Men and the muscular ideal:
an experimental study on the effect of non-muscular and
muscular models on body image.
Geke van der Voort (12317810) University of Amsterdam Master’s thesis
Graduate School of Communication
Master’s programme Communication Science Supervised by Christin Scholz
Abstract
Women have been the focus of research surrounding body-ideal images in the media and their effects on body image. However, the circumstances of these effects on men’s body
image remain to be investigated. What is the effect of exposure to male models with a muscular body versus male models with a non-muscular body on the body image of men?
And what are the boundary conditions of this effect? A diverse sample of 151 men were randomly assigned to be exposed to either a muscular fashion model or a non-muscular
fashion model. Contrary to previous research, no difference in body image was found between the two groups. Prior work suggested the hypothesis that effects of exposure to muscular male models on body image would be mediated by increased social comparison. In
this study marginally significant differences were found between the groups. Similarly, the effect of social comparison on body image was different with marginal significance between men with a low and those with a high discrepancy between their ideal and actual body types. This study provides an indication for the mediating effect of attitude towards clothing on the relationship between the exposure groups and social comparison. The results of this study suggest men and women have different boundary conditions for the effect of exposure to a body ideal on body image. The research indicates that men should be treated separately from
women in future research and intervention efforts and that future research should explore different aspects than just muscularity for determining the factors of advertisements that
Introduction
Advertisements and their effects on the psychological wellbeing of individuals has been a hot topic in research over the last decades. Advertisements tend to portray beautiful, flawless bodies that create an ideal that can be hard to live up to. Plenty of research shows that this can have negative effects on our well-being and in particular the way we view our body (e.g. Duggan & McCreary, 2004; Grogan, Williams & Connor, 1996; Groesz,
Levine & Murnen, 2002; Grabe et al., 2008). Although most of the popular and scientific attention has been focused on women, it has been shown that men are also affected by body-ideals presented in the media (e.g. Duggan & McCreary, 2004; Grogan, Williams & Connor, 1996; Agliata & Tantleff-Dunn, 2004, Barlett, Vowels & Saucier, 2008). It is largely
unknown how and under what circumstances these effects occur in men, but there is reason to believe that key relationships may differ from those found in women. For instance, whereas women tend to compare themselves to a thin-ideal, men are more inclined to compare themselves to a muscular ideal (Fisher, Dunn & Thompson, 2002). In other words, the advertisement environment is saturated with strong imagery of perfect male bodies that may have harmful effects on half the population, which are currently rather poorly understood.
This paper re-evaluates the effect of exposure to ideal model bodies on body image found in women in a male population and tests new hypotheses about the boundary
conditions and mechanisms of these effects that are specific to the male context.
Bessenoff (2006) found that the negative effect of exposure to thin-ideal images on body image among women, can be explained by the level of social comparison the women engage in after this exposure. She also found that the level of self-discrepancy women experience moderates this effect. To this date, no research has been done to investigate if these factors might also explain and influence the relationship between exposure to the
same moderator and mediator that applies to women, can also be applied to men or if this relationship is perhaps so different that we should look for other moderators and mediators to explain its effects. By gaining knowledge on which individuals are more likely to experience negative effects from exposure to the muscular-ideal, future policy efforts or interventions can be more targeted. Furthermore, if men do differ in how and when they are affected by body-ideal images, they should be treated separately in research efforts.
Lastly, this study can also provide awareness about the influence of muscular-ideal advertisements on men’s body image and can contribute towards a public discourse. For women, the increase in awareness about the effect of thin-ideal images on body image has sparked public discourse and public pressure has managed to make changes in the industry. As a result, female models have become more diverse in their shape and size in the
advertisement industry (Cheng, 2017). Increased awareness can contribute towards similar changes in the male advertisement industry.
RQ: What is the effect of exposure to male models with a muscular body versus male models
with a non-muscular body on the body image of men? And what are the boundary conditions of this effect?
Theoretical framework Body image
The concept of body image has been widely used in research to describe how people perceive and feel about their own body. One aspect of body image is body dissatisfaction, this occurs when an individual has negative thoughts and low esteem about their body (Schilder, 1950). Multiple studies have shown that this dissatisfaction can lead to a range of mental and physical health problems, such as depression, low self-esteem and eating
disorders (Dittmar, 2009; Cafri & Thompson, 2004) and related behaviors such as such as unhealthy diets, steroid abuse, excessive exercising or unnecessary cosmetic surgery (Cafri et al., 2015; Stice, Mazotti, Krebs, & Martin, 1998; Davis, 1995).
Research into the effects of the media on body image has been mainly focused on women. Meta-analyses show a consistent significant causal relationship between exposure to thin-ideal images for women and negative body image (Groesz, Levine & Murnen, 2002; Grabe et al., 2008). The body image of women was significantly more negative after
exposure to thin media images than after exposure to images of average size models, plus size models or inanimate objects. Although there is less research for men, there are some studies that have provided evidence that men can be similarly affected by body ideals presented in the media (e.g. Duggan & McCreary, 2004; Grogan, Williams & Connor, 1996; Agliata & Tantleff-Dunn, 2004; Barlett, Vowels & Saucier, 2008; Ogden & Mundray, 1996).
