Skinny Like You: Visual Literacy, Digital Manipulation and Young Women’s Drive to be Thin

[Timothy Gangwer]

Numerous studies have established a relationship between thin ideal media content and disordered eating

patterns in women. Many of the images viewed in the media that endorse the thin ideal are digitally

manipulated or computer-created. This experiment compared college women’s visual literacy–defined in

terms of their knowledge of digital manipulation in fashion and entertainment images–to their desire

to be thin, their desire to look like the model shown and four disordered eating subscales. Visual literacy

did not reduce participants’ desire to look like the model seen; however, entertainment media was a significant

predictor of greater body image distortion across the sample.

Introduction

A recent People magazine article highlighted the dieting secrets of the stars in the weeks prior to the 2004

Academy Awards show. One star bragged that starving herself for the week prior to the awards show

was the only way she could “look good,” and another star admitted to eating mostly proteins and egg

whites and exercising “like crazy” two weeks prior to the event (People Magazine, March 15, 2004). Several

articles published during the same week in People, Teen, touted similar messages: If you need to look good

for a special occasion, take drastic steps to drop those “last few pounds.”

There is certainly no shortage of tips for dieting, fitness and fashion all related to weight loss in fashion

and entertainment magazines targeted toward adolescent girls and young women. While these types of

tips inundate the media market, a few other types of media campaigns have begun to tout very different

messages (Media Awareness Network, March 28, 2004). These more recent public relations campaigns

promoting self-awareness and self-acceptance are targeting younger, more impressionable viewers who

may easily get duped into believing that starving themselves and fanatic exercising are the only way to

achieve the desired look. For example, the “Free to Be Me” program with 5th and 6th grade Girl Scouts

was used as a means of outreach to teach girls about the media’s influence on body image. The rationale

behind many of these campaigns is that if young girls can be taught to happy with their own body shape

and size, they will be more likely to have success academically, socially and athletically.

Despite the positive attributes of these campaigns, young women and girls are saturated with media

messages touting the rewards of being thin. Furthermore, because the body shape standard in the

media young women pay the most attention to largely represents a single body type (Bissell & Zhou,

2004), it becomes difficult for young women to be confident and comfortable with their body shape

and size. With eating disorders on the rise and an increasingly large number of young women and

girls being dissatisfied with their body shape and size, it is important to find ways to help these

young women be secure and comfortable with their own bodies. In this project, visual literacy is

used as an independent variable to determine if knowledge of digital manipulation in fashion photographs

has a short-term effect on women’s feelings about their bodies. While it is widely known that

many images, especially those in advertising, are manipulated or retouched digitally, the question is

whether the viewers of these messages process that knowledge as the images are interpreted and read.

Digital Manipulation of Media Images

Studies of body shape types represented in entertainment media indicate the body shape standard for

women has increasingly become thinner (Silverstein, Perdue, Peterson, & Kelly, 1986; Garner,

Studies in Media & Information Literacy Education, Volume 6, Issue 1 (February 2006), 1–14

# University of Toronto Press. DOI: 10.3138/sim.6.1.002

Garfinkel, Schwartz, & Thompson, 1980); however, what is not clear is whether the body shape of

models has really become thinner or if advances in digital technologies have enhanced the body

shape and size of models in advertisements and editorial content. Certainly computer doctoring of

photographs has become much easier in recent years (Wheeler, 2002). Wheeler reports of Newsweek

and Time magazine covers of the McCaughey septuplets’ parents in December 1997. Technicians in

the Newsweek lab straightened McCaughney’s teeth for the cover shot, while Time ran the shot of the

parents untouched. In news photographs, the expectation is that photographs are “real” and not

manipulated unless otherwise specified in the cutline as a “photo illustration” (Wheeler, 2002,

pg. 118). However, the perception of truth in editorial content in magazines has become fuzzier.

Wheeler says, “In fashion, drastically manipulated photography is taken for granted by art directors

and editors, and editorial layouts are sometimes barely distinguishable from advertising spreads”

(Wheeler, 2002, pg. 121).

Visual Literacy and Visual Persuasion

Messaris argues that visual literacy is a prerequisite for the proper comprehension of visual media and

defines visual literacy as the “familiarity with visual conventions that a person acquires through cumulative

exposure to visual media” (Messaris, 1994, pg. 3). Messaris’ argument, more simply, is that because

of the allure and appeal of visual media, it is very easy for a viewer to become duped by single or multiple

images if that viewer has little understanding of the ways in which visual media can “misinform,

distort, and manipulate” (Messaris, 1994, pg. 2).

. . .visual education might make a viewer more resistant to the manipulations

attempted by TV commercials, magazine advertising, political campaigns, and so

on. In other words, even if learning about the visual devices used in picture-based

media does not have any effect on a viewer’s comprehension of pictures or on

one’s other cognitive abilities, it might still make the viewer more aware of how

meaning is created visually-and therefore less likely to be taken in by abuses of

this process (Messaris, 1994, pg. 3).

