Abstract
Because research on body shame has predominantly focused on women, the consequences of male body shame for gender relations have been under-investigated. Following up on preliminary findings suggesting that men high on body shame were hostile toward women, in two experiments, we uniquely observed that body shame predisposes men to sexual aggression when they react negatively to masculinity threats. In Experiment 1, men rejected by a female confederate for being unattractive showed rape proclivity to the extent they were high on both body shame and post-rejection negative affect. In Experiment 2, the same pattern emerged on the part of men rejected by a female (but not a male) confederate for ostensibly being gay. In concert, the findings suggest that men’s body shame is an overlooked factor in sexual aggression, which has implications for extant rape theories and precarious manhood theory.
As one of the most common human rights abuses, sexual violence (rape and sexual assault) victimizes women at much higher rates than men. In the United States, approximately one in five female undergraduates are victimized by sexual assaults (B. S. Fisher, Cullen, & Turner, 2000; Karjane, Fisher, & Cullen, 2005), although prevalence is difficult to estimate because around 65% of sexual assaults are not reported to police (Langton, Berzofsky, Krebs, & Smiley-McDonald, 2012). Not all rape victims are female, but men are overwhelmingly the perpetrators of sexual assault and women its victims (99% and 91%, respectively; Centers for Disease Control and Prevention, 2008; Greenfield, 1997). Therefore, it is important to investigate factors that promote male sexual aggression.
Not surprisingly, hostility toward women is one such factor in male-on-female aggression, which is predicted by trait aggression and acceptance of interpersonal violence, but it is non-specific to sexual aggression (Anderson & Anderson, 2008; Malamuth, 1988). Because rape is a fully embodied act of violence, we might expect body orientations to characterize at least some male perpetrators. Indeed, men who automatically dehumanize women’s bodies (as animals or objects) also score high on rape proclivity (Rudman & Mescher, 2012). We might also expect men’s orientation to their own bodies to play a role; however, male body shame has been an overlooked element in sexual assault. Researchers typically employ men as a comparison group to women, finding that women suffer higher rates of body shame, which covaries with women’s psychological distress (e.g., depression and eating disorders; Calogero, Davis & Thompson, 2005; Fredrickson & Roberts, 1997; Moradi & Huang, 2008; Tiggemann & Kuring, 2004). By contrast, we propose that, as an individual difference variable, men’s body shame can be a risk factor for women by predisposing men to sexual aggression.
Masculinity, Body Shame, and Sexual Aggression
A muscular physique is so strongly associated with masculinity that men without it are viewed as feminine (not merely weak; Grogan & Richards, 2002; Weinke, 1998). Surveys indicate that 50% to 71% of male undergraduates are dissatisfied with their bodies, and that 90% desire to be more muscular (Frederick et al., 2007). While a drive for muscularity is associated with poor self-esteem in male adolescents (McCreary & Sasse, 2000), men of all ages stress the importance of having a muscular physique (E. Fisher, Dunn, & Thompson, 2002; Jones, 2001; McCreary & Sasse, 2002). Moreover, men report poor body image following exposure to images of idealized male bodies (Morry & Staska, 2001) as well as idealized female bodies (Lavine, Sweeney, & Wagner, 1999; see also Hunt, Gonsalkorale, & Murray, 2013). Given the ubiquity of such images, men—like women—may be at risk of developing chronic body shame (Frederick et al., 2007; Hunt et al., 2013). Indeed, body shame has been linked to internalized cultural standards of male attractiveness in heterosexual men (Parent & Moradi, 2011).
Some indirect evidence supports our thesis that body shame can predispose men for sexual aggression. Men’s drive for muscularity has been associated with endorsing sexist attitudes, traditional gender roles, and a tendency to objectify women (McCreary, Saucier, & Courtenay, 2005; Steinfeldt, Gilchrist, Halterman, Gomory, & Steinfeldt, 2011; Swami & Voracek, 2012). Moreover, male partners were more domestically violent to the extent their body image was poor (Shelton & Liljequist, 2002). In an experiment, men who were told they performed poorly on a strength task subsequently delivered more painful electric shocks to a criticizing female confederate, as compared to men paired with a silent female confederate (Richardson, Leonard, Taylor, & Hammock, 1984). However, we could find no prior research documenting a direct link between male body shame and sexual aggression.
Our discovery of this relationship was serendipitous, arising from research investigating dehumanization of women’s bodies (Rudman & Mescher, 2012). Reasoning that women’s own self-objectification might moderate dehumanization, we administered the Objectified Body Consciousness Scale (McKinley & Hyde, 1996) in two studies, which includes an index of body shame (described in the “Method” section). Measures of gender prejudice included hostile and benevolent sexism (Glick & Fiske, 1996), negative attitudes toward rape victims (Ward, 1988), rape myth acceptance beliefs (Burt, 1980), and rape proclivity, which combines two subscales (willingness to rape and force a sexual partner) from the Attraction to Sexual Aggression Inventory (Malamuth, 1989). Women did not support our hypothesis (i.e., women who self-objectified were not more likely to implicitly dehumanize women). Instead, unexpected positive relationships emerged between men’s body shame and all of these measures, presented in Table 1. That is, men high on body shame were also hostile toward women. For women, these relationships were either negligible (Study 1) or confined to hostile sexism and attitudes toward rape victims (Study 2). The novelty of this unexpected pattern for men warranted further examination.