Although both genders are thus affected by body-ideals presented in the media, they differ in how they compare themselves to these ideals. Whereas women have a stronger focus on the weight dimension of their body compared to the thin-ideal, men have a stronger focus on the muscle dimension and compare themselves to a muscular ideal (Fisher, Dunn & Thompson, 2002). This can be explained by the different cultural norm that is set for bodies of the different genders. Leit, Pope and Gray (2002) analyzed issues of the Playgirl centerfold models from 1973 to 1997. They found that the male models became increasingly more lean and muscular over time, suggesting that cultural norms of the male body ideal are growing increasingly more muscular. A similar study by Wiseman, Gray, Mosimann & Ahrens, (1992) studied Playboy centerfold and Miss America contestants between 1979 and 1988 and reported that the expected weight of these women significantly decreased during these years. These studies show that the cultural norm for physical appearance has become increasingly
While there have been quite a few studies that found evidence that men experience less body satisfaction after exposure to muscular ideal images compared to after exposure to neutral images without a model (e.g. Grogan, Williams & Connor, 1996; Leit, Pope & Gray, 2002), to my knowledge only two studies explored the difference between exposure to muscular ideal images and non-muscular body images (Lorenzen, Grieve & Thomas, 2004; Diedrich’s & Lee, 2010). Whereas Lorenzen, Grieve and Thomas found evidence for an increased body dissatisfaction after exposure to muscular ideal images compared to non-muscular body images, Diedrich’s and Lee found no significant differences between the effects of these two conditions. It is therefore hard to draw any conclusions based on these two studies alone.
This is problematic because although comparing exposure to a muscular ideal to a neutral image provides evidence for the negative effects of exposure to these images, it does not allow inferences about the specific elements of the image or aspect of the model’s body that cause the effect. Furthermore, it provides evidence for a problem but it does not provide an alternative. By comparing the effects of muscular models to non-muscular models, the actual effect of the muscularity aspect can be determined. If exposure to non-muscular models does not have a negative or even has a positive effect on body image, more non-muscular representation in the advertisement industry could be a potential solution to this problem.
H1: Exposure to muscular male fashion models will lead to a more negative body
image among men than exposure to a non-muscular fashion model.
Social comparison. The concept of Social comparison used in this study is based on the
compare themselves to social targets in situations in which there is no objective benchmark to compare oneself to, for instance in order to evaluate one’s performance on a task. They engage in social comparison, meaning that they compare themselves to similar others. Two types of social comparison were later defined: upward comparison and downwards
comparison (Latané, 1966). When individuals compare themselves to others who they believe are inferior to them in a certain domain, they engage in downward comparison. This usually leads to enhanced feelings of self-worth. Upward comparison however, means that
individuals compare themselves to others who they deem better than themselves in a certain domain. In this case, the comparison can lead to diminished feelings of self-worth (Wood, 1989).
In the context of exposure to advertising containing imagery of ideal bodies, most instances of social comparison in the target group can be described as upward social comparison. This upward comparison has also been shown to play a part in explaining the negative effect of exposure to media on self-worth and specifically body image. Richins (1991) found that exposure to idealized images in advertisement raised comparison standards for attractiveness and lowered satisfaction with the own attractiveness of women.
Furthermore, Tiggemann and McGill (2004) found that the amount of social comparison women engage in plays an important role in the effect of media advertisements on body dissatisfaction and is increased when exposed to an image of an ideal body. Exposure to an ideal body makes one’s own body concerns more salient, resulting in higher levels of social comparison (Bessenoff, 2006).
H2: The effect of exposure to muscular male models (compared to a model with a
exposure to non-muscular male fashion models, men who highly engage in social comparison will experience more negative effects on body image than men who engage in little social comparison.
Self-discrepancy. Upward comparison to muscular models, is associated with a discrepancy
between the ideal body and the perceived self among men (Bessenoff, 2006). The concept of Self-discrepancy used in this study is based on the self-discrepancy theory by Higgins (1987). According to this theory, individuals compare themselves to internalized standards of their ideal selves. A discrepancy, or gap, between the actual self and the ideal self, can lead to negative outcomes, such as disappointment and dissatisfaction. In later studies appearance-related self-discrepancy was also found to be appearance-related to body image and body dissatisfaction (Jung, Lennon & Rudd, 2001; Vartanian, 2012). Bessenoff (2006) found that social
comparison is a mediator for the effect of exposure to thin-ideal advertising on symptoms of depression, weight-related thoughts, and weight regulatory thoughts and that physical discrepancy moderated this mediating effect. According to this study, women with high self-discrepancy were almost twice as likely to engage in social comparison after being exposed to thin-ideal images, than women with low self-discrepancy.
If an individual has a high level of self-discrepancy, an ideal body presented in the media is likely more relevant to them because they are more preoccupied with the way their body looks. Being exposed to the ideal body they feel they do not live up to, makes the issue more salient. Therefore they are more likely to use the body presented in the media as a standard for comparison (Bessenoff, 2006).
Some studies investigating male self-discrepancy found that men did not differ significantly in the amount of self-discrepancy they experienced compared to women, others found that men experienced less self-discrepancy than women. However, these mixed
findings can be explained by looking at how self-discrepancy was measured. Unlike women, who generally tend to have an ideal body image that is thin, men are just as likely to want to be leaner as they are to want to be bigger and more muscular. When body discrepancy is measured compared to one preset ideal of size, the results may cancel each other out.
Therefore, it is important to measure the absolute discrepancy between men’s ideal body and their actual body, without taking into account if they would like their body to be bigger or smaller. Some studies also measured self-discrepancy by only using figures that increased in overall body size and fat, whereas men are more focused on their level of muscularity instead of their weight (Fisher, Dunn & Thompson, 2002). Studies that measured absolute
discrepancy and looked at muscularity concluded that there is no significant difference
between the level of self-discrepancy men and women experience (Vartanian, 2012; Halliwell & Dittmar, 2006).
Besides the moderating effect of self-discrepancy on the relationship between exposure to thin-ideal images and social comparison, self-discrepancy was also found to moderate the effect of social comparison on symptoms of depression, weight-related thoughts, and weight-regulatory thoughts. Self-discrepancy thus moderated the effect of exposure condition on the mediator as well as the effect of the mediator on the dependent variables. Whereas women with high self-discrepancy felt worse after comparing themselves to the thin-ideal images, women with low self-discrepancy actually felt better after comparing themselves to the thin-ideal images (Bessenoff, 2006).