Certainly, in the case of advertisements and fashion photographs, we, as viewers, may acknowledge that

images are retouched, but we may not process this information as we view a single image if the message

is not reinforced prior to viewing. Messaris (1997) further suggests that because we often respond to

visual imagery on an affective or emotional level, the imagery evokes a stronger response in us,

which may in turn make the ad or image more persuasive. Messaris asserts that critically processing

an image may require aptitude in general critical thinking, thus it is questionable the degree in which

young girls process knowledge related to the digital manipulation of ads. Specifically, do the powerful

messages promoting the thin-ideal trump all knowledge of digital manipulation in fashion magazines?

Because answers to these questions are largely unknown, one goal of this study is to determine how

or if women process visual literacy statements that help cue them to process visual images in a particular

way. Thus, the first series of hypotheses is advanced:

H1: Exposure to a visual literacy statement will be related to decreased levels of body image distortion in

women.

H1b: Exposure to a visual literacy statement will reduce female respondents’ desire to look like the

model shown, when controlling for exposure to entertainment media.

The Promotion of the Thin Ideal in Advertising

Where do young women learn about the thin ideal and from what sources do young women feel the

most pressure to be thin? Many scholars studying the social effects of mass media as it relates to

body image distortion in women suggest the media are at least partially responsible for promoting

thin-ideal body types as the norm for women, and to a large degree, this representation of a single

Studies in Media & Information Literacy Education, Volume 6, Issue 1 (February 2006), 1–14

# University of Toronto Press. DOI: 10.3138/sim.6.1.002

2 ideal for women is seen in advertising. Furthermore, because many researchers suggest that editorial

content nowadays is so similar to advertising content, it has becoming even more difficult for

readers and viewers to make distinctions between the two. This is especially relevant for younger

girls who read teen magazines and may not be all that knowledgeable about influences of advertising.

Kilbourne (1999) says, “Far from being a passive mirror of society, advertising is an effective and pervasive

medium of influence and persuasion, and its influence is cumulative, often subtle, and primarily

unconscious” (pg. 67). It is for this reason, and many others, that researchers have been concerned with

the potential link between advertising and disordered eating in women. Pollay (1986) reports that advertising

has been accused of unintentionally imposing a “sense of inadequacy” on women’s self-concepts

because advertising may play a role in reinforcing a preoccupation with physical attractiveness (pg. 34).

Furthermore, because women and girls are exposed to so many ads throughout the course of the day,

the repeated promotion of the thin ideal becomes a message that is difficult to escape from.

Hesse-Biber (1989) and Strober (1986) report that chronic dieting is a direct result of the social pressure

on American women to be unnaturally thin. Stephens and Hill (1994) argue that “advertising has been

vilified for upholding-perhaps even creating-the emaciated standard of beauty by which girls are taught

from childhood to judge the worth of their own bodies” (pg. 137). Cash and Hicks (1990) report that

women who are dissatisfied with their body shape and size also tend to have lower levels of self-esteem

and perceive their social well being to be worse than others. While research has established a very clear

link between exposure to TDP media and body image distortion in women and girls, what isn’t known is

the degree in which women and young girls recognize that many images in fashion and entertainment

magazines are digitally manipulated or crafted by a computer. Subsequently, the question this study

addresses is, does visual literacy, defined in this study as knowledge of digital manipulation of a

fashion photograph, reduce the likelihood of social comparison or the likelihood of a subject to idealize

an unrealistic image? Theorists suggest this phenomenon of body shape comparison between self and

mediated images is a result of social comparison, a means of evaluating oneself based on a self-comparison

to others. Along these lines, cultivation theory allows us to understand how women and

girls’ beliefs about an ideal body image have been cultivated over time through repeated exposure to

thin ideal media content.

Theoretical Foundation

Social comparison theory has been used in recent mass communication research examining the social

effects of mass media as it relates to body image distortion in women and young girls. As Festinger

(1954) and Goethals (1986) suggest, social comparison theory provides a theoretical framework for

understanding the way people look to media images they perceive to be realistic and attainable and

make comparisons between themselves, others and idealized images. Festinger suggests that because

humans are motivated to improve themselves, they look to models, sometimes in the media, that

would aid them in their drive for improvement. Wood and Taylor assert that when women make comparisons

between themselves and idealized images, their beliefs about the importance of being thin are

confirmed, and they become motivated to achieve that goal. Therefore, as women read a fashion magazine,

they compare their looks to the looks of models in editorial content and in advertisements, and

when they perceive a discrepancy between the two, they do whatever they can to narrow the gap

between their own image and the idealized image.

The premise of Festinger’s argument related to social comparison is that social comparison will only

occur when comparisons are made to similar others-i.e., those who are similar with respect to skin

color, age, status, abilities and opinions. Suls (1977) advanced this line of thinking by suggesting that

comparisons with similar others make it easier for an individual to assess herself/himself and determine

what the gap is between the self and the target other. Suls also says that the similarity between the self

and the target other will allow an individual to evaluate his or her own position, in short, engaging in

self-evaluation. Thus, if an individual perceives a model in an advertisement to be similar, it is more

likely the individual will engage in the social comparison process to determine how he or she compares.

