Brief research report
Women’s exposure to thin-and-beautiful media images:
body image effects of media-ideal internalization
and impact-reduction interventions
Yuko Yamamiyaa, Thomas F. Cashb,*, Susan E. Melnykb, Heidi D. Posavacc,
Steven S. Posavacd
aUniversity of South Florida, Tampa, FL, USA
bDepartment of Psychology, Old Dominion University, Norfolk, VA 23529, USA
cRochester, NY, USA
dUniversity of Rochester, Rochester, NY, USA
Received 18 June 2004; received in revised form 1 November 2004; accepted 1 November 2004
Exposure tomedia images of thin-and-beautiful women negativelyaffects the body image and mood states of youngwomen.
However, not all women are equally susceptible to these effects. The present experimental investigation with 123 young college
of two brief interventions (i.e., media literacy information with and without a dissonance-induction procedure). Results
indicated that relative to a control group, the exposure to thin-and-beautiful media images adversely influenced the state body
image of participants with high internalization levels. Media-literacy psychoeducation prior to the media exposure prevented
this adverse effect. Adding a pre-exposure dissonance-induction procedure did not significantly enhance the preventativeeffects
relative to psychoeducation alone. These results and their implications for the treatment and prevention of body image
disturbances are discussed in the context of the empirical literature on the media’s effects on body image.
# 2004 Elsevier B.V. All rights reserved.
Keywords: Body image dissatisfaction; Body image states; Media exposure effects; Thin-ideal internalization; Media literacy; Dissonance
Body image is not strictly a stable trait, but is also a
variable state influenced by specific contextual events
(Cash, 2002a, 2002b), such as exposure to media
images and messages. Researchers have found that in
the United Sates, 94% of female characters in
television programs are thinner than the average
American woman (e.g., Gonzalez-Lavin & Smolak,
1995), with whom the media frequently associate
happiness, desirability, and success in life (Tigge-
mann, 2002). The media alsoexplicitly instruct howto
Body Image 2 (2005) 74–80
* Corresponding author. Tel.: +1 757 683 4439;
fax: +1 757 683 5087.
E-mail address: email@example.com (T.F. Cash).
1740-1445/$ – see front matter # 2004 Elsevier B.V. All rights reserved.
attain thin bodies by dieting, exercising, and body-
contouring surgery, encouraging female consumers to
believe that they can and should be thin. However, an
idealization of thinness is positively correlated with
body image dissatisfaction (Stice, Schupak-Neuberg,
Shaw, & Stein, 1994), which is often accompanied by
social anxiety, depression, eating disturbances, and
poor self-esteem (Cash & Pruzinsky, 2002).
According to Groesz, Levine, and Murnen’s (2002)
meta-analysis, women are significantly more body
dissatisfied after viewing thin-and-beautiful media
images versus average-size, oversize, or nonbody
images. At least two factors account for varying body
image responses to media exposure: (1) level of
internalization of the media-promoted thin-ideal
(Cattarin, Thompson, Thomas, & Williams, 2000;
Halliwell & Dittmar, 2004; Tiggemann & McGill,
2004), and (2) social comparison processes (Heinberg
& Thompson, 1992; Tiggemann & McGill, 2004).
These variables may be viewed within the broader
construct of ‘‘appearance schematicity,’’ or one’s
cognitive structures vis-a `-vis one’s physical appear-
ance, which organize and influence the processing of
self-relevant information (Cash, 2002b; Tiggemann,
2002). Thompson developed the Sociocultural Atti-
tudes Towards Appearance Questionnaire to measure
women’s media-ideal internalization and comparison
(Thompson, van den Berg, Roehrig, Guarda, &
Heinberg, 2004). High-internalization (HI) women
are especially likely to use high profile persons (e.g.,
media images) as upward comparison targets (Hein-
berg & Thompson, 1992) and feel inferior for not
meeting social ‘‘norms’’ of attractiveness.