The meta-analysis by Vartanian (2012) concluded that, similar to women, higher levels of self-discrepancy are also associated with increased body dissatisfaction among men. Therefore, it is expected that similar results to the study by Bessenoff (2006) on women will be found for men with regards to self-discrepancy as a moderator (see figure 1).
H3: The effect of exposure to muscular models compared to non-muscular models on
social comparison will be stronger for men with a high level of self-discrepancy compared to men with a low level of self-discrepancy.
H4: The effect of social comparison on body image will be stronger for men with a
high level of self-discrepancy compared to men with a low level of self-discrepancy.
This study will explore difference between the effect of exposure to non-muscular and muscular models on body image and measure if and how social comparison explains this effect. Furthermore, the influence of self-discrepancy on this effect will be determined. The results of this study will provide evidence on how men are influenced by the muscular ideal and whether this has a more negative impact on their body image than exposure to a non-muscular body. Furthermore, this study will determine whether men should be targeted similarly or differently than women in future research and intervention efforts.
Figure 1. Conceptual model.
Method
Exposure to
model
Body image
Social
comparison
Self-discrepancy
Participants
151 people provided informed consent to participate in the study. The sample was recruited through personal connections, survey exchange and social networking sites, using
convenience and snowball sampling techniques. Only men who are over the age of 18 were able to take part in this study.
7 participants had partial data, their missing data were treated as missing values. One participant was removed from the study because he had consistently chosen the same answer, indicating a lack of attention to the survey, and 11 participants did not pass the screening criteria and were sent to the end of the questionnaire after the screening questions. This data was also removed from the study. Analyses were thus conducted using a final sample of 139 participants. Participants’ ages ranged from 18 to 78 years (M = 27.83, SD = 9.67).
Study design. This study has a 2 x 2 full-factorial design, with exposure condition (2
levels, namely: muscular model and non-muscular model) as between-subjects experimental and self-discrepancy (2 levels, namely: low self-discrepancy and high self-discrepancy) as between subjects quasi-experimental factor (see table 1).
Table 1
Experimental design
Low self-discrepancy High self-discrepancy
Photo of muscular model N = 47 N = 23
Photo of non-muscular model
Cover story. At the start of the survey a cover story is presented. This is important
because knowing what the study is about might influence how participants answer the questions. They might answer what they think is desired or consciously reconsider their answer. Participants are told they are participating in a study researching the effects of using different male models on advertisement effectiveness.
Exposure to Model. The main independent variable is exposure to a fashion model.
This variable has two conditions, namely a model with a muscular body and a model with a non-muscular body. To operationalize this variable, participants were randomly assigned to be exposed to a series of five photos of either muscular models or non-muscular models. Across conditions, models ‘sell’ a similar item and the photographs are similar in background and in the pose of the model to minimize influence of third variables. The photos are obtained from an actual fashion site, namely Asos.com (http://www.asos.com), to increase the
ecological validity. Any tattoos or notable jewelry were removed with photo editing. To avoid case-by-category confounds and avoid systematic biases due to small differences in the clothes, poses or skin colors of the different models across conditions, the participants are exposed to a series of photos instead of just one in each condition. By
exposing them to various models and clothing items the focus is less on the specific model or clothing item, and more on the body type in general.
16 photos of muscular models and 16 photos of non-muscular models were pre-tested on a sample of 6 respondents. The participants were asked to rate the muscularity of the models on a scale of 1 (not muscular at all) to 5 (extremely muscular). The five photos that scored lowest on this scale (M=1.5, SD=0.28) and the five photos that scored highest on this scale (M=3.9, SD=0.58) were selected as the stimulus material for this experiment (see figure 2).
Figure 2. Example of stimulus material for the muscular condition and non-muscular condition, respectively.
Attitude towards the clothing of the model. To make the cover story credible and
measure differences in the attitude towards the clothing the models were ‘selling’ between the conditions, respondents were asked to rate 5 statements on a scale of 1 (not at all) to 5 (very much) after being exposed to the models. Examples of these statements are “How likely would you be to buy this clothing item if you saw it in a store?” and “How much do you like this clothing item?”
For each of the five attitude scale items, the mean was computed across model images within participant. A factor analysis was done to see if these 5 new variables loaded on the same factor. The Kaiser-Meyer-Olkin measure of sampling adequacy was 0.783 and
Bartlett’s test of sphericity was significant (χ2 (10) = 667.222, p < .001). Only the first factor had an eigenvalue above 1 and explained 75% of the variance. One factor was extracted and all items had a factor loading of above 0.8.
To test the internal consistency of this scale, a reliability analysis was conducted. With a Cronbach’s alpha of 0.92, the internal consistency is high. None of the items needed to be excluded or reversed. To compute this new scale, the mean score of all 5 items was calculated for each participant. The computed variable has a minimum score of 1.04 and a maximum score of 4.56 (M = 2.47, SD = 0.69).
Discrepancy. The moderator that is used in this experiment is Physical
Self-discrepancy. This is measured using the Visual Body Scale for Men (VBSM) figural rating scales (Talbot, Cass & Smith, 2019). This includes two scales, the VBSM-Muscularity (VBSM-M) and the VBSM-Body Fat (VBSM-BF). The VBSM-M scale consists of 10 male bodies increasing in muscularity. The VBSM-BF consists of 10 male bodies increasing in body fat percentage. The participants are asked the following questions: “Please indicate which figure most resembles your own body on the dimension of muscularity/fat percentage” and “Please indicate which figure most resembles your ideal body on the dimension of muscularity/fat percentage (i.e., the body you would most like to have as your own).” Since this is a quasi-experimental design, it is not possible to use this variable to assign the participants to different conditions beforehand.
