The relation of body dissatisfaction to salience of particular body sizes

Abstract

Objective

Research examining body image schemas has found that individuals high in body dissatisfaction direct increased attention toward body shapes relative to neutral stimuli. However, it is not known whether attention is attracted to particular body shapes over others (e.g., thin or obese). The present study examined whether body dissatisfaction would moderate the extent to which women find thin, average, and/or obese body sizes salient.

Method

Women with high and low body dissatisfaction (N=32) completed an indirect cognitive task assessing the relative salience of different body sizes (thin, average, obese).

Results

Degree of body dissatisfaction was inversely related to frequency estimates of obese body sizes; highly dissatisfied participants found obese body sizes less salient.

Conclusion

These results highlight the importance of examining the salience of specific body sizes. The present study has implications for understanding cognitive aspects of social comparison and body dissatisfaction.

Introduction

Eating disorders remain significant and well-recognized problems in the United States and other western cultures (1-3). Body dissatisfaction, a state of troubled, negative feelings about the size or shape of one's body (4), has been shown to be associated with likelihood of eating disorders (5-7). Although such dissatisfaction is not a problem solely faced by women, nearly half of women in one national survey reported having negative feelings about their appearance and a preoccupation with weight issues (8).

Given the prevalence of body dissatisfaction and its relationship with eating disorders, it is important to understand factors that contribute to its occurrence. There are a variety of cognitive biases centered around attention to weight and eating information (e.g., 9, 10, 11), attributions for life events (e.g., 12) which may be related to likelihood of body dissatisfaction and disordered eating [for a review see (13)]. Recent work implicates body image schemas (generalized knowledge structures similar to self-schemas that contain information about one's shape and weight) as a contributing factor in body dissatisfaction because of the way that they direct some individuals' attention toward appearance-related information and stimuli (14-19).

Empirical research has supported this hypothesis. In many studies, attention toward shape-related (e.g., plump, flabby) and control words (e.g., fables, crown) was examined in populations known to be related to body dissatisfaction (e.g., restrained eaters or individuals with eating disorders). These populations allocate greater attention to shape-related words than to control words (20-22). Of particular importance to the present study, one study examined the attention of individuals with vs. without eating disorders toward body silhouettes and neutral images (e.g., sporting equipment). Participants with eating disorders allocated greater attention to body silhouettes relative to those without eating disorders (23).

This study suggests that individuals with eating disorders find body shapes especially salient and attention-grabbing; however, this specific study does not tell us whether that attention is allocated toward particular (i.e., thin, average, or obese) body sizes. It is possible that the particular body sizes that women attend to may be influenced by the way they feel about their own bodies and, thus, by the content of their body-image schemas. Women who are satisfied with their bodies may attend to different body sizes than will women with high levels of body dissatisfaction. When a variety of body sizes are available for an individual to attend to, which ones are most salient and receive the most attention?

This question is an important one because the body images that are most salient and attract the most attention are most likely to be the ones used as comparison standards when making judgments. Engaging in social comparisons to others' bodies is related to body dissatisfaction (24, 25). For example, adolescent girls use both peers and media models as comparison targets to evaluate their own bodies (26). Similarly, adults exposed to appearance-related television commercials are significantly more dissatisfied with their appearance, relative to those who view non-appearance related commercials (16) [see (27) for a meta-analytic review]. Indeed, exposure to even a single thin target may be sufficient to lead to decreased body satisfaction and reduced confidence in one's attractiveness (26).