Media literacy interventions involving critical
analyses of contents of the media messages have
been advocated to prevent internalization and social
comparison processes (Levine & Piran, 2004). In an
experiment by Posavac, Posavac, and Weigel (2001),
women received two types of psychoeducational,
media-literacy information. The ‘‘Artificial Beauty’’
condition argued that media images of females are
inappropriate ‘‘standards’’ because their flawless
looks are created by various techniques, including
make-up and air-brushing. The ‘‘Genetic Realities’’
condition argued that genetics influence body weight/
shape and that most women are biologically predis-
posed to be heavier than women in the media.
Results indicated that exposure to thin-and-beautiful
media images increased women’s weight concerns.
Providing either or both psychoeducational conditions
comparably prevented the effect.
Although the present study was modeled after
Posavac et al. (2001), certain changes were made.
Second, theyused a trait measure of weight concern as
their dependent measure, but we used a state measure
that taps a range of current body image evaluations
and feelings. Whereas the two variables are quite
likely related, a state measure is more appropriate to
capture an immediate effect of media exposure (Cash,
2002a). Third, because the male narrator in their
videotaped materials explicitly urged the audience
not to compare themselves to media images, it is
uncertain whether it was substantive content of the
message that was effectiveor an instructional demand.
Thus,we adapted message contentwithouttheexplicit
demand, using an audiotaped presentation by a male
narrator. In both studies, he was identified as a
psychologist and expert on the topic.
Investigating specific procedures to reduce thin-
ideal internalization, Stice, Chase, Stormer, and Appel
(2001) used an intervention based on dissonance
theory. This framework maintains that when people
have inconsistent cognitions, they experience psycho-
logical discomfort and are motivated to change their
cognitions to restore consistency. Females who had
internalized the thin-ideal were asked to voluntarily
take a stance against thin-ideals by discussing ways to
help adolescent girls avoid internalization. Post-test
data revealed a resultant decrease in both thin-ideal
internalization and body image dissatisfaction.
The two aforementioned studies were the bases for
this experiment. We also examined a potentially
important moderator variable—the disposition for
hypothesized that: (1) without the psychoeducational,
media-literacy information described above, women’s
state body image experiences would be negatively
affected by thin-and-beautiful media image exposure
(Groesz et al., 2002); (2) giving the psychoeduca-
tional, media-literacy information would reduce these
media-exposure effects; (3) having individuals con-
struct arguments against thin-ideal based on the
media-literacy information would also reduce the
effects; (4) adding this dissonance-induction techni-
que to the provision of the psychoeducational
Y. Yamamiya et al./Body Image 2 (2005) 74–8075
information would have a greater effect than the
information alone. In all conditions, only those with a
stronger disposition for internalization and social
comparison were expected to be significantly affected
by the experimental manipulations.
Participants were 123 White female students at Old
Dominion University. They were between 18 and 29
years old (M = 21.4, SD = 2.86) and volunteers for the
study in exchange for extra credit in psychology
courses. White women were studied given that the
less relevant social comparison standards for other
ethnic groups. Participants’ average body mass index
(BMI = kg/m2) was 24.1 (SD = 5.72).
Experimental stimuli were 20 pictures of young
White fashion models taken from various women’s
magazines. Control stimuli were 20 slides of auto-
mobiles. Stimuli were those used by Posavac et al.
time of each slide was 15 s, followed by a 15 s
exposure to a blank slide (i.e., the response interval).
Control information concerned parenting skills and
child behavior management, presented as important
not only for current or future parents but also for
anyone having contact with children. Psychoeduca-
tional information in the experimental conditions
‘‘Genetic Realities’’described above, as adapted from
Posavac et al. (2001). Both experimental and control
messages were presented on a 7 min audiotape.