Studies in Media & Information Literacy Education, Volume 6, Issue 1 (February 2006), 1–14

# University of Toronto Press. DOI: 10.3138/sim.6.1.002

3

One point this study addresses is will the individual be as likely to engage in social comparison if she

feels she is dissimilar to the media model based on reading a statement about the validity of an image.

In short, does reading a statement about digital manipulation affect the processing of an image on a

short-term basis? The important issue here is that women and girls may be engaging in social comparison

with images they perceive to be real but are in fact manipulated or manufactured.

Major, Testa, and Blysma argue that as women strive for self-improvement, they look to superior others

as the ideal target for comparison. When this happens, upward social comparison occurs, which in turn

can have a negative effect on mood and self. But, what is not known is if individuals are making upward

or downward comparisons with the target model, and this type of comparison may be determined based

on similarity. As more studies have been carried out in this area of study, researchers have gradually realized

that upward and downward comparisons are not intrinsically linked to a particular effect–on the

contrary, “either direction has its ups and downs.” Specifically in upward comparison, the presence of

superb others may have at least two implications: First, you are not as well off as the superb others;

second, it is possible that you can be one of them. To individuals who feel optimistic about narrowing

the discrepancy revealed by comparison, the comparison process may be inspiring for making progress.

Myers and Biocca proposed that the feeling of self-control is a potential factor to explain the positive

effects of upward comparison found in their study, in which participants exposed to body image commercials

evaluated their own body image more favorably. The researchers suggested that those participants

might view the ideal images as attainable and within reach. Thus, participants may translate

commercials’ meaning of “you can be thin” into “I am getting thin” and even into “I will be thin.”

Bower expressed a similar sentiment: “The more easily improved a body part is perceived to be, the

less likely a comparer is to feel negatively as a result of the comparison.” On the contrary, individuals

who do not hold such optimism may experience a negative effect and self-deflation following

upward comparison. To them, the presence of better off others only highlights their own weakness

that is considered to be unchangeable.

Social comparison plays an important role in the development of body self esteem and body satisfaction/

dissatisfaction because through continual comparisons with target others, individuals are reminded

that they are not living up to the ideal as presented to them in the media. The process of comparison

over the course of time, arguably, is occurring with the help of cultivation. Gerbner (1969, 1998) argues

that television plays an important role in the development of perceptions about social reality through

repeated and long-term exposure to the mediated world presented to us. The underlying assumption

of cultivation theory is that through repeated exposure to “consistent media portrayals,” i.e. the portrayal

of an ideal body type for women, television viewers are more likely to adopt the beliefs, ideologies, and

perceptions that are similar to the television world. Certainly in the representation of women’s body

types found in the media, we can argue that the mediated portrayal has never been parallel to the

“real world.” Subsequently, as women and girls are exposed to these images repeatedly and over the

course of the lifespan, they could be in a position where they are continually comparing themselves

to the ideal other, which may leave them feeling dissatisfied with their own body shape and size.

The issues related to body image distortion, the media’s role in this phenomenon and the potential

for some awareness to moderate the more harmful effects of the media led to the following series of

research questions and hypotheses:

RQ1: Does exposure to a visual literacy statement affect feelings of similarity to the model shown?

RQ2: Does exposure to a visual literacy statement affect evaluations of thinness and attractiveness to

each model?

H2: Respondents with a lower degree of similarity to the model shown will be more likely to have

greater body image distortion.

Social comparison theory can also be used to understand the relationship between exposure to media

content and attitudes based on that exposure with regard to race. Social comparison theory suggests

Studies in Media & Information Literacy Education, Volume 6, Issue 1 (February 2006), 1–14

# University of Toronto Press. DOI: 10.3138/sim.6.1.002

4

that through the prevalence of thin females in the media, young women may attempt to model what they

see because they are presented with the ideal of a thin body and compare their body shape to what’s seen

in the media. However, it is possible that comparisons may not take place if the media model is not

similar in race to the individual. As McGuire and Pitts, Whalen, O’Keefe and Murray have suggested,

ads tends to be the most effective when the characters, values and symbols in the ad are drawn from the

intended audience’s cultural environment. Several studies published in the last decade support the

finding that the mass media having a greater influence on White women than non-White women as

it relates to body image distortion. Even though improvements have been made in the representation

of non-white women in primetime television, many of the dominant female characters in primetime

programs are Caucasian.

Based on the literature in this area, the following hypothesis is advanced:

H3: Black respondents will have lower degrees of body image distortion, independent of exposure to the

visual literacy statement.

Method

Participants

Participants were 64 freshmen, 27 sophomore, 22 junior and 11 senior women enrolled in undergraduate

mass communication and journalism classes at a university in the southeast (N 1/4 124). Of the 124

participating women, 72 percent were White, 21 percent were Black, and 7 percent were multiracial or

Hispanic. Participants in this project ranged in age from 18–24.

Design

Students participating in this project were randomly selected to be in one of three groups as part of an

experiment testing the effectiveness of a visual literacy statement in the processing of thin-ideal media

images. In this study, two control groups and one experimental group were used. One objective for this

project was to determine how or if reading a visual literacy statement affected responses to the stimulus.

