological traits
mediated the above relationships. A mediated path would
need to meet four conditions. First, a direct relationship
must exist between a change in state VAS score and the
measure of unhealthy body change. This condition was
tested in the hierarchical analyses described in the previous
paragraph. Second, a change in state VAS score must
correlate with a putative mediator (i.e., with the psycho-
logical trait of interest). This condition was met if significant
DR 2 values were obtained when each trait score was
regressed on each preexposure state VAS score in Step 1
and on each postexposure state VAS score in Step 2. Third,
the putative mediator must also correlate with the measure
of unhealthy body change in question (cf., Table 4). Finally,
traits that correlated both with change in the relevant state
VAS score (identified in Condition 2) and with the relevant
measure of unhealthy body change (identified in Condition
3) could be tested as mediators by determining whether their
inclusion in the regression equation reduces to nonsignifi-
cance the direct path (of Condition 1) between change in
state VAS and unhealthy body change. To test this final
condition, original hierarchical regressions were repeated as

Table 4
Separate correlations for females and males (Pearson’s r)
BMI BUL DFT MUSC EXER INTERN COMP DISS SELFEST BDI IDCONF
Women
Age .3444 .05 .08 .15 .13 .16 .19 .04 .204 .18 .2644
BMI – .12 .11 .04 .08 .06 .04 .36 .08 .04 .09
BUL – .6044 .2944 .2644 .4344 .4944 .4344 .4144 .3444 .3744
DFT – .3244 .4744 .6744 .6344 .6744 .4644 .2744 .3844
MUSC – .5144 .2844 .3444 .13 .2344 .184 .16
EXER – .3344 .3944 .15 .194 .09 .174
INTERN – .7244 .5044 .4144 .204 .2944
COMP – .5044 .4244 .2644 .3544
DISS – .5744 .3944 .3744
SELFEST – .6844 .6144
BDI – .6344
Men
Age .4044 .05 .14 .18 .10 .3144 .20 .12 .14 .12 .18
BMI – .02 .3444 .02 .13 .234 .10 .18 .214 .11 .214
BUL – .4144 .244 .3244 .18 .18 .224 .20 .18 .3444
DFT – .264 .3944 .2844 .3744 .4144 .04 .04 .15
MUSC – .5844 .4444 .5644 .20 .05 .01 .254
EXER – .4344 .5144 .15 .20 .06 .3244
INTERN – .7944 .4844 .3144 .17 .5344
COMP – .4144 .224 .07 .3944
DISS – .3444 .3044 .3544
SELFEST – .6944 .5344
BDI – .5044
4 Significant at P b.05.
44 Significant at P b.01.
 S. Cahill, A.J. Mussap / Journal of Psychosomatic Research 62 (2007) 631 – 639
three-step regressions, with each putative mediator included
individually in the second step of the regression—between
each preexposure state VAS score (included in the first step)
and each postexposure state VAS score (included in the
third step).
For women, the significant direct path between change in
state anxiety and bulimia (DR 2=.05, Pb.01) was reduced to
nonsignificance by the inclusion of internalization
(DR 2=.01, P=.17) and comparison (DR 2=.02, P=.11); the
path between state depression and bulimia (DR 2=.04,
Pb.05) was reduced to nonsignificance by the inclusion of
trait body dissatisfaction (DR 2=.01, P=.33), internalization
(DR 2=.01, P=.26), trait self-esteem (DR 2=.00, P=.57), trait
depression (DR 2=.01, P=.15), and trait identity confusion
(DR 2=.01, P=.21); the path between change in state body
satisfaction and bulimia (DR 2=.03, Pb.05) was reduced to
nonsignificance by the inclusion of trait body dissatisfaction
(DR 2=.00, P=.38), comparison (DR 2=.01, P=.18), and
internalization (DR 2=.01, P=.38); and the path between
change in state anger and drive for thinness (DR 2=0.09,
Pb.01) was reduced to nonsignificance by the inclusion of
comparison (DR 2=.01, P=.24).2
For men, the only significant direct path obtained—
between change in state body satisfaction and strategies to
increase muscles (DR 2=.06, Pb.05)—was reduced to non-
significance only by the inclusion of internalization
(DR 2=.03, P=.09) and comparison (DR 2=.01, P=.20).
