Eating disorders represent serious public mental health problems that are attracting increasing attention (Varsou, Reference Varsou2010). The prevalence of eating disorders can be underestimated when only individuals who seek treatment and meet the diagnostic criteria for anorexia nervosa (AN) and bulimia nervosa (BN) are considered (Faravelli et al, Reference Faravelli, Ravaldi, Truglia, Zucchi, Cosci and Ricca2006; Stice, Marti, Shaw, & Jaconis, Reference Stice, Marti, Shaw and Jaconis2009). Therefore, models that explain partial symptoms, rather than models for complete AN or BN, as described in the Diagnostic and Statistical Manual of Mental Disorders (DSM-IV-TR), should be developed (American Psychiatric Association, 2000). A few models have been proposed to identify risk factors for and explain the development of eating disorders (Collier & Treasure, Reference Collier and Treasure2004; Gleaves, Brown, & Warren, Reference Gleaves, Brown and Warren2004; Kim, Reference Kim2012). The two-pathway Stice model (Stice, Reference Stice1994; Stice, Reference Stice2001) is most likely one of the most comprehensive measures to explain the development of bulimic symptoms and integrates the following factors: (1) sociocultural pressure to be thin; (2) ideal-thin body internalization; (3) body dissatisfaction; (4) negative emotions; and (5) restrained eating (Figure 1). The dual pathway model refers to the negative emotions and restrained eating that represent the final predictors of bulimic symptoms (Stice, Reference Stice2001; Stice, Bohon, Marti, & Fisher, Reference Stice, Bohon, Marti and Fischer2008). According to this model, social pressure and thin body internalization promotes body dissatisfaction, which increases negative emotions and dieting behaviors that trigger bulimic pathology (Ouwens, van Strien, van Leeuwe, & van der Staak, Reference Ouwens, van Strien, van Leeuwe and van der Staak2009; Stice, Reference Stice2001). The model predicts that bulimic symptoms are associated with the social pressure to have a thin body leads to the internalization of the ideal of the thin body and therefore body dissatisfaction associated with negative emotions and restrained eating as a mean to reduce such discrepancy. Restrained eating and dieting behaviors, in particular, have drawn increased attention in the literature with respect to the development of clinical symptoms (for example, Ouwens et al., Reference Ouwens, van Strien, van Leeuwe and van der Staak2009; Stice, Pressnell, & Spangler, Reference Stice, Pressnell and Spangler2002). Numerous studies have noted the core importance of eating restraint itself as a risk factor for the development and maintenance of eating disorders (Andrés & Saldaña, Reference Andrés and Saldaña2014; Cogan & Ernsberger, Reference Cogan and Ernsberger1999; Stice, Davis, Miller, & Marti, Reference Stice, Bohon, Marti and Fischer2008; Stice, Martinez, Presnell, & Groesz, Reference Stice, Martinez, Presnell and Groesz2006). Indeed, many symptoms of eating disorders are directly associated with dieting status as described in the “starvation study” by Keys (Reference Keys1950).

Figure 1.

The two pathways in Stice’s model.

The influence of the media and peers on the social pressure to be thin has been intensively studied. An extensive review of 25 studies analyzing the influence of the thin ideal model presented in mass media reveled a significant increase in body dissatisfaction after viewing thin media images rather than average or super-size models or non-human objects (Groesz, Levine, & Murnen, Reference Groesz, Levine and Murnen2002). Verbal peer pressure to be thin has also been reported in an experimental study in which 120 females exposed to so-called “fat talk” reported increased body dissatisfaction compared with a control group in which a neutral topic was discussed (Stice, Maxfield, & Wells, Reference Stice, Maxfield and Wells2003; Stice & Whitenton, Reference Stice and Whitenton2002). Similarly, exposure to negative information or feedback about body weight or appearance has been found to positively correlate with body dissatisfaction compared with exposure to control (non-appearance related) information (Lavin & Cash, Reference Lavin and Cash2000).