The importance of schemas and attentional differences toward potential social comparison targets in explaining body dissatisfaction suggests that it may be important to examine the specific stimuli that individuals who vary in their degree of body dissatisfaction attend to. One recent study measured eye gaze (i.e., a measure of selective attention) toward beautiful and ugly body parts in individuals who varied in their degree of eating disorder symptomology (28). In this study, selective attention toward others' beautiful body parts would be an example of a focus on upward (i.e., more favorable as compared to the self) social comparison targets. In contrast, focus on others' ugly body parts would illustrate attention toward downward (i.e., less favorable) social comparison targets. The results of this study showed that individuals with low eating disorder symptomology attended to beautiful body parts when they were viewing their own bodies and ugly body parts when they viewed others' bodies (i.e., they focused on downward comparison targets). In contrast, individuals with high eating disorder symptomology attended to their own ugly body parts and others' beautiful body parts (28). Arguably, these findings illustrate that a bias toward downward social comparisons can be self-serving and protective against body dissatisfaction, similar to research demonstrating that self-serving biases have general mental health benefits [see (29) for a meta-analysis], and thus provides further support for our hypothesis that attention to specific types of body images might differ as a function of individuals' body dissatisfaction.

Study overview

These findings lead to the question: when individuals who vary in their degree of body dissatisfaction are presented with social comparison targets (i.e., body shapes) of various sizes, which size(s) do they focus on the most? The first purpose of the present study was to provide an initial test of this question. Specifically, the present study examined whether women with varying degrees of body dissatisfaction differed in the extent to which they found thin, average, and obese body sizes salient. Second, we introduce a new methodology to complement the indirect, cognitive methods used in previous research (19). Modeled after Tversky and Kahneman's (30) famous names study, the present methodology indirectly requires individuals to estimate the frequency of thin, average, and obese body sizes. This methodology allows us to examine individuals' attention toward particular body sizes, rather than body shapes or eating-related stimuli in general.

To examine differential focus on thin, average, and obese body images across levels of body dissatisfaction, women who varied in their degree of body dissatisfaction completed a task designed to provide an implicit measure of salience of particular body images. This task drew on the logic of Tversky and Kahneman's (30) classic study of salience and the availability heuristic. In their study, participants listened to a list of male and female names and later were asked to estimate how many male names and how many female names they heard. In each trial, either the male or female names were predominantly composed of names of famous people. Participants erroneously judged the names of the sex containing the famous names as more frequent than they actually were. Tversky and Kahneman argued that the overestimation occurred due to the greater salience of the famous names, which led them to be more available in memory. Because the sex and the famousness of the names covaried, by asking participants to estimate gender, one could examine the salience of famous names without overtly asking about the proportion of famous names in the list.

Following this logic, participants in our study viewed serial presentations of thin, average, and obese male and female body silhouette drawings presented in different colors and then estimated the frequency of images of each color. We systematically covaried body silhouette size and the color in which the image appeared (e.g., obese silhouettes presented in the color red). Thus, the color of the body shapes served the same purpose as the gender of the names in Tversky and Kahneman's (30) study, and the size of the body shapes was analogous to the famous vs. non-famous dichotomy. If particular body types are especially salient, participants should have higher estimates of the frequency of images presented in the color associated with that body type. We hypothesize that women who are dissatisfied with their bodies would focus more on thin and average body sizes and less on obese body sizes. In contrast, we hypothesize that women who are low in body dissatisfaction would show the opposite pattern and focus more on obese body sizes and less on thin and average body sizes because focusing on a downward comparison target would be ego-protective (31) and self-esteem enhancing (32).

Method

Design and participants

This study employed a 3 (Silhouette Body Size: Thin, Average, Obese) × 2 (Silhouette Sex: Male, Female) within-participants design. Thirty-two undergraduate women (mean age =19.5; 90% European American) from a mid-sized Midwestern university participated in exchange for extra course credit. This study was approved by the institutional review board of that academic institution. Because of our interest in examining the salience of body sizes across various levels of body dissatisfaction, we selected women who scored low and high on the Body Dissatisfaction subscale of the Eating Disorder Inventory-2 [(33) see “Measures”]. Students enrolled in Psychology courses took part in a mass testing session at the beginning of the semester; the EDI-2 was included in this mass testing. Based on scores on the EDI-2 pretest, 16 participants exhibiting high body dissatisfaction (defined as a score above 15 on the Body Dissatisfaction subscale M=24.7, sd=3.24) and 16 exhibiting low body dissatisfaction (defined as a score below 9 on the subscale; M=1.63, sd=1.54) (33) participated in the study¹. Based on EDI-2 norms, a score of 9 is approximately the minimum of the “Normal College Female” range (16^th percentile), and a score of 15 is approximately the maximum (84^th percentile).