Discussion
The proposition that emotional reactions to media images
of idealized bodies are relevant to the impact of sociocultural
pressures on body image and unhealthy body change
attitudes and behaviors was tested. As expected on the basis
of previous research [13,21], exposure to thin female models
significantly decreased state body satisfaction and increased
state anger in women, while exposure to muscular male
models decreased state body satisfaction in men. These
effects were obtained after only very brief exposures and after
exposure to only six photographs—far less than an individual
might witness from reading even a single fashion or sports
magazine. More importantly, individual differences in these
emotional reactions were relevant to unhealthy body change
attitudes and behaviors. For women, postexposure increases
in state anxiety and depression, and decreases in state body
satisfaction, were associated with an elevated drive for
thinness (accounting for between 5% and 10% of variance
in VAS scores) and with bulimic symptoms (accounting for
between 2% and 5% of variance in VAS scores). For men,
postexposure change in state body satisfaction was associated
2 symptomatology [41].
As noted by an anonymous reviewer, since trait depression correlated
with many other trait measures (self-esteem and body dissatisfaction in
particular; see Table 4), it is possible that some of the abovementioned
effects were due to a common underlying influence of depression.
637
with strategies to increase muscles (accounting for 6% of
variance in VAS scores).
These results are consistent with individual differences
in psychological vulnerability to images of idealized
bodies. To test this interpretation, a range of psychological
traits were tested as potential mediators of the relationships
between change in emotional state and unhealthy body
change. For women, internalization of the thin ideal
mediated the relationships between increased state anger,
anxiety, depression, and decreased body satisfaction
following exposure to thin models, and symptoms of
unhealthy weight loss. For men, internalization of the
athletic/muscular ideal mediated the relationship between
decreased state body satisfaction and participation in
strategies to increase muscles. These relationships support
the dual-pathway model’s contention that exposure to
media images of idealized bodies is not necessarily
sufficient to induce clinically relevant changes in psycho-
logical state; the media images must be valued by the
individual [38]. That internalization was relevant not only
for women but also for men further suggests that the
model is relevant to concerns with muscularity, as well to
as concerns with thinness.
Frequency of body comparison was identified as a
mediator of the relationships between both increased state
anxiety and decreased state body satisfaction and bulimic
symptoms in women, and also mediated the relationship
between decreased state body satisfaction in men and
strategies to increase muscles. These relationships suggest
that individuals predisposed to body comparisons are more
likely to psychologically engage with images of thin/
muscular bodies, thereby increasing the salience of these
images. Previous research has found that body comparison
mediates the relationship between real-world media expo-
sure and disordered eating [39]. The present study extended
this finding to men and the pursuit of muscularity.
In women, trait body dissatisfaction was found to
mediate the relationships between increased state depression
and both drive for thinness and bulimic symptomatology,
and trait depression mediated the relationship between
increased state depression and bulimic symptoms. Body
dissatisfaction and depression are known to predict disor-
dered eating in women [22,40]. The absence of mediating
effects involving these trait measures for men suggests that
stable body dissatisfaction in men is not as relevant to male
body change attitudes and behaviors as is malleability of
state body dissatisfaction. Similar arguments have been
made in the context of women and unhealthy attitudes and
behaviors associated with the desire for thinness and weight
loss, where it has been found that unstable (rather than
simply negative) self-evaluations made by women following
exposure to thin bodies are predictive of disordered eating
Low trait self-esteem in women was identified as a
mediator of the relationship between increased state
depression and both drive for thinness and bulimic
638 S. Cahill, A.J. Mussap / Journal of Psychosomatic Research 62 (2007) 631 – 639
symptomatology. Trait self-esteem is known to be associ-
ated with body dissatisfaction and symptoms of disordered
eating [23]—an observation that is consistent with the idea
that self-esteem reflects, at least in part, an individual’s
evaluation of one’s own appearance. It has also been
suggested that individuals with low self-esteem may attempt
to increase their self-esteem by striving for social accept-
ance, perhaps by endeavoring to meet culturally mandated
standards of physical appearance [42]. The results of the
present study lend partial support to this idea. Trait identity
confusion was also found to mediate the relationship
between increased state depression and bulimic symptoms.