In sum, body dissatisfaction, which has been defined in various ways, is a key variable in Stice’s model. Keeton, Cash, and Brown (Reference Keeton, Cash and Brown1990) defined body dissatisfaction as the discrepancy between the self body image and the ideal body image. Body dissatisfaction differs from body image distortion, which can be understood as the discrepancy between the self body image and the real body image. Cash, Edwin, and Deagle (Reference Cash, Edwin and Deagle1997) defined body dissatisfaction as a cognitive judgment bias rather than a perceptual distortion based on the overestimation of body size. Body dissatisfaction refers to the attitudinal dimension of the body image and body distortion to the perceptual dimension (Skrzypek, Wehmeier, & Remschmidt, Reference Skrzypek, Wehmeier and Remschmidt2001).

The high prevalence of body dissatisfaction was first reported almost three decades ago and was referred to as “normative discontent” (Rodin, Silberstein, & Striegel-Moore, Reference Rodin, Silberstein, Striegel-Moore and Sonderegger1985). A decade later, Brumberg (Reference Brumberg1997) reported that 13.53% of American girls were dissatisfied with their body at 13 years old; by age 17, the percentage had risen to 78%. In the United States, the ideal body is thin, as suggested by the trend in Playboy centerfolds models from 1959 to 1978, which exhibited significant decreases in bust and hip size (Garner, Garfinkel, Schwartz, & Thompson, Reference Garner, Garfinkel, Schwartz and Thompson1980). Consequently, expenses related to the goals of weight loss and meeting the cultural standards of thinness and beauty in western society have increased. For example, the number of cosmetic procedures has tripled in the past decade according to the American Academy of Cosmetic Surgery (Shroff & Thompson, Reference Shroff and Thompson2006). The drive for thinness is less pronounced in males according to the trend reflected by male Hollywood stars, who are more muscular now than actors in the 1940s and 1950s (Goldfield, Blouin, & Woodside, Reference Goldfield, Blouin and Woodside2006).

With respect to gender differences, literature has traditionally focused on females more than males (Shepherd & Ricciardelli, Reference Shepherd and Ricciardelli1998; Stice, Reference Stice1998; Stice, Reference Stice2001; van Strien, Reference van Strien1996) because females exhibit higher body dissatisfaction than do males (Brennan, Lalonde, & Bain, Reference Brennan, Lalonde and Bain2010; Cohane & Pope, Reference Cohane and Pope2001; Zaccagni, Masotti, Donati, Mazzoni, & Gualdi-Russo, Reference Zaccagni, Masotti, Donati, Mazzoni and Gualdi-Russo2014). However, men are not free from body dissatisfaction. Three decades ago, Cash, Winstead and Jada (Reference Cash, Winstead and Janda1986) found that the level of body dissatisfaction among men had increased since 1972; therefore, data must be continuously updated, and the male level of body dissatisfaction also should be analyzed in Stice’s Model, which has been suggested to also be useful for males (Halliwell & Harvey, Reference Halliwell and Harvey2006).

Gender differences have also been observed with respect to eating restraint or dieting behaviors, which are the best predictors of developing eating disorders, although negative emotion pathways also contribute (van Strien, Engles, van Leeuwe, & Snoek, Reference van Strien, Engels, van Leeuwe and Snoek2005). Drewnowski and Yee (Reference Drewnowski and Yee1987) reported that the percentage of women who followed a low-calorie diet in the previous month was more than double the number of males. Although women tend to diet to lose weight, men exercise to control weight and gain muscle.

Stice’s model focuses on the development of bulimic symptoms, and such behaviors are also observed among the non-clinical population. However, the application of this model in the non-clinical population would benefit our current knowledge of the risk factors associated with eating disorders and would therefore benefit future programs for the prevention of eating disorders both in males and females.

The aim of this study was to explore and analyze gender differences in the selection of variables involved in Stice’s Model of eating disorders in a non-clinical population.

The following hypotheses were addressed. Females will report higher body dissatisfaction (BSQ score) and dieting behaviors (DIS score). Females will perceive more social pressure to be thin (PSPS score) and will report higher thin-body internalization (IBSS score). Furthermore, differences in body dissatisfaction, dieting behavior and the perception of the pressure to be thin will be analyzed with respect to BMI levels. First, participants with higher BMIs will score higher with respect to body dissatisfaction regardless of their gender and will report more weight loss behaviors, such as dieting (DIS score), and to perceive more pressure to be thin than those subjects with lower BMIs.