Procedure

Participants completed a computerized frequency estimation task to assess the salience of body sizes (see Measures). They also completed a second administration of the Eating Disorder Inventory-2, and they indicated their actual and ideal body size (see “Measures”). After these two body dissatisfaction measures, their weight and height was measured. Order of presentation of these tasks was counterbalanced such that half of the participants completed the computer task prior to the EDI-2 and weight/height measures, and half completed the computer task following those measures. The participants were fully debriefed after completing all study tasks.

Measures

Frequency estimation task

Participants completed 18 trials of the frequency estimation task; half of the trials contained silhouettes of men, and half contained silhouettes of women. Forty male or female body silhouettes were presented in each trial. Silhouettes were selected from the Stunkard, Sorensen, & Schulsinger (34) scale. For the frequency estimation task, 2 thin (silhouettes 1 and 2 in Stunkard's numbering system), 2 average (silhouettes 4 and 5), and 2 obese (silhouettes 7 and 8) were selected.

Each 40-silhouette trial included 2 target body sizes (e.g., thin and obese). The target body sizes were counterbalanced across trials. In each trial, the 2 target body sizes were presented 8 times each, in separate target colors. For example, one trial consisted of the presentation of thin and obese target silhouettes. This trial consisted of 8 thin silhouettes presented in one target color (e.g., red) and 8 obese silhouettes presented in another target color (e.g., green). To disguise the color-size covariation, in addition to the 16 target silhouettes, the trial also included 2 distraction silhouettes presented in each of the target colors (e.g., 2 red and 2 green silhouettes were randomly selected from the other body sizes). Finally, each trial included 2 “distraction colors” (i.e., different from the 2 target colors). In the first distraction color, the thin and obese silhouettes each were shown twice, and the average silhouettes were shown 3 times, for a total of 14 silhouettes. In the second distraction color, each of the silhouettes was shown once, for a total of 6 silhouettes. Thus, the distraction colors contained all of the body sizes, with a mean body size being the average body size. Across the participants, within each body dissatisfaction condition, each target color was presented equally often with each body shape. All silhouettes in a given trial were presented in a random order both within and across participants.

The frequency estimation task was computerized. To begin each trial, participants pressed the spacebar. A fixation point (+) was presented for 500 ms in the center of the screen, followed by a 500 ms blank screen. Each body silhouette was then presented one at a time on the center of the screen for 2000 ms. There was no inter-stimulus interval between the presentation of each silhouette. At the end of each 40-silhouette trial, participants were asked to estimate how many images appeared in each of the two target colors (i.e., red and green in the above sample trial). Thus, each trial provided an assessment of the relative attentional focus on body image silhouettes of the 2 target sizes for that trial. Participants were never told to focus on particular colors or body shapes or that the colors and shapes covaried.

Body dissatisfaction measures

Participants completed two body dissatisfaction measures: the body dissatisfaction subscale of the Eating Disorder Inventory-2 (33) (α=0.98) and an actual-ideal body size rating using the Stunkard et al. (34) silhouettes. The body dissatisfaction subscale was used to select participants with a range of body dissatisfaction. It consists of 9 items assessing how satisfied or dissatisfied an individual is with both overall body shape and the size/shape of specific body parts (e.g., “I think that my stomach is too big.”). Participants indicated their answer for each item on a scale consisting of 6 options: “always,” “usually,” “often,” “sometimes, “rarely,” “never”. Garner's (33) scoring algorithm was used to compute an overall body dissatisfaction score with a possible range of 0 (lowest body dissatisfaction) to 27 (highest body dissatisfaction).