Thus, the experimental group (group 1) was exposed to the stimuli (see below for description) and read

the visual literacy statement plus answered questions related to other measures. The first control group

(group 2) received an instrument containing all items from the group 1 instrument except the visual

literacy statements. A second control group (Group 3) was used so that a solid basis of comparison

for the other dependent variables could be determined. In short, participants in the second control

group did not read a visual literacy statement nor did their instrument contain the visual stimulus.

Group 3 responses were used as a basis for comparison for the other dependent variables because

the pre-test indicated exposure to the stimulus did predict higher scores on the four disordered

eating subscales.

The experimental group (Group 1; N 1/4 40) received an instrument with questions related to media

exposure, media use, and disordered eating symptomatology. Subjects selected for the experimental

group (Group 1) and one control group (Group 2) were shown three images of swimsuit models

and were asked a series of questions related to each model’s attractiveness, thinness, their feelings of

similarity to the model shown, and the desire to look like the model shown. Participants in the experimental

group (Group 1) read a visual literacy statement prior to viewing any photographs. Group 2

(N 1/4 41) participants received the same instrument but did not read a visual literacy statement.

Group 3 (N 1/4 43) participants received an instrument without a visual literacy statement and

without seeing images of swimsuit models (stimulus). However, all media exposure, media use and

disordered eating symptomatology questions were the same regardless of the group. A pre-test was conducted

on the instrument to test for question ordering, responses to the stimuli, and to test the wording

of the literacy statement.

Studies in Media & Information Literacy Education, Volume 6, Issue 1 (February 2006), 1–14

# University of Toronto Press. DOI: 10.3138/sim.6.1.002

5

Procedures

Stimulus. Participants in the experimental group (Group 1) read the following statement prior to viewing

images of swimsuit models:

As you may know, many images seen in fashion and entertainment magazines have

been digitally manipulated and/or altered to enhance the overall look of the models

appearing in the ad. This happens for male and female models. In some cases,

“models” have been created using digital software to create the “perfect” female

or male. In many cases, photo-editing software is used to enhance a model’s body

shape or size and to eliminate physical imperfections.

The next page of the instrument contained a photograph of a swimsuit model with a second statement

related to the digital manipulation of the image:

The image below has been digitally manipulated to enhance the model’s appearance.

Participants read the statement above and then were shown a 7 by 9 inch image of a swimsuit model.

Participants then answered a series of questions related to their perceptions of the model’s thinness and

attractiveness and the degree of similarity they felt toward the model shown. Participants were also asked

to report how much they would like their body shape to resemble that of the model’s shown. Participants

viewed three photographs of swimsuit models, each with the digital manipulation statement above it, and

then answered the same series of questions related to each model. An earlier pre-test was conducted to

determine which of the images of swimsuit models would be most appropriate for the study. In the pretest,

20 women evaluated the thinness, attractiveness, and feelings of similarity to nine swimsuit models.

Within this pool, two models were African American, 1 was Hispanic and another was Asian American.

Mean responses for attractiveness and thinness were tabulated and then used to divide the models into

three groups. Since evaluations of thinness and attractiveness are quite subjective, the goal was to select

three photographs of models that yielded different responses which so that the images of models chosen

for the final instrument would be more heterogeneous and would represent a variety for the target other

with regard to attractiveness and thinness.

The independent variable for many of these hypotheses was the exposure to the digital manipulation

statement.

Exposure to entertainment media was other important variable used as an IV or as a control variable in

some of the hypotheses. Several studies examining the social effects of mass media have examined

exposure to print media (magazines) and broadcast media and its relationship to disordered eating.

In this study, analysis of the exposure variables and their potential relationship to disordered eating

was done separately and reported separately for television and magazines, even though the hypotheses

group these exposure measures together. Exposure to entertainment TV was measured by asking

respondents to record the total number of minutes per day they spent viewing entertainment television.

Female respondents reported viewing between 0 and 450 minutes of entertainment television per day

(M 1/4 149.60, SD 1/4 84.43). Exposure to entertainment magazines was measured by asking participants

to record their frequency of reading several types of magazines, using the following responses

(0 1/4 never, 5 1/4 regularly). Participants in this study reported reading on average four magazines a

month (M 1/4 3.61, SD 1/4 3.01) and indicated that most of the magazines read were entertainment

(M 1/4 3.18, SD 1/4 1.34) or fashion magazines (M 1/4 3.75, SD 1/4 1.56).

In order to measure exposure to thin-ideal TV, respondents were asked to indicate how frequently they

watched specific programs by using a 5-point scale (0 1/4 never, 5 1/4 regularly). Forty-seven television

shows airing on six networks-NBC, ABC, CBS, Fox, UPN, and WB-were selected from the fall

2003-spring 2004 season. The programs chosen represented shows in the top 50 viewed by 18–24

year olds, and they represented a diverse sampling of body types. In an earlier pilot study, 20 college

students from another sample used a 1–5 scale (1 1/4 conspicuously thin, 3 1/4 about average, 5 1/4

conspicuously fat) to assign a body shape code to the primary female characters in each of the 47 programs.