The scores on these two questions were extracted from each other to find the amount of self-discrepancy. To make sure absolute discrepancy was measured, any negative scores were made positive so that the direction of the discrepancy was not relevant. The two variables were then transformed into variables with two groups (low discrepancy and high discrepancy). For the muscle discrepancy variable this was done by taking the median (2) and assigning all scores of two and lower to the low discrepancy group (M = 1.22, SD = 0.75) and all scores above two to the high discrepancy group (M = 3.73, SD = 0.75). For the weight discrepancy group this was also done by taking the median (2) and assigning all scores of
lower than two to the low discrepancy group (M = 0.61, SD = 0.49) and all scores of two and higher to the high discrepancy group (M =2.66, SD = 1.02).
Body image. The dependent variable of this study is Body Image. After being
exposed to one of the exposure conditions, the participants are asked fill in the Body Image State Scale (BISS) (Cash, Fleming, Alindogan, Steadman & Whitehead, 2002). This scale measures body image state with six statements that are answered on a 9-point Likert-type scale. Examples of statements are “Right now, how physically attractive do you feel?” (1 = Extremely unattractive to 9 = Extremely attractive), or “Right now, how do you feel about your looks compared to how you normally feel about your looks?” 1 = A great deal worse, 9 = A great deal better).
To create this variable, a scale was constructed of five items. To establish that these five items measure the same factor, a factor analysis with oblique rotation was conducted. The Kaiser-Meyer-Olkin measure of sampling adequacy was .80, above the commonly recommended value of 0.60, and Bartlett’s test of sphericity was significant (χ2 (10) = 422.451, p < .001). Only the first factor had an eigenvalue above 1 and explained 71% of the variance. One factor was extracted and all items had a factor loading of above 0.7.
To test the internal consistency of this scale, a reliability analysis was conducted. With a Cronbach’s alpha of 0.88, the internal consistency is high. None of the items needed to be excluded or reversed. To compute this new scale, the mean scores on all four items were calculated. The computed variable has a minimum score of 1.60 and a maximum score of 8.40 (M = 5.55, SD = 1.46).
State social comparison. The mediator in this study is Social Comparison, which is
measured using a variation of the Extent Thought questionnaire used by Bessenoff (2006). This questionnaire assesses thoughts relating to social comparison with regards to weight,
while looking at and rating the advertisements. Participants respond on a 5-point Likert scale ranging from 1 (not at all) to 5 (to an extreme degree) to the following questions: “To what extent were your thoughts related to comparing yourself to something or someone in the ads?”, “To what extent did you have thoughts related to comparing your own muscularity to the muscularity of the models?”, and “To what extent did you have thoughts related to comparing your own body shape to the body shapes of the model?” In the original version of the questionnaire developed by Bessenoff, the last two questions were related to weight. However, as this questionnaire was designed for women and men tend to have a stronger focus on muscles, these questions have been adapted. Additional questions, such as “To what extent did you think about the products and their features?” will also be asked to maintain the cover story.
To create this variable, a scale was constructed of three items. To establish that these three items measure the same factor, a factor analysis with an oblique rotation was conducted. Firstly, the Kaiser-Meyer-Olkin measure of sampling adequacy was 0.76, above the
commonly recommended value of 0.60, and Bartlett’s test of sphericity was significant (χ2 (3) = 245.234, p < .05). Only the first factor had an eigenvalue above 1 and explained 85% of the variance. One factor was extracted and all items had a factor loading of above 0.9.
To test the internal consistency of this scale, a reliability analysis was conducted. With a Cronbach’s alpha of 0.90, the internal consistency is high. None of the items needed to be excluded or reversed. To compute this new scale, the mean scores on all four items were calculated. The computed variable has a minimum score of 1.00 and a maximum score of 5.00 (M = 2.74, SD = 1.11).
Control variables. Two additional dependent variables that are measured are Positive
affect, Negative affect and Trait social comparison. To measure positive and negative affect, the Positive and Negative Affect Schedule (PANAS) scale will be used (Watson, Clark &
Tellegen, 1988). Participants are shown 10 words that describe positive affect feelings and emotions and 10 words that describe negative affect feelings and emotions. For each word the participants is asked to indicate to what extent they are currently experiencing this feeling or emotion on a 5-point likert scale ranging from 1 (very little or not at all) to 5 (extremely).
To create the variable Positive affect, a scale was constructed of ten items. To establish that these ten items measure the same factor, a factor analysis with an oblique rotation was conducted. Firstly, the Kaiser-Meyer-Olkin measure of sampling adequacy was 0.88, above the commonly recommended value of 0.60, and Bartlett’s test of sphericity was significant (χ2 (45) = 518.607, p < .001). There were two factors that had an eigenvalue above 1, the first factor explained 47% of the variance and the second factor 10%. Two factors were extracted. These factors could be interpreted as attention (interested, alert, attentive, active) and positive affect (excited, strong, enthusiastic, proud, inspired, determined). Two new separate factor analyses were run for both of these dimensions.
To test the internal consistency of this scale, a reliability analysis was conducted. With a Cronbach’s alpha of .073, the internal consistency is high. None of the items needed to be excluded or reversed. To compute this new scale, the mean score of all 4 items was calculated for each participant. The computed variable has a minimum score of 1.00 and a maximum score of 5.00 (M = 2.64, SD = 0.79).
To test the internal consistency of this scale, a reliability analysis was conducted. With a Cronbach’s alpha of 0.86, the internal consistency is high. None of the items needed to be excluded or reversed. To compute this new scale, the mean score of all 6 items was calculated for each participant. The computed variable has a minimum score of 1.00 and a maximum score of 4.33 (M = 246, SD = 0.84).
rotation was conducted. Firstly, the Kaiser-Meyer-Olkin measure of sampling adequacy was .91, above the commonly recommended value of 0.60, and Bartlett’s test of sphericity was significant (χ2 (45) = 784.943, p < .001). Only the first factor had an eigenvalue above 1 and explained 59% of the variance. One factor was extracted and all items had a factor loading of above 0.6.