Sociocultural Attitudes Towards Appearance
Questionnaire-3 (SATAQ-3; Thompson et al., 2004)
The full 30-item scale measures four facets of
societal influences on personal body image standards,
using a 5-point Likert-type scale. This study used the
Internalization–General subscale, which assesses the
extent to which one idealizes (‘‘I would like to look
like ...’’) and compares oneself (‘‘I compare my
appearance ...’’) to movie stars, television, and
magazine models. Based on median-split from a local
sample of 284 college women, participants were
classified as ‘‘high internalization’’ and ‘‘low inter-
nalization’’ (LI) individuals. Internal consistency of
the subscale in this study was .96.
Body Image States Scale (BISS; Cash, Fleming,
Alindogan, Steadman, & Whitehead, 2002)
Using a 9-point Likert-type scale, the 6-item BISS
measures one’s current (‘‘at this moment’’) dissatis-
faction–satisfaction with aspects of his/her physical
appearance. Higher scores indicate a more favorable
body image state. Its internal consistency was .87 in
Evaluation of Educational Audiotape Form
This questionnaire asked about the clarity, impor-
tance, usefulness, and agreeability of the information
either listed the key information in the tape or wrote
persuasive arguments against the media-transmitted
thin-ideal based on the tape’s information. Persuasive
arguments in the dissonance induction were to help a
teenage girl, whom participants knew and caredabout.
They were to help her not to take seriously or be
affected by beautiful, thin models in the media.
In a study several days earlier, participants
completed a demographics form and the SATAQ-3,
embedded among other questionnaires. They were
randomly assigned to one of four conditions: Control-
Info/Control-Slides(n = 30),
Slides (n = 33), Media-Info/Model-Slides (n = 28),
(n = 32). Approximately half of each group scored
above or below the SATAQ-3 median. In this between-
groups design, we did not conduct an immediate
pre-test assessment of state body image satisfaction,
to prevent sensitization or reactivity effects on post-
In the laboratory, small groups of no more than five
(typically only one or two) women participated, to
reduce the likelihood of influences of close physical
Y. Yamamiya et al./Body Image 2 (2005) 74–8076
proximity or overt reactions to the materials. A female
experimenter presented the study as consumer
research on young women’s evaluations of new
educational programs and certain products. They then
received either the psychoeducational or control
information and completed the Evaluation of Educa-
tional Audiotape Form. The latter also required that
they list facts from the tapes. In the dissonance-
induction condition they were to construct persuasive
written arguments, as described above. Participants
were given 5 min for fact listing or argument
construction. Next, under the guise of a ‘‘different
consumer research study,’’ the slides of either fashion
models or automobiles were presented. The experi-
menter gave participants a bogus evaluation ques-
tionnaire to rate how much they liked the products
(models’ clothing or automobiles) to ensure that
participants attended to slide contents and to reinforce
the consumer-research pretense. Finally, participants
completed the BISS, presented as a ‘‘third study.’’
Preliminary data analyses revealed no significant
age or BMI differences across conditions. LI group
scores on SATAQ-3 Internalization were comparable
across conditions, as were the HI group scores.
Planned contrasts tested the study’s four hypoth-
eses within a 2 (high versus low internalization) ? 4
(conditions) general linear model analysis of variance.
We examined post-experimental differences as a
function of condition and internalization level.
Because multiple contrasts were carried out that were
not completely orthogonal, the alpha level for
significance was set at p < .01. Degrees of freedom
for each contrast were 1 and 115. Collectively, the
omnibus test of contrasts among the LI participants
was not significant, F(3, 115) = 1.82, p < .15, partial
h2= .045, but among the HI participants it was
significant,F(3, 115) = 7.90,
that, across conditions, HI participants reported
generally less favorable body image states than did
LI participants, F(1, 115) = 11.31, p < .001, partial
h2= .09. Table 1 provides the means and standard
deviations on the BISS for the cells of the
Internalization ? Condition design.
The test of Hypothesis 1 compared BISS scores of
the Control-Info/Control-Slides group and the Con-
trol-Info/Model-Slides group as a function of inter-
nalization level. Results supported Hypothesis 1; only
among HI participants, BISS scores in the latter group
were significantly lower than their counterparts in the
ence = 2.03, p < .001; effect size, d = 1.25). This
confirms the adverse impact of the exposure to thin/
attractive images on body image for HI women.