Participants also indicated their perceived actual and ideal body sizes using the well-validated body figures rating scale developed by Stunkard et al. (34). The scale consists of 9 female silhouettes systematically ranging in size from very thin [body mass index (BMI) of approximately 18] (35) to very obese (BMI of approximately 45), and has been successfully used in previous research as a measure of individuals' body dissatisfaction and attitudes toward body shapes (36, 37). To assess participants' body dissatisfaction, participants' ideal body size was subtracted from their actual body size such that larger scores indicate greater degrees of body dissatisfaction (M=1.62, sd=1.29; range=−1.00-4.00).

Body size

Measurements of the participants' height and weight were used to compute participants' BMI (M=24.4, sd=5.49). Participants' height was measured using a yardstick fixed to the wall, and their weight measurements were taken using a bathroom scale.

Analytic strategy

Although two measures of body dissatisfaction were taken (i.e., the EDI and the actual-ideal measurement), the actual-ideal discrepancy is more closely related to the medium used in the frequency estimation task (i.e., both the body dissatisfaction measure and the frequency estimation task involved the use of body silhouettes). Additionally, although participants who scored high or low on the EDI were selected for participation, their scores on the actual-ideal discrepancy measure were more normally distributed (actual-ideal skewness =−0.01 and kurtosis =−0.77, compared to skewness =0.07 and kurtosis =−2.00 for the EDI) and thus can be included as a continuous predictor variable in general linear model analyses. This provides greater statistical power for our hypothesis tests than the categorical split on EDI scores. Thus, the actual-ideal discrepancy measure was used as an independent variable in all analyses².

Results

A mixed-factor General Linear Model (GLM) was used to examine body dissatisfaction as a predictor of participants' frequency estimations. The sex of the silhouettes (male, female) and the size of the silhouettes (thin, average, obese) were included as within-participants factors. The centered actual-ideal body dissatisfaction measure was included as a continuous between-participants predictor of participants' frequency judgments. Based on the logic of the availability heuristic, if participants found particular images relatively more salient, then they should overestimate those images' corresponding target color.

Consistent with the idea that silhouette salience would differ by body image dissatisfaction, there was a significant silhouette size X body dissatisfaction interaction, F(2, 29)=4.86, p<0.05, ε²=0.25. There were no other significant main effects or interactions.

To explore the nature of this interaction effect, we separately examined the relationship between body dissatisfaction and estimates for each of the 3 body sizes. Follow-up bivariate correlational analyses showed that body dissatisfaction did not influence participants' frequency ratings for thin or average silhouettes; rs(30)<±0.20, ns. However, as hypothesized, as participants' degree of body dissatisfaction increased they found obese male and female silhouettes less frequent, as indicated by a significant negative correlation, r(32)=−0.35, p<0.05. According to Cohen (38), this is a medium-sized effect.

Discussion

Using our frequency estimation task, we examined how the relative salience of thin, average, and obese social comparison targets differed for women varying in their degree of body dissatisfaction. Supporting our initial hypotheses, body dissatisfaction had an inverse relationship with women's frequency estimations of obese body sizes. Women with higher degrees of body dissatisfaction found obese body sizes less salient than women with lower body dissatisfaction.

These findings extend and qualify research examining body image schemas, which shows that restricted eaters and individuals with eating disorders direct their attention toward shape-related words (20-22), and body silhouettes (23) relative to neutral stimuli. The findings from the present study advance this line of research by showing that women with high levels of body dissatisfaction directed their attention away from obese body sizes. One potential explanation for these findings is that both women who are high and low in body dissatisfaction have body image schemas but they contain different information. Thus, these schemas may direct attention toward or away from different body sizes. Research examining appearance schemas suggests a second possibility: it is possible that only the highly dissatisfied women have salient body image schemas (15). These schemas may have directed highly dissatisfied women's attention away from obese body sizes; whereas women who were low in body dissatisfaction who may not have salient body image schemas may not have been subject to this influence.