To test the internal consistency of this scale, a reliability analysis was conducted. With a Cronbach’s alpha of 0.90, the internal consistency is high. None of the items needed to be excluded or reversed. To compute this new scale, the mean scores on all 10 items were calculated. The computed variable has a minimum score of 1.00 and a maximum score of 4.30 (M = 1.74, SD = 0.73).
To measure Trait social comparison, the Physical Appearance Comparison Scale-Revised (PACS-R) was used (Schaefer & Thompson, 2014). Participants were asked to indicate how often they engage in physical social comparison on a scale of 1 (never) to 5 (always) for 11 statements. Examples of statements include “When I'm out in public, I compare my physical appearance to the appearance of others” and “When I meet a new person (same sex), I compare my body size to his/her body size.”
To create the variable Trait social comparison, a scale was constructed of eleven items. To establish that these eleven items measure the same factor, a factor analysis with varimax rotation was conducted. Firstly, the Kaiser-Meyer-Olkin measure of sampling adequacy was 0.84, above the commonly recommended value of 0.60, and Bartlett’s test of sphericity was significant (χ2 (55) = 554,004, p < .001). Two factors had an eigenvalue above 1, the first factor explained 51% of the variance and the second 10%. Two factors were extracted. These factors could be interpreted as Trait social comparison on the muscle dimension and Trait social comparison on the weight dimension. Two new separate factor analyses were run for both of these dimensions.
To test the internal consistency of the Trait social comparison on the muscle
dimension scale, a reliability analysis was conducted. With a Cronbach’s alpha of 0.89, the internal consistency is high. None of the items needed to be excluded or reversed. To
compute this new scale, the mean score of all 6 items was calculated for each participant. The computed variable has a minimum score of 1.00 and a maximum score of 5.00 (M = 2.93, SD= 0.86).
To test the internal consistency of Trait social comparison on the weight dimension scale, a reliability analysis was conducted. With a Cronbach’s alpha of 0.87, the internal consistency is high. None of the items needed to be excluded or reversed. To compute this new scale, the mean score of all 5 items was calculated for each participant. The computed variable has a minimum score of 1.00 and a maximum score of 5.00 (M = 2.41, SD = 0.90).
Manipulation check. In order to make sure the bodies of the models in the two
conditions are actually perceived as intended, that is muscular models were perceived as more muscular than non-muscular models and social comparison with muscular models is likely to indicate upward social comparison rather than downward social comparison, a manipulation check was conducted. Respondents are asked to indicate how muscular the bodies in the photos are and how muscular they think they are themselves on a 5-point Likert scale of 1 (not muscular at all) to 5 (extremely muscular). Manipulation of the photos was successful if the mean score of perceived muscularity in the muscular condition is
significantly higher than the mean score of perceived muscularity in the non-muscular condition and if perceived muscularity in the muscular condition is significantly higher than the own perceived muscularity. This is tested with an independent and a dependent t-test, respectively. To check if the cover story worked and the true purpose of the study remained concealed, the respondents are also asked to briefly describe what they thought the research
Procedure. After receiving the online invitation to participate in this study,
participants are shown a screen that explains the contents of the study, the university’s
privacy policy and the contact information of the researcher. They are asked to give informed consent if they would like to continue in this study. They are then asked to provide
information about their gender and age to check if they are allowed to take part in this study. Next, they are asked to indicate the body that resembles their own body and their ideal body on the aspect of weight and muscularity. Afterwards, they are exposed to five photos of either non-muscular or muscular models and asked to answer 5 questions about their attitude
towards the clothing for each model. Next, they either answer the body image scale or the state social comparison scale. The order of this was randomized to control for influence of this order on the results. After this, they continue to the state social comparison scale and the positive and negative affect scale. Lastly, they are asked to indicate how muscular they thought each of the models were that they were exposed to, how muscular they think they are themselves and what they think the study is about. They are then thanked for their
participation and debriefed about the true purpose of this study.
Data analysis. Two analyses are used to test the hypotheses. To test hypotheses 1,2
and 3, a regression analysis using PROCESS model 7 is run, with exposure condition as an independent variable, body image as a dependent variable, state comparison as a mediating variable and body discrepancy as a moderator of the effect of exposure condition on state comparison. To test hypothesis 4, a separate regression analysis using PROCESS model 14 is run, with exposure condition as an independent variable, body image as a dependent variable, state comparison as a mediating variable and body discrepancy as a moderator of the effect of state comparison on body image.
Randomization check
In order to check if age was equally distributed across the groups for both the exposure conditions and the self-discrepancy groups, two independent T-tests were conducted. The test for exposure condition had condition (non-muscular = 0 vs. muscular = 1) as a grouping variable, and age as a testing variable. The T-test showed that participants’ mean age in the non-muscular condition (M = 27.75 years, SD = 10.73) was not significantly different from participants’ mean age in the muscular condition (M = 28.16 years, SD = 8.56),t (137) = -0.245, p = .807.
The test for self-discrepancy had self-discrepancy (high vs. low) as a grouping variable and age as a testing variable. The T-test showed that participants’ mean age in the low self-discrepancy group (M = 28.69 SD = 10.96) and the participants’ mean age in the high self-discrepancy group (M = 26.71, SD = 6.98) were not significantly different, t (136) = 1.16 p = .248. Randomization was thus successful for age across both variables.
Manipulation check. An independent T-test showed a significant difference in
muscularity between models in the non-muscular condition (M = 1.73, SD = 0.71) and the muscular condition (M = 3.53, SD = 0.62), t (132) = -15,575, p < .001. This means that there was enough difference between the conditions and the models in the muscular conditions were actually perceived to be more muscular than the models in the non-muscular condition.