Studies in Media & Information Literacy Education, Volume 6, Issue 1 (February 2006), 1–14

# University of Toronto Press. DOI: 10.3138/sim.6.1.002

6

Coders were instructed to list the character or characters they considered to be “primary,” then assign a

body shape code using the 1–5 scale to those primary characters. Using Harrison’s framework of obtaining

an index of thin-ideal television, coders were instructed to consider a body size as “conspicuous” if at least

one character on the show drew attention to her body shape because she was either thin or fat. Intercoder

reliability tests were run on each of the 47 programs used in the pilot study. The overall reliability score was

.87, using Scott’s pi. The body-size rating assigned to each of the 47 programs was used with the frequency

of viewing scores to create a scale representing thin-ideal television viewing. The descriptive results indicate

the respondents did not watch some programs frequently. In data analysis, the 13 most-watched programs

were selected, which translated into selecting shows where the mean frequency of viewing was a 1 or higher

on the 0–4 scale (see Table 1). The 13 programs used in statistical analysis were shows such as Friends, Will

& Grace, America’s Next Top Model and Law & Order: SVU. All 13 programs received a mean thinness code

between 1.13 and 3.01. At that point, each respondent’s self-reported viewing frequency was multiplied by

each show’s thinness code, which resulted in a scale called the frequency of viewing scale. Six of the 13

most frequently viewed programs received a thinness code between 1.13 and 2.66. Using the same computation,

a “thin-ideal TV viewing” index was created.

Dependent Variables. One of the primary dependent variables for this project was eating-disorder symptomatology.

This measure represented four subscales-anorexia, bulimia, drive for thinness, and body dissatisfaction.

The questions on the instrument came from the Eating Attitudes Test and the Eating

Disorders Inventory. The responses to 43 questions on the survey were used to create an additive

scale for each dimension-drive for thinness, body dissatisfaction, bulimia, and anorexia. Responses to

all questions related to the dependent variable ranged from never 1/4 1 to always 1/4 6.

The second series of dependent variables related directly to respondents’ feelings of similarity to the

models shown. Four questions measured respondents’ feelings of similarity to the model shown and

respondents’ desire to be similar to the model shown: How similar do you feel to this swimsuit

model? How much would you like to look like this swimsuit model? How much would you like your

body shape to look like this swimsuit model? How much do you like this swimsuit model?

Responses ranged from 0 (not at all) to 4 (very much). Respondents were also asked to assign a thinness

and attractiveness score to each model. Responses ranged from 0 (not at all attractive) to 9 (extremely

attractive) and 0 (extremely overweight) and 9 (extremely thin).

Results

Participants randomly selected to one of the three experimental groups did not report significant differences

in their exposure to entertainment media, in their age, their weight, their exposure to sports media,

or their race. The only significant difference between the three groups was that participants in the second

control group (Group 3) reported exercising significantly less than participants in first control group

(Group 2) (Group 2 X 1/4 4.34, SD 1/4 1.39; Group 3 X 1/4 3.16, SD 1/4 1.19, p , .05). The television

programs most frequently viewed across groups were Friends, Joan of Arcadia, The OC, That 70s Show,

America’s Top Model, Will & Grace, Everybody Loves Raymond, Everwood, 7th Heaven, The Bachlorette, and all

three Law & Order programs, Law & Order: SVU, Law & Order: Criminal Intent, and Law & Order. There

were no significant differences in viewing patterns for the above shows among the three groups except

for 7th Heaven. Group 3 reported viewing this program significantly more than the other two groups.

Hypothesis 1 predicted that exposure to the visual literacy statement would reduce respondents’ likelihood

to have higher scores on the four disordered eating subscales. ANOVA tests indicate no significant

differences between the three groups’ scores on the bulimia, anorexia, drive for thinness, and body dissatisfaction

scores. While no significant differences emerged, the results were in the predicted direction

between the two groups exposed to the visual stimuli-experimental (Group 1) and the first control group

(Group 2). When looking at responses from these two groups alone, the control group (Group 2) consistently

had higher scores on all four scales, even though these differences were not significant.

Hypothesis 1b predicted that exposure to a visual literacy statement would reduce respondents’ desire to

look like the model shown. In this case, respondents reading the visual literacy statement (Group 1) had

a greater desire to look like the model shown and had a greater desire to have a body similar to each

model’s in all three cases. Subsequently, this hypothesis was also not supported.

RQ1 tested the notion of similarity based on the experimental group each respondent was selected into.

The line of thinking for this research question was to test whether direct and specific knowledge of

digital manipulation-i.e., exposure to the visual literacy statement–affected the way respondents interpreted

or processed images when viewed. If a fashion photograph or fashion advertisement is perceived

as unrealistic or fake, it is possible young women may be less likely to make upward or downward comparisons,

which may in turn affect the way they perceive their own body image. Independent t-tests indicate

no significant differences between the experimental or control group (Groups 1 & 2) in their

feelings of similarity to the model shown and this was found when controlling for exposure to entertainment

television and entertainment and fashion magazines.

RQ2 examined a related question: how does direct and immediate knowledge of digital manipulation in

fashion photographs or advertisements affect immediate processing of the image? In this case, respondents

were asked to evaluate three models on attractiveness and thinness (each separate questions).