The test of Hypothesis 2 compared BISS scores
from the Control-Info/Model-Slides group with those
fromthe Media-Info/Model-Slidesgroup asa function
of internalization level. Hypothesis 2 was supported;
only among HI participants, BISS scores of those in
Media-Info/Model-Slides group were significantly
more positive than scores of those in Control-Info/
Model-Slides group (M difference = ?2.42, p < .001;
effect size, d = ?1.49). This confirms that the
provision of media-literacy information prior to
exposure to thin/attractive images reduced their
adverse effect on body image among HI women.
To test Hypothesis 3, BISS scores of the Control-
Info/Model-Slides group and the Media-Info/Disso-
nance-Induction/Model-Slides group were compared
within internalization level. Supporting Hypothesis 3,
BISS scores for the latter group were significantly
more positive than those for the Control-Info/Model-
p < .001,partial
Y. Yamamiya et al./Body Image 2 (2005) 74–8077
The BISS means and standard deviations of participants with low and high internalization levels in four conditions
ConditionsHigh internalization mean (SD)Low internalization mean (SD)
Control-Info/Control-Slides (n = 30)
Control-Info/Model-Slides (n = 33)
Media-Info/Model-Slides (n = 28)
Model-Slides (n = 32)
Note: The higher the score, the more favorable the body image state.
Slides group (M difference = ?1.46, p < .007; effect
size, d = ?.90), but only when internalization level
was high. Therefore, for these women, the negative
body image impact of thin/attractive image exposure
was also less when the combined psychoeducational
information and dissonance-induction procedures
The test of Hypothesis 4 compared the Media-Info/
Model-Slides group and the Media-Info/Dissonance-
Induction/Model-Slides group. Results failed to
support our hypothesis; the BISS scores of those in
both groups did not differ significantly, regardless of
internalization levels (M difference = .430, p < .414
forLIparticipants;Mdifference = .961,p < .08forHI
participants). Thus, adding the dissonance-induction
activity to the provision of media-literacy information
failed to augment the effect of the latter.
Finally, we conducted analyses to explore reasons
that Hypothesis 4 was not supported. We counted the
number of facts that participants wrote down in each
condition, as well as the number of arguments made
with these facts in the dissonance-induction proce-
dure. On average, the Media-Info/Model-Slides group
stated 7.9 facts (SD = 2.2) from 14 facts given on the
tape, whereas the Media-Info/Dissonance-Induction/
Model-Slides group gave only 3.8 facts (SD = 2.1).
Significantly fewer facts were conveyed in the latter
condition, F(1, 56) = 54.79, p < .001. Furthermore,
dissonance induction produced an average of only 3.2
arguments (SD = 2.1) using those facts. Thus, while
creating arguments as requested, they did not reiterate
as much of the ‘‘discounting’’ information as did
participants in the other condition.
The study’s results indicate that even a 5 min
exposure to thin-and-beautiful media images results in
a more negative body image state than does exposure
women with high media-ideal internalization levels
and social comparison tendencies. The adverse effect
of the media exposure was significantly reduced
among high-internalization women when they were
given media-literacy information and either asked to
recall and write down the information or induced to
make written arguments against the media’s thin-
ideals based on the information. Low-internalization
women were unaffected by any of the study’s
We failed to find the addition of a dissonance-
induction procedure to be more beneficial than media-
literacy information alone. Our data suggest some
plausible explanations for the null finding. The
dissonance-induction procedure led to only about
three arguments from a reiteration of fewer facts than
were listed in the media-literacy condition alone.