The present findings also are consistent with past research showing that women who were high in eating disorder symptomology focused on other individuals' beautiful body parts, whereas low symptomatic women focused on others' ugly body parts (28). Similarly, in the present study, women who were highly dissatisfied found the obese body sizes (i.e., downward comparison targets) less salient than women who were satisfied with their bodies. Together with previous research, the present study suggests that women's body image schemas may direct attention toward both general (e.g., body shapes) and more specific (e.g., particular body shapes and body parts) stimuli.

This is an important advancement because women's body image schemas function like other mental representations and direct attention toward schema-consistent stimuli (14), including potential social comparison targets. Although the present study did not include a direct measurement of participants' social comparison to different body sizes, the salience of particular body sizes is important for understanding social comparison processes because individuals who focus more on social comparison targets are more affected by them (39, 40). Thus, the particular body sizes that women focus on, and potentially compare themselves to, may have implications for the perpetuation of the feelings they have about their bodies. For example, if women who are satisfied with their bodies find obese bodies relatively more salient and, thus, engage in downward comparison with obese individuals, their body satisfaction may be perpetuated. Downward comparisons with obese bodies can be ego-protective (41), and such comparisons may be important for the maintenance of body satisfaction. In contrast, body dissatisfaction may be perpetuated in women with high body dissatisfaction because downward comparison targets (i.e., obese bodies) are less salient.

Our findings are similar to research on attentional differences in anxious and depressed individuals which shows that these individuals tend to focus more on depressive or threatening stimuli (42, 43), as well as consistent with research on maladaptive attributions for life events in those with eating disorders (12, 44). Similarly, individuals with large degrees of body dissatisfaction may engage in fewer social comparisons with obese individuals, which could lead to a perpetuation of their negative feelings about their bodies. When presented with both upward and downward comparison targets, individuals who were high in body dissatisfaction appeared to selectively focus less on downward comparison targets. Thus, the cognitive processes exhibited by these individuals may become cyclical: high body dissatisfaction may lead to fewer downward comparisons, which may leave these women vulnerable to further body dissatisfaction. An empirical test of this hypothesis is necessary; however, the present findings, in conjunction with previous research documenting a link between social comparison and body dissatisfaction (26, 45) certainly suggest that this cycle is possible.

Limitations

There are, of course, limitations to the present study. First, the small sample size may leave the present study underpowered; however, the within-participants factors (i.e., body silhouette sex and size) help diminish this problem. In addition, the fact that significant differences in attention were observed is important to note. Further, even given the small sample size, the present study yielded important findings, and, at the very least, it demonstrated the utility of a new cognitive methodology. This study should be regarded as an initial examination and should be followed with larger-scale studies of the effect. Second, the design of the present study does not allow for causal inferences; the hypothesis about the potential cyclical nature of body dissatisfaction and social comparison warrants future causal examinations. Research examining exposure to thin images in the media suggests that people, particularly women, may compare themselves to these media images, which leads to increased body dissatisfaction (16, 27, 46-48). However, this research has not yet examined the particular body sizes that women would attend to, if a variety of body sizes were made available through the media.

Conclusions

Body dissatisfaction is a well-established problem in western cultures. The present study demonstrated that body dissatisfaction relates to the relative salience of different body sizes. Women who were dissatisfied with their bodies found obese body sizes less salient than body women who were satisfied with their bodies. Given that focusing on thin bodies has been shown to contribute to body dissatisfaction, the findings suggest that ignoring the prevalence of more obese bodies (i.e., possible downward social comparison targets) could be a potentially important cognitive contributor to the occurrence and maintenance of women's body dissatisfaction.