Presumably, respondents in the experimental group (Group 1) might not give models as favorable an

evaluation based on the fact they were shown a statement indicating each image was enhanced to

improve the overall appearance of the model. In some instances, significant differences were found

between group 1’s evaluation of each model and group 2’s evaluation of each model. However, the

differences were not in the predicted direction. Model 1 received a mean thinness score of 7.75

(SD 1/4 .89, p , .05) from Group 1 respondents and a mean thinness score of 7.36 (SD 1/4 .87) from

Group 2 respondents. Similar differences were found in the two groups’ evaluations of Model 2’s attractiveness

(Group 1: X 1/4 6.65, SD 1/4 2.21, p , .05; Group 2: X 1/4 5.64, SD 1/4 1.74). While each

group’s evaluation of the three models’ thinness and attractiveness were not significantly different,

quite consistently, group 1 did report higher scores on all evaluations.

Studies in Media & Information Literacy Education, Volume 6, Issue 1 (February 2006), 1–14

# University of Toronto Press. DOI: 10.3138/sim.6.1.002

8

Hypothesis 2 predicted that respondents who reported not feeling similar to the model shown would be

more likely to have greater body dissatisfaction. The line of thinking here was that if respondents felt

similar to the target other, they might be less likely to make comparisons between themselves and

the media model. Similarity was measured on a five-point scale and was tested on all three models.

The variable was then collapsed into a three-point scale representing little similarity to the model

shown (1), some similarity to the model shown (2), a great deal of similarity to the model shown (3).

ANOVA and regression analysis indicate this hypothesis was supported. Respondents with a higher

degree of similarity to the model shown had significantly lower scores on the drive for thinness

scale, the body dissatisfaction scale and the anorexia scale than those who lower degrees of similarity.

For example, respondents classified in the “great deal of similarity” group reported a mean body dissatisfaction

score of 21.70 (SD 1/4 3.52) on a scale of 9–54; respondents classified in the “some similarity

group” reported a mean body dissatisfaction score of 22.15 (SD 1/4 6.47), and respondents in the “little

similarity” group reported a mean body dissatisfaction score of 26.82 (SD 1/4 5.20, p , .01). Significant

differences were also found between all three groups on the drive for thinness and anorexia scales (see

Table 2).

Drive for thinness was a seven-item scale ranging from a low of 7 to a high of 42. Body dissatisfaction

was a nine-item scale ranging from a low of 9 to a high of 54. Anorexia was a seven-item scale ranging

from a low of 7 to a high of 42. In all cases, the higher the number on the scale, the greater the feelings

of drive for thinness, body dissatisfaction or the more likely respondents are to have anorexic tendencies.

Hypothesis 3 predicted that Black respondents would have lower levels of BID than White respondents,

independent of exposure to the visual literacy statement. Regression analysis indicated a significant

difference between Black and White respondents’ scores on all four disordered eating subscales, independent

of exposure to the visual literacy statement (drive for thinness 1/4 .19, p , .05; body

dissatisfaction 1/4 .39, p , .001; anorexia 1/4 .17, p , .05; bulimia 1/4 .22, p , .001). Independent

t-tests also confirmed that Black respondents had significantly lower scores on all four scales.

For example, Black respondents’ mean score on the drive for thinness scale was an 8.15

(SD 1/4 6.58) and White respondents’ mean score on the same scale was 13.67 (SD 1/4 5.89,

p , .001). Both statistical tests indicate this hypothesis was supported.

Drive for thinness was a seven-item scale ranging from a low of 7 to a high of 42. Body dissatisfaction

was a nine-item scale ranging from a low of 9 to a high of 54. Bulimia was a six- item scale ranging from

a low of 6 to a high of 36. In all cases, the higher the number on the scale, the greater the feelings of

drive for thinness, body dissatisfaction or the more likely respondents are to have bulimic tendencies.

Discussion

A great deal of research exists documenting the social advantages of being thin and of being attractive.

Other studies have documented the relationship between media exposure and women’s desire to be thin.

What isn’t known is the degree in which women engage in social comparison when they know the image

they are viewing has been digitally manipulated. This study tested that premise: did knowledge of digital

manipulation affect a respondents’ desire to achieve a particular body shape seen in a photograph?

Evidence collected in this study suggests the answer to this question is no. Respondents with knowledge

Table 2. One-way ANOVA for similarity to model and disordered eating subscales. (N 1/4 124)

Group 1 Group 2 Group 3

Dependent variable Mean (SD) F df p value

DFT* 15.35 (6.71) 10.12 (6.19) 11.40 (5.62) 5.65 74 p , .01

BD** 26.82 (5.20) 22.15 (6.47) 21.70 3.52 6.16 74 p , .01

Anorexia*** 16.18 (5.43) 10.10 (5.99) 14.40 (5.62) 9.34 74 p , .001

*Post-hoc bonferroni tests found significant differences between groups 1 and 2 and groups 1 and group 3.