This is consistent with the proposition that women with an
unstable or unclear self-concept may be more inclined to
look at culturally prescribed ideals to find their sense of
identity [42].
Given the variety of mediators identified in the present
study, it would be worthwhile for future research to assess
other factors that meet the basic prerequisites for mediation
(that they correlate both with emotional reactivity and with
measures of unhealthy body change). Such factors include
sexual orientation and gender–role orientation [43–45], fear
of negative appearance evaluation [46], obsessiveness and
perfectionism [47], impulsivity [48], and participation in
sporting activities [49].
As to limitations, it is important to note that the study
possessed only limited ecological validity (cf., Thompson et
al. [1]); a brief and isolated exposure to idealized bodies
does not replicate a lifetime of repeated exposure to such
bodies, which one experiences in the real world, nor does it
simulate the positive outcomes (success, happiness, popular-
ity, etc.) that the media associate with these bodies. This
limitation is exacerbated by having the timing of each
component of the study and the intervals between these
components under the control of participants. Furthermore,
the use of VAS scores, while ideally suited to the short
exposure durations involved in this study (because VAS
scales are unmarked), does not permit interpretation of the
clinical significance of changes in psychological state. Thus,
while causal relationships between changes in state VAS
scores and exposure to same-sex models can be stated with
some degree of certainty due to the experimental design
employed, the implications of these effects on unhealthy
body change behaviors must be made cautiously—the
absence of longitudinal data; the use of an adult sample
where any evidence of unhealthy body change preexists; the
short exposure durations employed; the short-lived effects
of exposure measured in the study; and the lack of clinically
validated measures of state change make it impossible
to confirm that it is an accumulation of emotional reactions
to idealized bodies that contributes to the development
of unhealthy body change [50]. To do so would require
further research to confirm that the inherent instability of
psychological states actually precedes the development of
negative body image and unhealthy body change attitudes
and behaviors.
It is also impossible to dismiss the influence on the
results of demand characteristics. The body change meas-
ures related to weight and muscles provided participants
with sufficient information to conclude that we expected
negative emotional reactions to the models. However, it is
unlikely that this could have produced the complex pattern
of relationships obtained between emotional reactions and
unhealthy body change attitudes and behaviors—relation-
ships that were mediated by specific psychological traits.
In conclusion, there is a multidimensional and gendered
relationship between exposure to media images of idealized
bodies, psychological reactions to these images, and
attitudes and behaviors that reflect a desire to conform to
these images. While the media’s effects on men appear to be
limited to their current sense of body satisfaction and appear
to contribute to (or interact with) their desire to increase
muscles, the media’s effects on women influence their
current sense of anger, anxiety, depression, and body
satisfaction, and appear to contribute to (or interact with)
their desire to decrease weight. Thus, the results not only
confirm that men and women differ in their body image
concerns, attitudes, and body change behaviors [51] but also
point to differences in the psychological processes that
determine the effects of media exposure on these concerns,
attitudes, and behaviors.
Acknowledgments
This research was supported by a grant from the
Australian Research Council.
References
[1] Thompson JK, Heinberg LJ, Altabe M, Tantleff-Dunn S.
Exacting beauty. Washington (DC)7 American Psychological Associ-
ation, 1999.
[2] van den Berg P, Thompson JK, Obremski-Brandon K, Coovert M.
The tripartite influence model of body image and eating disturb-
ance: a covariance structure modeling investigation testing the
mediational role of appearance comparison. J Psychosom Res
2002;53:1007 – 20.
[3] Cash TF, Brown TA. Body image in anorexia nervosa and
bulimia nervosa: a review of the literature. Behav Modif 1987;11:
487 – 521.
[4] Olivardia R, Pope HG, Hudson JI. Muscle dysmorphia in
male weightlifters: a case–control study. Am J Psychiatry 2000;157:
1291 – 6.