Thus, the addition of dissonance induction, while
comparably effective, may have been less involving
than it should have been. For example, a stronger
induction might entail asking participants first to list
all the facts and then construct at least 10 arguments
that derive from them, and/or make these arguments
aloud to another person in a role-playing context
(Stice et al., 2001). Private delineation of arguments in
a research context may be weaker than a ‘‘voluntary’’
media-literacy information per se maximally pre-
vented the impact of media exposure for the high
internalization participants and could not have been
improved upon. The mean body image level for
control participants was 5.53 and that of the media-
literacy group after media-ideal exposure was slightly
even more positive (5.92).
We used 20 model slides as stimuli, presented for a
total duration of 5 min. Groesz et al. (2002) found that
as the number of stimuli exceeded 10, viewers were
somewhat less influenced, probably due to habitua-
tion. Still, we clearly confirmed the significance of the
exposure effect among high internalization women.
Moreover, although the media-literacy intervention
effectively prevented negative effects of media
exposure, such brief interventions would be unlikely
to have lasting power much beyond the study’s
duration, just as acute exposure effects dissipate.
Womenwith a strong drivefor thinness experience not
only body dissatisfactionbutalso negativeaffect for at
least 2 hours after exposure to idealized media images
(Hausenblas, Janelle, Gardner, & Focht, 2004).
Nevertheless, media exposure in daily life is
considerable and effects may be cumulative. One
prospective study (Hargreaves & Tiggemann, 2003)
concluded that girls who are acutely affected most
over a 2-year period.
Y. Yamamiya et al./Body Image 2 (2005) 74–8078
A number of variables may moderate or perhaps
mediate the impact of media exposure to extreme
ideals of beauty and thinness (Tiggemann & McGill,
2004)—for example, media-ideal internalization,
drive for thinness, social and physical comparison
tendencies, and appearance schematicity. All these
dimensions are facets of the overriding body image
investment construct (Cash, 2002b; Cash et al., 2004),
which refers to the cognitive and behavioral impor-
tance that persons place on their appearance,
especially its salience in their sense of identity and
self-worth. Such investment is predictive of greater
body image reactivity and instability in daily life
(Cash et al., 2002; Melnyk, Cash, & Janda, 2004).
Relativeto males, females in our society are clearly
more body dissatisfied (Feingold & Mazzella, 1998;
Striegel-Moore & Franko, 2002), partly due to the
strong media emphasis on women’s thinness and
attractiveness. Societal and institutional changes are
needed to de-emphasize unrealistic physical ‘‘stan-
dards.’’ The media could use average-sized models
instead of thin ones, as the former may be equally
effective in ads without adversely affecting the body
image of women with high internalization (Halliwell
& Dittmar, 2004). Regardless of the targeted level of
negative body image prevention and improvement
efforts, from individual interventions (Cash &
Hrabosky, 2004) to institutional and societal change
strategies (Levine & Piran, 2004), such efforts are
imperative in appearance-obsessed societies. As
Levine and Smolak (2001) state, ‘‘mental health
and a variety of professionals ... must work together
to extend, integrate, apply, and carefully evaluate all
... models of prevention’’ (p. 256).
Cash, T.F. (2002). Beyond traits: Assessing body image states. In T.
F. Cash & T. Pruzinsky (Eds.), Body image: A handbook of
In T. F. Cash & T. Pruzinsky (Eds.), Body image: A handbook of
theory, research, and clinical practice (pp. 38–46). New York:
Cash, T. F., & Hrabosky, J. I. (2004). The treatment of body-image
disturbances. In J. K. Thompson (Ed.), Handbook of eating
disorders and obesity (pp. 515–541). New York: Wiley.
Cash,T. F., & Pruzinsky, T. (Eds.). (2002).Body image: A handbook
of theory, research, and clinical practice. New York: Guilford
Cash, T. F., Fleming, E. C., Alindogan, J., Steadman, L., & White-
head, A.(2002).Beyond body image asa trait: The development
and validation of the Body Image States Scale. Eating Disor-
ders: The Journal of Treatment and Prevention, 10, 103–113.