**Post-hoc bonferroni tests found significant differences between groups 1 and 2 and groups 1 and 3.

***Post-hoc bonferroni tests found significant differences between all three groups.

Studies in Media & Information Literacy Education, Volume 6, Issue 1 (February 2006), 1–14

# University of Toronto Press. DOI: 10.3138/sim.6.1.002

9

of digital manipulation had no less desire to look like a swimsuit model seen in a photograph and did

not have lower scores on any of the disordered eating subscales. Furthermore, after reading a statement

about a fashion model photograph being enhanced, respondents seemed even more likely to assign

higher attractiveness and thinness scores to the three models shown. This suggests some very important

things about the way women actively process images seen in editorial and advertising content and further

suggests the visual literacy statements used in this study were not enough to help young women process

images in a way that is less harmful to their own body image. Kilbourne suggests the processing of

images in advertising can be unconscious, and even though particular information might help

viewers understand the image being viewed is not real, the comparison between oneself and an idealized

media model still seems to take place.

A particularly useful finding in this study is the relationship between knowledge of digital manipulation

and the four disordered eating subscales. While respondents randomly sampled into the experimental

group did have scores in the predicted direction, they were not significantly lower than respondents

in either control group. This suggests a few things about the visual literacy statement and the possible

processing of the statement prior to viewing images of attractive and thin swimsuit models. First, it is

possible some of the respondents failed to read the two statements or failed to pay much attention to it.

This idea does resonate with the way people respond to and process visual media. Newspaper readership

studies indicate many readers look at the dominant images first before reading much of the text. It is

possible respondents in this study did just that. It is also possible respondents did in fact read the statement

but the statement wasn’t written in strong enough language to make much of a difference when

viewing powerful images. One of the primary goals of this study was to determine if visual literacy or

education about the power of visual media could help mediate some of the effects of the same media.

The findings here suggest the stimulus needs to be even more obvious and overt than what was used in

this study.

Another plausible explanation to the lack of support is that the effects of long-term exposure to entertainment

media may be more persuasive in the general attitudes women have about their bodies.

Cultivation theory would support this conclusion. As Gerbner (1998) says, “cultivation is both dependent

on and a manifestation of the extent to which television’s imagery dominates viewers’ sources of

information” (pg. 182). Post-hoc analysis of the data indicated that respondents who had high exposure

to entertainment media were much more likely to have significantly higher scores on all four disordered

eating subscales than those who were grouped into the low exposure category and the medium exposure

category. This finding suggests that exposure to thinness depicting and promoting (TDP) media is still a

crucial risk factor for disordered eating patterns and behavior. Bissell and Zhou found that females who

were exposed to “thin ideal” television scored fairly high on four eating disorder subscales and further

found that the type of entertainment media women reported watching the most frequently was the

strongest predictor of anorexia, bulimia, body dissatisfaction and drive for thinness. This trend

seemed to be especially prevalent for White women.

Table 3. Regression analysis of sports media exposure, frequency of exercise, and the four disordered eating subscales

Drive for thinness

Variable Beta T sig

Exp to men’s & women’s televised sports 2.33 23.79 p , .001

Frequency of exercise 2.28 23.12 p , .01

Body dissatisfaction

Exp to men’s & women’s televised sports 2.33 23.73 p , .001

Frequency of exercise 2.12 21.39 p , .05

Bulimia

Exp to men’s & women’s televised sports 2.32 23.70 p , .001

Frequency of exercise 2.24 22.82 p , .01

Anorexia

Exp to men’s & women’s televised sports 2.21 22.41 p , .01

Frequency of exercise 2.31 23.53 p , .001

Studies in Media & Information Literacy Education, Volume 6, Issue 1 (February 2006), 1–14

# University of Toronto Press. DOI: 10.3138/sim.6.1.002

10

Another important finding here was respondents’ desire to look like the model shown. When asked how

much they would like to look like the model shown and how much they would like their bodies to look

like the model shown, respondents in the experimental group consistently had a greater desire to do so,

even though the differences were not significant. This suggests the visual literacy statement had no effect

at all on the way young women processed the images of swimsuit models. While the findings could be

different if the cell size were increased, it seems as if the problem seems to be more with the placement

of the visual literacy statement and/or the wording of it.

Theoretical Implications

The findings from this study suggest participants in the experimental group may not have processed the

visual literacy statement in the way predicted, as the processing of this information was predicted to decrease

feelings of similarity, decrease social comparisons made with the media models and lower scores on the

disordered eating subscales. The reason for this may be related to the way participants in this study processed

the verbal and visual information. Cook (1992) suggests that when consumers or viewers have to

process visual and verbal components of advertisements or related media, the complete processing of information

may not occur because one form of information, i.e. the visual information, may be more interesting,

dynamic, compelling, or interesting, which would subsequently reduce the likelihood of processing the

textual information. Participants may have engaged in the first of Shimp and Gresham’s (1983) proposed

eight stages of processing an advertisement (e.g. one containing textual and visual information), but that

before any cognitive, affective or behavioral effects are seen, the consumer must attend to the information

and then comprehend it. The comprehension of the verbal information may not have taken place simply

because of the so-called competition with visual information. Along these lines, Messaris (1994; 1997)

points out that the implicit nature of visual imagery is to attract viewer’s attention, and once this has happened,

the imagery can “give rise to some emotional disposition toward the product, politician, social cause,

or whatever else the ad is about.” The visual stimuli may have proved to be so interesting or compelling that

participants in this study failed to really process what was presented to them via text.