[5] Ricciardelli LA, McCabe MP. Psychometric evaluation of the Body
Change Inventory: an assessment instrument for adolescent boys and
girls. Eat Behav 2002;2:1 – 15.
[6] Smolak L, Murnen SK, Thompson JK. Sociocultural influences and
muscle building in early adolescent boys. Psychology of Men and
Masculinity 2005;6:227 – 39.
[7] Cusumano DL, Thompson JK. Body image and body shape ideals in
magazines: exposure, awareness and internalization. Sex Roles
1997;37:701 – 21.
 S. Cahill, A.J. Mussap / Journal of Psychosomatic Research 62 (2007) 631 – 639
[8] Botta RA. For your health? The relationship between magazine
reading and adolescents’ body image and eating disturbances. Sex
Roles 2003;48:389 – 99.
[9] Harrison K. Television viewing, fat stereotypical, body shape stand-
ards, and eating disorder symptomatology in grade school children.
Communic Res 2000;27:617 – 40.
[10] Botta RA. Television images and adolescent girls’ body image
disturbance. J Commun 1999;49:22 – 41.
[11] Hatoum IJ, Belle D. Mags and abs: media consumption and bodily
concerns in men. Sex Roles 2004;51:397 – 407.
[12] 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;21:1 – 16.
[13] Monteath SA, McCabe MP. The influence of societal factors on
female body image. J Soc Psychol 1997;137:708 – 27.
[14] Henderson-King E, Henderson-King D. Media effects on women’s
body esteem: social and individual difference factors. J Appl Soc
Psychol 1997;27:399 – 417.
[15] Posovac HD, Posovac SS, Posovac EJ. Exposure to media images of
female attractiveness and concern with body weight among young
women. Sex Roles 1998;38:187 – 201.
[16] Levine MP, Smolak L, Hayden H. The relation of sociocultural factors
to eating attitudes and behaviours among middle school girls. J of
Early Adolescence 1994;14:472 – 91.
[17] Lorenzen LA, Grieve FG, Thomas A. Exposure to muscular
male models decreases men’s body satisfaction. Sex Roles
2004;51:743 – 8.
[18] Grogan S, Williams Z, Conner M. The effects of viewing same
gender photographs on body esteem. Psychol Women Q
1996;20:569 – 75.
[19] Agliata D, Tantleff-Dunn S. The impact of media exposure on males’
body image. J Soc Clin Psychol 2004;23:7 – 22.
[20] Hoek JW, van Hoeken D. Review of prevalence and incidence of
eating disorders. Int J Eat Disord 2003;34:383 – 96.
[21] Durkin SJ, Paxton SJ. Predictors of vulnerability to reduced body
satisfaction and psychological well-being in response to exposure to
idealized female media images in adolescent girls. J Psychosom Res
2002;53:995 – 1005.
[22] Stice E, Agras WS. Predicting onset and cessation bulimic behaviors
during adolescence: a longitudinal grouping analysis. Behav Ther
1998;29:257 – 76.
[23] Abell SC, Richards MH. The relationship between body shape
satisfaction and self-esteem: an investigation of gender and class
differences. J Youth Adolescence 1996;25:691 – 703.
[24] Tiggemann M, McGill B. The role of social comparison in the effect
of magazine advertisements on women’s mood and body dissat-
isfaction. J Soc Clin Psychol 2004;23:23 – 44.
[25] Posovac HD, Posovac HD, Weigel RG. Reducing the impact of media
images on women at risk for body image disturbance: three targeted
interventions. J Soc Clin Psychol 2001;20:324 – 40.
[26] Agliata D, Tantleff-Dunn S. The impact of media exposure on males’
body image. J Soc Clin Psychol 2004;23:7 – 22.
[27] Zalta AK, Keel PK. Peer influence on bulimic symptoms in college
students. J Abnorm Psychol 2006;115:185 – 9.
[28] Garner DM. Eating Disorder Inventory-2 professional manual. Odessa
(FL)7 Psychological Assessment Resources, 1991.
[29] Thompson JK, van den Berg P, Roehrig M, Guarda AS, Heinberg LJ.