Cash, T. F., Melnyk, S. E., & Hrabosky, J. I. (2004). The assessment
of body image investment: An extensive revision of the Appear-
ance Schemas Inventory. International Journal of Eating Dis-
orders, 35, 305–316.
Body image, mood, and televised images of attractiveness: The
role of social comparison. Journal of Social and Clinical
Psychology, 19, 220–239.
Feingold, A., & Mazzella, R. (1998). Gender differences in body
image are increasing. Psychological Science, 9, 190–195.
Gonzalez-Lavin, A., & Smolak, L. (1995, March). Relationships
between television and eating problems in middle school girls.
Paper presented at the meeting of the Society for Research in
Groesz, L., Levine, M. P., & Murnen, S. K. (2002). The effect of
experimental presentation of thin media images on body satis-
faction: A meta-analytic review. International Journal of Eating
Disorders, 31, 1–16.
model’s body size on women’s body-focused anxiety and adver-
tising effectiveness. Journal of Social and Clinical Psychology,
Hargreaves, D. A., & Tiggemann, M. (2003). Longer-term implica-
tions of responsiveness to thin-ideal television: Support for a
cumulative hypothesis of body image disturbance. European
Eating Disorders Review, 11, 465–477.
Hausenblas, H. A., Janelle, C. M., Gardner, R. E., & Focht, B. C.
(2004). Viewing physique slides: Affective responses of women
at high and low drivefor thinness. Journal of Social and Clinical
Psychology, 23, 45–60.
Heinberg, L. J., & Thompson, J. K. (1992). Social comparison:
Gender, target importance ratings, and relation to body image
disturbance. Journalof SocialBehaviorand Personality, 7, 335–
Levine, M. P., & Piran, N. (2004). The role of body image in the
prevention of eating disorders. Body Image: An International
Journal of Research, 1, 57–70.
Levine, M. P., & Smolak, L. (2001). Primary prevention of body
image disturbances and disordered eating in childhood and early
adolescence. In J. K. Thompson & L. Smolak (Eds.), Body
image, eating disorders, and obesity in youth: Assessment,
prevention, and treatment (pp. 237–260). Washington, DC:
American Psychological Association.
Melnyk, S. E., Cash, T. F., & Janda, L. H. (2004). Body image ups
and downs: Prediction of intra-individual level and variability of
women’s daily body image experiences. Body Image: An Inter-
national Journal of Research, 1, 225–235.
Y. Yamamiya et al./Body Image 2 (2005) 74–8079
impact of media images on women at risk for body image
disturbance: Three targeted interventions. Journal of Social
and Clinical Psychology, 20, 324–340.
Stice, E., Chase, A., Stormer, A., & Appel, A. (2001).A randomized
trial of a dissonance-based
program. International Journal of Eating Disorders, 29,
Stice, E. M., Schupak-Neuberg, E., Shaw, H. E., & Stein, R. I.
(1994). Relation of media exposure to eating disorder sympto-
Abnormal Psychology, 103, 836–840.
Striegel-Moore, R. H., & Franko, D. L. (2002). Body image issues
among girls and women. In T. F. Cash & T. Pruzinsky (Eds.),
Body image: A handbook of theory, research, and clinical
practice (pp. 183–191). New York: Guilford Press.
Thompson, J. K., van den Berg, P., Roehrig, M., Guarda, A. S., &
Heinberg, L. S. (2004). The Sociocultural Attitudes Towards
Appearance Scale-3 (SATAQ-3): Development and validation.
International Journal of Eating Disorders, 35, 293–304.
Tiggemann, M. (2002). Media influences on body image develop-
ment. In T. F. Cash & T. Pruzinsky (Eds.), Body image: A
handbook of theory, research, and clinical practice (pp. 91–98).
New York: Guilford Press.
Tiggemann, M., & McGill, B. (2004). The role of social comparison
in the effect of magazine advertisements on women’s mood and
body dissatisfaction. Journal of Social and Clinical Psychology,
Y. Yamamiya et al./Body Image 2 (2005) 74–80 80