A dependent T-test also showed that both conditions differed significantly from how the participants rated their own muscularity (M =2.47, SD = 0.88). Participants rated the models in the non-muscular condition as being significantly less muscular then they rated themselves, t (66) = -8.509, p < 0.001. They rated the models in the muscular condition as being significantly more muscular than they rated themselves, t (66) = 13.940, p < 0.001.
effect of exposure condition on social comparison, a regression analysis using PROCESS model 7 was run, with exposure condition as an independent variable, body image as a dependent variable, social comparison as a mediating variable and the moderating effect of muscular body discrepancy on the effect of exposure condition on state social comparison.
This analysis showed that the direct effect of exposure condition on body image was not significant, b = 0.02, t(131) = 0.78, p = .93, 95% CI [-0.50, 0.55]. Thus, hypothesis 1 was not supported by the data.
The effect of exposure condition on state social comparison was marginally significant, b = 0.36, t(130) = 0.19, p = .07, 95% CI [-0.50, 0.55]. The difference was
trending in the hypothesized direction, as state social comparison was higher in the muscular exposure condition (M = 2.60, SD = 1.08) than in the non-muscular exposure condition (M =2.93, SD = 1.13)
State social comparison was a marginally significant predictor of body image, b = -0.203, t(131) = -1.71, p = .090, 95% CI [-0.44, 0.03]. The effect was in the hypothesized direction. Participants who scored higher on state social comparison, scored lower on body image. Indirect effect = -0.07, SE = 0.07, 95% CI [-0.24, 0.01]. There was no mediation effect since the confidence interval includes zero. Thus, there was no support found for hypothesis 2.
There was no significant interaction effect of exposure condition and body discrepancy on state social comparison, b = 0.09, t(130) = 0.22, p = .823. 95% CI [-0.70, 0.88]. Body discrepancy did not significantly moderate the effect of exposure condition on state social comparison. Thus, there was no support found for hypothesis 3.
To test hypothesis 4, a separate regression analysis using PROCESS model 14 was run testing the moderating effect of body discrepancy on the effect of state social comparison on body image. There was no significant interaction effect of state social comparison and
self-discrepancy on body image, b = -0.32, t(129) = -1.43, p = 0.15, 95% CI [-0.76, 0.12]. Self-discrepancy did not significantly moderate the effect of state social comparison on body image, but the effect was in the hypothesized direction. Those participants who scored higher on state social comparison and belonged to the high discrepancy group, scored lower on body image than participants who scored lower on state social comparison and belonged to the low discrepancy group.
Exploratory analyses. Since there were no significant results in the testing of the
hypotheses, some exploratory analyses were done with the variables attitude toward the clothing items, age and affect.
Attitude towards clothing items. In order to test whether attitude towards clothing
items was perhaps a better moderator of the mediating effect of social comparison than self-discrepancy, a regression analysis using PROCESS model 7 was run, with exposure condition as an independent variable, body image as a dependent variable, state comparison as a
mediating variable and attitude towards clothing as a moderator of the mediating effect. No moderating effect for attitude towards clothing was found, but there was a significant effect found of attitude towards clothing on state social comparison, b = 0.545, t(131) = 4.04, p < 0.001, 95% CI [-0.28, 0.81]. Those participants who scored higher on attitude towards the clothing, also scored higher on state social comparison.
Next, a regression analysis using PROCESS model 6 was done, testing the indirect effect of exposure condition on body image with attitude towards clothing items and social comparison as a mediator. This showed that the effect of exposure condition on attitude towards clothing items was significant, b = 0.264, t(132) = 2.25, p = .026. The effect of attitude towards clothing items on social comparison was significant, b = 0.527, t(132) = 3.9, p < 0.001. Lastly, the effect of social comparison on body image was marginally significant,
b = -0.25, t(131) = -1.96, p = .052. Indirect effect = -0.03, SE = 0.03, 95% CI [-.10,0.00]. There was however no mediation effect since the confidence interval includes zero.
Another analysis was done to test whether attitude moderates the main effect of exposure condition on body image, but this yielded no significant results
Age. To see if age was a possible moderator for the relationship between exposure
condition and body image, a regression analysis using PROCESS model 1 was executed. This analysis provided no support for age as a moderator, but did find that age was a significant predictor of body image, b = -0.043, t(131) = -2.92, p = 0.004, 95% CI [-0.07, 0.01]. Participants with a higher age, scored lower on body image.
Affect. For affect there were three variables, namely positive affect, attention and
negative affect. To see if any/all of these variables possibly mediate the effect of exposure condition on body image, a regression analysis using PROCESS model 4 was done. There was no significant effect of exposure condition on any of these variables. However, this test did indicate that positive affect (b = 0.78, t(129) = 4.10, p < 0.001, 95% CI [0.40, 1.16]), negative affect (b = -0.70, t(129) = -3.88, p < 0.001, 95% CI [-1.05, -3.40]) and attention (b = -0.48, t(129) = -2.39, p < 0.018, 95% CI [-0.88, 0.08]) were significant predictors of body image. Participants who scored higher on positive affect, also scored higher on body image. Participants who scored higher on negative affect and attention scored lower on body image.
Conclusion and discussion Hypotheses
The present study explored the difference between the effect of exposure to non-muscular and muscular models on body image and if and how social comparison explains this effect. Furthermore, the influence of self-discrepancy on this effect was investigated. The results of this study do not provide full support for the formulated hypotheses.
First of all, there was no evidence found that exposure to muscular models leads to a more negative body image in men than exposure to non-muscular models. These results contradict the study by Lorenzen, Grieve and Thomas (2004) who did find that exposure to a muscular ideal resulted in more body dissatisfaction than exposure to a non-muscular model. However, their results were based on before and after measurements of body image, which might explain the discrepant findings. Perhaps there was a significant difference in body image before and after exposure to the (non-)muscular models that this study did not find because a pre-test was not included. It would be interesting to repeat the current experiment and include a measurement of body image before exposure to the (non)-muscular models to make a good comparison to the study by Lorenzen, Grieve and Thomas.