As Messaris (1994; 1997) points out, if viewers are attending to the visual imagery more than the text

and then have an emotional response to the image, the effect of the information in textual form may

decline significantly. As respondents are attending to the visual stimuli–images of swimsuit models–

they may be engaging in the social comparison process by assessing themselves compared to the

target other. As this assessment is taking place, participants are making comparisons between themselves

and the target other to determine how similar they are in age, ethnicity, social standing, body shape and

size, attractiveness, etc., and as Milkie (1999) points out, the similarity between oneself and the target

other will affect the degree in which social comparison occurs. Results from this study indicate that participants

in the experimental and control group did not feel altogether similar to any of the three swimsuit

models. Singer (1981) and Rosenberg (1986) suggest that social comparison theories assume some

degree of freedom in choosing a comparison referent, and these target others are chosen in part by

motivations for self-enhancement. If there was no real point of comparison between study participants

and the swimsuit models, as indicated by the similarity scores, we would not expect to find support for

earlier hypothesis predictions. Thus, this study’s findings help us better understand the relationship

between processing textual and visual media, especially with a younger participant pool, and the

results help us better understand how similarity affects persuasive forms of media.

While this study’s findings set the stage for future research, this project is not without its limitations. The

visual literacy statement could certainly be improved. It could also be one reason some of the hypotheses

were not supported. Even though the stimulus was tested in a pilot study, it is possible the placement of

it on the instrument led respondents not to pay attention to it or even read it because there were no other

items on that page needing attention. While the visual literacy statement could be one variable that helps

young women process visual imagery with knowledge that will allow them to do so in less harmful ways,

the statement may need to be refined in a manner that actually forces young girls to process messages

that stay with them through viewing experiences. Certainly manipulating the placement of the visual

Studies in Media & Information Literacy Education, Volume 6, Issue 1 (February 2006), 1–14

# University of Toronto Press. DOI: 10.3138/sim.6.1.002

11

literacy statement is one way to proceed with this line of research. It might also be useful to incorporate

the textual information with the visual information, as is done in advertisements, in the hopes that both

will be processed. It is also possible repeated exposure to a similar statement is needed in order to really

compete with the more harmful messages in the media young women attend to. A pre-test/post-test

design or repeated measures design might allow for a more appropriate test of the visual literacy statement.

The findings here suggest that exposure to entertainment media, especially thin-ideal media, still predicts

higher scores on four disordered eating subscales. The findings also affirm what we have learned in

earlier studies-college women, to a large degree, are still dissatisfied with their body shape and size

and in some cases act on the dissatisfaction by engaging in dangerous eating behaviors. Future studies,

similar in nature, are underway to test these presumptions. Visual literacy and/or media literacy still

seem to be an important factor in helping young girls better understand the social effects of mass

media. The key is in developing a visual literacy campaign that reaches appropriate audiences and that educates

them about the digital manipulation often taken place in entertainment media. Even though women

and young girls may “know” images they see in the media are manipulated, it seems as if health and

medical advocates could find some way to demonstrate how fictional the images are, young women

and girls might not be as inclined to compare themselves to unrealistic images. The window on the

world that girls and young women are looking through indeed appears to be distorted. Other studies

suggest that encouraging young girls and women to view more sports and to participate more in sports

may help their overall attitudes toward their bodies. That said, these same girls are pummeled with “objectifying

images” in a media that has “culturally conditioned them (girls) to hate their bodies” (Pipher, 1994,

pg. 184). The goal for future research is to continue to find ways to combat the negative messages often

found in entertainment media and to find ways to keep girls from destroying their physical and mental

health at such early ages.

COLUMBIA ONLINE CITATION: HUMANITIES STYLE

Bissell, Kimberly. “Skinny Like You: Visual Literacy, Digital Manipulation and Young Women’s Drive to

be Thin.” Studies in Media & Information Literacy Education 6.1 (2006).

http://www.utpress.utoronto.ca/journal/ejournals/simile (insert access date here)

COLUMBIA ONLINE CITATION: SCIENTIFIC STYLE

Bissell, K. (2006). Skinny Like You: Visual Literacy, Digital Manipulation and Young Women’s Drive to

be Thin. Studies in Media & Information Literacy Education 6(1).

http://www.utpress.utoronto.ca/journal/ejournals/simile (insert access date here).

BIOGRAPHICAL NOTE

Kim Bissell is an associate professor in the Department of Journalism at the University of Alabama. Her

research interests include the social effects of mass media as it relates to disordered eating in women

and girls, with emphasis on the relationship between entertainment and sports media, sports

participation and body image distortion.

AUTHOR CONTACT INFORMATION

Kimberly L Bissell

Assistant Professor

Department of Journalism

College of Communication and Information Sciences

University of Alabama

Box 870172

Tuscaloosa, AL 35487-0172

Telephone: [phone]

Fax: [phone]

E-mail: [e-mail]

- Kimberly L. Bissell

University of Alabama