The Sociocultural Attitudes Towards Appearance Scale-3 (SATAQ-3):
development and validation. Int J Eat Disord 2004;35:293 – 305.
639
[30] Fisher E, Dunn M, Thompson JK. Social comparison and body image:
an investigation of body comparison processes using multidimen-
sional scaling. J Soc Clin Psychol 2002;21:566 – 79.
[31] Rosenberg M. Society and the adolescent self-image. Princeton (NJ)7
Princeton University Press, 1965.
[32] Beck AT, Steer RA, Brown GK. Manual for the Beck Depression
Inventory, 2nd ed. San Antonio7 Psychological, 1996.
[33] Campbell JD, Trapnell PD, Heine SJ, Katz IM, Lavallee LF, Lehman
DR. Self-concept clarity: measurement, personality correlates and
cultural boundaries. J Pers Soc Psychol 1996;70:141 – 56.
[34] Thompson J, Pasman L. The Obligatory Exercise Questionnaire.
Behavior Therapist 1991;14:137.
[35] Garner DM, Olmstead M, Polivy J. Development and validation of a
multi-dimensional Eating Disorder Inventory for anorexia nervosa and
bulimia. Int J Eat Disord 1983;2:15 – 34.
[36] Cronbach LJ, Furby L. How should we measure bchangeQ—or should
we? Psychol Bull 1970;74:68 – 80.
[37] Baron RM, Kenny DA. The moderator–mediator variable distinction
in social psychological research: conceptual, strategic, and statistical
considerations. J Pers Soc Psychol 1986;51:1173 – 82.
[38] Thompson JK, Stice E. Thin-ideal internalization: mounting evidence
for a new risk factor for body-image disturbance and eating pathology.
Curr Dir Psychol Sci 2001;10:181 – 3.
[39] Stice E, Shaw HE. Adverse effects of the media portrayed thin-ideal
on women and linkages to bulimic symptomatology. J Soc Clin
Psychol 1994;13:288 – 308.
[40] Polivy J, Herman CP. Causes of eating disorders. Annu Rev Psychol
2002;53:187 – 213.
[41] Waller G, Barnes J. Preconscious processing of body image cues
Impact on body percept and concept. J Psychosom Res 2002;
53:1037 – 41.
[42] Stice E. Review of the evidence for a sociocultural model of bulimia
nervosa and an exploration of the mechanisms of action. Clin Psychol
Rev 1994;14:633 – 61.
[43] Conner M, Johnson C, Grogan S. Gender, sexuality, body image and
eating behaviours. J Health Psychol 2004;9:505 – 15.
[44] Borchert J, Heinberg L. Gender schema and gender role discrepancy
as correlates of body image. J Psychol 1996;130:547 – 59.
[45] Hospers HJ, Jansen A. Why homosexuality is a risk factor for eating
disorders in males. J Soc Clin Psychol 2005;24:1188 – 201.
[46] Lundgren JD, Anderson DA, Thompson JK. Fear of negative
appearance evaluation scale: psychometric properties and validation.
Eat Behav 2004;5:75 – 84.
[47] Maida DM, Lee Armonstrong S. The classification of muscle
dysmorphia. International Journal of Men’s Health 2005;4:73 – 91.
[48] Fernandez-Aranda F, Aitken A, Badia A, Gimenez L, Solano R,
Collier D, Treasure J, Vallejo J. Personality and psychopathological
traits of males with an eating disorder. Eur Eat Disord Rev 2004;
12:367 – 74.
[49] Cafri G, van den Berg P, Thompson J. Pursuit of muscularity
in adolescent boys: relations among biopsychosocial variables
and clinical outcomes. J Clin Child Adolesc Psychol 2006;
35:283 – 91.
[50] Hargreaves D, Tiggemann M. Longer-term implications
of responsiveness to bthin-idealQ television: support for a cumulative
hypothesis of body image disturbance? Eur Eat Disord Rev
2003;11:465 – 77.
[51] Cohane GH, Pope HG. Body image in boys: a review of the literature.
Int J Eat Disord 2001;29:373 – 9.