Other studies, for example by Grogan, Williams and Connor (1996), Leit, Pope and Gray (2002) and Diedrichs and Lee (2010) found that men reported a more negative body image after exposure to muscular models than after exposure to a neutral image without a model. Based on these earlier findings, the present results suggest that this decrease in body image may not be exclusively caused by the exposure to muscular models but perhaps by other aspects of the imagery such as facial beauty. This would explain why there were no differences found in exposure to muscular models versus exposure to non-muscular models. It would be interesting to examine this further by exploring the difference in body image after exposure to a neutral image, a non-muscular model and a muscular model or by manipulating other image aspects instead of muscularity.
Furthermore, although in the present study the difference in the level of muscularity in the non-muscular was significant (an of average of 1.7 vs. 3.5 on a 5-point Likert scale), it would be interesting to compare models who are on either ends on this spectrum. This may cause more significant differences in the effect of exposure to these images.
Moreover, state social comparison was found to play a marginally significant part in explaining the relationship between exposure to the non-muscular versus muscular condition and body image. Men who were exposed to a muscular model did report higher levels of social comparison than men who were exposed to a non-muscular model, meaning the effect was in the expected direction. The level of social comparison men engaged in also had a marginally significant effect in the expected direction on body image. Men who engage in higher levels of social comparison scored lower on body image. These results are interesting, because although they are not fully significant they do provide some inclination that social comparison has a mediating effect on the relationship between exposure to (non-)muscular models and body image. With a larger sample, these effects could prove to be significant.
Lastly, the study found no evidence that body discrepancy has any effect on the relationship between exposure to the different conditions and state social comparison and the relationship between state social comparison and body image. Comparing these results to the study by Bessenoff (2006), it suggests that men have different boundary conditions than women when it comes to exposure to a body ideal and its effect on body image. However, drawing any real conclusions on the boundary conditions is hard based on these results, as there was no support for the main effect of exposure to the two conditions on body image. It is possible that in different circumstances men do have similar boundary conditions as women. This is worth further exploring in follow-up studies.
Exploratory results. Next to planned analyses, several exploratory models were
estimated in order to provide additional information to follow-up research. Diedrichs and Lee (2010) found that the clothes advertised by average-sized models were equally appealing as clothes that were advertised by muscular models. However, the present study found that participants have a more positive attitude towards clothes advertised by a muscular model than clothes advertised by a non-muscular model. It should be noted though, that these
contrary results could be due to the fact that Diedrichs and Lee used exactly the same model and clothing and manipulated the model in Photoshop, whereas the present study used different models and similar but different clothing items. However, models and clothing items were chosen and edited to be as similar as possible.
Furthermore, the exploratory tests also found that exposure to a muscular model had a significant positive effect on the attitude towards the clothing, which in turn had a significant positive effect on social comparison. As mentioned before, social comparison had a
marginally significant effect on body image. This suggests that the attitude towards clothing could be a mediator of the effect of exposure condition on social comparison. A possible explanation for this could be that if one likes a clothing item, one starts thinking more about how this clothing item would look on themselves and for that purpose starts comparing themselves to the model. Although this was an unexpected effect, it provides an interesting new area for future research. It indicates that the focus should perhaps be less on the model’s body but more on the clothing aspect.
Next, these results also suggest that age and affect are correlated with body image. The scores of participants on body image decreased as age increased. A possible explanation for this could be that as people get older, their actual body becomes further removed from the ideal body presented in the media. However, further research is needed to draw any real conclusions on this relationship.
Lastly, the exploratory tests found that participants who scored higher on positive affect scored higher on body image, whereas participants who scored higher on attention and negative affect scored lower on body image. Especially the negative effect of attention on body image is interesting here, as attention was originally part of the positive affect scale. One would therefore expect it to have a positive effect on body image similar to the positive
and focused felt more negative about their body. Attention was also tested as a moderator of the effect of the exposure condition on body image, but this yielded no significant results. A possible explanation for this is that those participants that had a higher level of attention put more thought into answering the body image scale and this caused them to report lower scores independent of the condition they were exposed to.
Limitations. Since existing photos were used to enhance the external validity of the
study, some concessions had to be made. Although it would have likely been beneficial to the experiment if the models had shown more skin, such as a bare torso, such images are almost never used in real-world advertisements for non-muscular models. Therefore, in order to keep the photos in the conditions similar, both sets of photos are of models wearing t-shirts or sleeveless tops. The tops they wear are not exactly the same, but they are mostly neutral and in one color.
Furthermore, the original aim was to get 160 respondents for this study. However, since recruiting participants proved to be more difficult than expected and some participants had to be removed from the study, the final sample consisted of 139 participants, which puts limitations on the ability of this study to find small effects.
Lastly, about 60% of the participants answered correctly when asked what they thought the research was about. Most of them had at least the idea it had something to do with body image and muscularity. This means the cover story was not entirely successful in hiding the true purpose of the study. It is possible that because participants were aware of what was being measured, they became too conscious of their answers and did not answer the questions intuitively.
Conclusion. This study provides no evidence of a direct effect of exposure condition
on body image, but it does indicate that there is a possible mediating effect of social
research and suggests that perhaps it is not the muscularity aspect of models that elicits social comparison and has a negative effect on body image but that we should look for alternative explaining factors of advertisement’s effect on body image such as the clothing the models wear.
Furthermore, this study suggests that men and women are not directly comparable when looking at the effect of advertisements on body image and the boundary conditions of this effect. This means that further research needs to be done to find out exactly how this relationships works for men, but it is clear that men should be treated separately from women in future research and intervention efforts.
Although this research could unfortunately not provide a concluding answer to the research question, I am confident that this study is a helpful contribution to the further
exploration on how men are affected by body ideals and what the boundary conditions of this effect are.
References
Agliata, D., & Tantleff-Dunn, S. (2004). The impact of media exposure on males' body image. Journal of Social and Clinical Psychology, 23(1), 7-22.