RESULTS

Gender effect

Table 1 presents gender differences in the mean and standard deviation of each measurement.

Table 1.

Gender differences in the means and standard deviations for every scale

Note: * = p < 0.5.

Females scored significantly higher than males with respect to body dissatisfaction, as measured in BSQ scores (F _1,151 = 57.09, p < .05). Females also scored higher in dieting as a weight loss measure compared with males. This difference in DIS scores was significant (F _1,151 = 26.35, p < .05). Similarly, females scored higher than males with respect to perceived social pressure (PSPS scores). The ANOVA revealed significant gender differences (F _1,151 = 44.91, p < .05). However, no significant gender differences were observed with respect to body thin internalization (IBSS-R score) when an ANOVA was conducted (F _1,151 = 0.0, p > .05).

BMI effect

Table 2 presents gender differences in variables for high or low BMI.

Table 2.

BMI differences in the means and standard deviations for every scale

Note: * = p < .5.

High BMI males and females reported higher body dissatisfaction (BSQ scores) than did low BMI males and females. An ANOVA revealed that there was a significant BMI effect F _1,151 = 21.65, p < .05. In general, subjects with greater BMI perceived more social pressure to be thin compared with low BMI subjects. An ANOVA revealed that this difference was significant: F _1,151 = 15.08, p < .05. No significant differences were reported with respect to thin-body ideal internalization for subjects with low or high BMI: F _1,151 = 1.45, p > .05.

Interaction between gender and BMI

The mean BMI of the males was 25.65 (SD = 5.96), whereas the mean BMI of the females was 22.85 (SD = 4.95), representing a medium effect size with Cohen’s d = 0.51.

An interaction was observed among gender, BMI and dietary restraint. Only a few males reported dieting, and those with higher BMI values (M = 15.4, SD = 6.4) reported more weight loss-oriented behaviors (higher score on DIS) than did males with low BMI values (M = 12.2, SD = 5.2). However, females engaged in dieting behaviors independent of their BMI values. Both females with high (M = 20.7, SD = 7.0) and low BMI (M = 20.1, SD = 8.0) values reported weight loss behaviors (higher score on DIS): F _1,151 = 2.24, p > .05. With respect to the social pressure to be thin, perceived social pressure from the media was significantly higher among females with high BMI values (M = 26.0, SD = 9.2) than among those with lower BMI values (M = 19.6, SD = 6.8) and higher than either high BMI (M = 16.3, SD = 5.7) or low BMI (M = 13.7, SD = 3.8) values (F _1,151 = 4.44, p < .05).

Gender differences and Model fit by SEM

Model fit was satisfactory. Chi-square statistic was non-significant (CMIN = 11.023; df = 6; n.s.; CMIN/DF = 1.837), and values for other fit indexes were acceptable (GFI = .969; AGFI = .895; AIC = 39.023; RMSEA = .074). With respect to gender differences, all coefficients were similarly moderate and statistically significant for both men and women, with the exception of the regression weight of IBSS score on BSQ score, which was positive, and statistically significant for women (r = .18; p < .05), but close to zero and non-significant for men (r = –.05; n. s.), thus indicating a weak but significant effect of Ideal thin body internalization on Body dissatisfaction for women, but no effect at all for men. This makes squared multiple correlation for BSQ a bit larger for women (R ² = .37) than for men (R ² = .32), given that both PSPS and IBSS scores are significant predictors of BSQ for women, whereas PSPS is the only predictor of BSQ scores for men. Anyway, this difference disappears when examining the squared multiple correlation for DIS, which is slightly larger for men (R ² = .46) than for women (R ² = .44) (Figure 2).

Figure 2.

Path diagram for the variables of the Stice’s model included in the study, with coefficients obtained for each gender.

* = p < .05; *** = p < .001.

In summary, the path analysis results support the notion that the Stice’s Dual Pathway model looks appropriate for women, but not for men, given that Ideal body internalization is not a significant factor in the development of body dissatisfaction for men. These results are consistent with the correlations obtained in Table 3, which show positive and significant correlations between IBSS and both BSQ (r = .334; p < .01) and DIS (r = .320; p < .01) scores for women, but not for men.