Barlett, C. P., Vowels, C. L., & Saucier, D. A. (2008). Meta-analyses of the effects of media images on men's body-image concerns. Journal of social and clinical psychology, 27(3), 279-310.
Bessenoff, G. R. (2006). Can the media affect us? Social comparison, self-discrepancy, and the thin ideal. Psychology of women quarterly, 30(3), 239-251.
Brennan, M. A., Lalonde, C. E., & Bain, J. L. (2010). Body image perceptions: Do gender differences exist. Psi Chi Journal of Undergraduate Research, 15(3), 130-138. Cafri, G., & Thompson, J. K. (2004). Evaluating the convergence of muscle appearance
attitude measures. Assessment, 11(3), 224-229.
Cafri, G., Thompson, J. K., Ricciardelli, L., McCabe, M., Smolak, L., & Yesalis, C. (2005). Pursuit of the muscular ideal: Physical and psychological consequences and putative risk factors. Clinical psychology review, 25(2), 215-239.
Cash, T. F., Fleming, E. C., Alindogan, J., Steadman, L., & Whitehead, A. (2002). Beyond body image as a trait: The development and validation of the Body Image States Scale. Eating Disorders, 10(2), 103-113.
Cheng, A. (2017, October 3) There's a Reason You're Seeing More Curvy Models in Mass Market Fashion Campaigns. Retrieved from
https://www.glamour.com/story/curvymodels-commercial-fashion-casting
Davis, K. (2013). Reshaping the female body: The dilemma of cosmetic surgery. Routledge. Diedrichs, P. C., & Lee, C. (2010). GI Joe or Average Joe? The impact of average-size and
muscular male fashion models on men's and women's body image and advertisement effectiveness. Body image, 7(3), 218-226.
Duggan, S. J., & McCreary, D. R. (2004). Body image, eating disorders, and the drive for muscularity in gay and heterosexual men: The influence of media images. Journal of homosexuality, 47(3-4), 45-58.
Dittmar, H. (2009). How do “body perfect” ideals in the media have a negative impact on body image and behaviors? Factors and processes related to self and identity. Journal of Social and Clinical Psychology, 28(1), 1-8
Festinger, L. (1954). A theory of social comparison processes. Human relations, 7(2), 117-140.
Fisher, E., Dunn, M., & Thompson, J. K. (2002). Social comparison and body image: An investigation of body comparison processes using multidimensional scaling. Journal of Social and Clinical Psychology, 21(5), 566-579.
Groesz, L. M., Levine, M. P., & Murnen, S. K. (2002). The effect of experimental presentation of thin media images on body satisfaction: A meta‐analytic review. International Journal of eating disorders, 31(1), 1-16.
Grogan, S., Williams, Z., & Conner, M. (1996). The effects of viewing same-gender photographic models on body-esteem. Psychology of Women Quarterly, 20(4), 569-575.
Hargreaves, D. A., & Tiggemann, M. (2009). Muscular ideal media images and men's body image: Social comparison processing and individual vulnerability. Psychology of Men & Masculinity, 10(2), 109.
Higgins, E. T. (1987). Self-discrepancy: a theory relating self and affect. Psychological review, 94(3), 319.
Halliwell, E., & Dittmar, H. (2006). Associations between appearance-related
emotional eating: A comparison of fixed-item and participant-generated self-discrepancies. Personality and Social Psychology Bulletin, 32(4), 447-458.
Jung, J., Lennon, S. J., & Rudd, N. A. (2001). Self-schema or self-discrepancy? Which best explains body image?. Clothing and Textiles Research Journal, 19(4), 171-184. Karazsia, B. T., & Crowther, J. H. (2009). Social body comparison and internalization:
Mediators of social influences on men's muscularity-oriented body dissatisfaction. Body Image, 6(2), 105-112.
Latané, B. (1966). Studies in social comparison—Introduction and overview. Journal of Experimental Social Psychology, 1, 1-5.
Leit, R. A., Pope Jr, H. G., & Gray, J. J. (2001). Cultural expectations of muscularity in men: The evolution of Playgirl centerfolds. International Journal of Eating Disorders, 29(1), 90-93.
Leit, R. A., Gray, J. J., & Pope Jr, H. G. (2002). The media's representation of the ideal male body: A cause for muscle dysmorphia?. International Journal of Eating
Disorders, 31(3), 334-338.
Lorenzen, L. A., Grieve, F. G., & Thomas, A. (2004). Brief report: Exposure to muscular male models decreases men’s body satisfaction. Sex Roles, 51(11-12), 743-748. Ogden, J., & Mundray, K. (1996). The effect of the media on body satisfaction: The role of
gender and size. European Eating Disorders Review: The Professional Journal of the Eating Disorders Association, 4(3), 171-182.
Schilder, P. (1950). The image and appearance of the human body. New York: International Universities Press
Silvia, P. J., & Duval, T. S. (2001). Objective self-awareness theory: Recent progress and enduring problems. Personality and Social Psychology Review, 5(3), 230-241. Stice, E., Mazotti, L., Krebs, M., & Martin, S. (1998). Predictors of adolescent dieting
behaviors: A longitudinal study. Psychology of addictive behaviors, 12(3), 195. Talbot, D., Cass, J., & Smith, E. (2019). Visual Body Scale for Men (VBSM): Validation of a
new figural rating scale to measure perceived‐desired body discrepancy in men. Journal of clinical psychology, 75(3), 462-480.
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, 23(1), 23-44.
Vartanian, L. R. 2. (2012). Self-discrepancy theory and body image. Encyclopedia of body image and human appearance, 2, 711-717.
Wiseman, C. V., Gray, J. J., Mosimann, J. E., & Ahrens, A. H. (1992). Cultural expectations of thinness in women: An update. International Journal of Eating Disorders, 11(1), 85-89.
Wood, J. V. (1989). Theory and research concerning social comparisons of personal attributes. Psychological bulletin, 106(2), 231