Correlates of Body Shame by Participant Gender (Pilot Research).
Note. ATRV is the negative attitudes toward female rape victims. Rape proclivity is the willingness to rape or force a sexual partner against his or her will (item was worded differently for men and women). BS = benevolent sexism; RMA = rape myth acceptance.
p < .05. **p < .01. ***p < .001.
Additional preliminary research also revealed that men high on body shame were low on self-esteem (Heatherton & Polivy, 1991), r(210) = −.62, p < .001, and masculinity, using the Personality Attributes Questionnaire (PAQ; Spence, Helmreich, & Stapp, 1975), r(40) = −.53, p < .001. They did not score high on PAQ femininity items, except for seeks social approval, r(40) = .31, p < .05, and feelings easily hurt, r(40) = .48, p = .001. We also found a positive relationship between men’s body shame and precarious manhood beliefs (e.g., “It is fairly easy for a man to lose his status”; Burnaford, Weaver, Bosson, & Vandello, 2009), r(124) = .21, p = .02. 1 The overall pattern suggests that men high on body shame may be chronically at risk of masculinity threat.
Masculinity Threat and Aggression
Precarious manhood theory posits that, in contrast to womanhood, men are not assured of their masculinity status as a result of puberty; rather, it must be continuously earned and defended (Vandello, Bosson, Cohen, Burnaford, & Weaver, 2008). Consequently, it is easier to threaten men’s gender identity than it is to threaten women’s (e.g., Bosson, Prewitt-Freilino, & Taylor, 2005; Rudman, Dohn, & Fairchild, 2007). Moreover, masculinity threat can cause men to aggress. Men given feminine feedback on a bogus personality test threw harder punches, as measured by an electronic sensor, compared with controls (Vandello et al., 2008; see also Bosson & Vandello, 2011; Bosson, Vandello, Burnaford, Weaver, & Wasti, 2009). Women who challenge male dominance may be especially at risk in this situation (Dall’Ara & Maass, 2000). For example, masculinity-threatened men harassed a feminist female confederate more so than a traditional female confederate (by emailing pornographic photos; Maass, Cadinu, Guarnieri, & Grasselli, 2003). If scoring low on a strength task provides a masculinity threat, then men’s delivery of painful shocks to a criticizing female confederate is evidence that masculinity threat promotes male-on-female aggression (Richardson et al., 1984).
However, our concern is with the link between men’s body shame and hostility toward women shown in Table 1, observed without any acute threat. We propose that men high on body shame may be under chronic masculinity threat, but for this idea to be tenable, masculinity threat should have consequences for men’s body image. Consistent with this view, men under acute masculinity threat (via false feedback on a personality test) showed a poorer body image than men who experienced affirmed masculinity (Hunt et al., 2013). Furthermore, men who lost to a female confederate in an anagram competition showed lowered body satisfaction, relative to those who lost to a male confederate (Mills & D’Alfonso, 2007). The present research afforded a test of whether acute masculinity threats have an effect on male body shame, but our overarching goal was to investigate whether men chronically high on body shame would react to such threats with sexual aggression when the perpetrator was female.
The Role of Female Rejection
Female sexual rejection appears prominently in extant rape theories (for a review, see Baumeister, Catanese, & Wallace, 2002) and male rapists typically report that conflicts with women arising from rejection (actual or perceived) precipitated sexual assaults (McKibben, Proulx, & Lusignan, 1994). However, whether female rejection heightens masculinity threat for men high on body shame is unknown. Rape theorists have posited that men high on narcissism (Baumeister et al., 2002), male entitlement (Hill & Fischer, 2001), or hyper-masculinity (Mosher & Anderson, 1986; Mosher & Sirkin, 1984) may be especially likely to respond to female rejection with rape. These theories do not seem applicable to men high on body shame, given our pilot work suggesting that they are low on self-esteem and masculinity, while high on approval seeking and easily hurt feelings. However, Malamuth speculated that insecure men fear rejection from women and may engage in sexual aggression to reassert control because “The power that a woman may have by virtue of her sex appeal may be particularly threatening to [such men]” (Malamuth, Linz, Heavey, Barnes, & Acker, 1995, p. 354). To our knowledge, the current Experiment 1 provides the first test of this hypothesis. Specifically, we subjected some men to female rejection (refusal to partner him for a “dating study”) to examine whether men high on body shame who reacted negatively to rejection would respond with rape proclivity.
Another factor likely to heighten masculinity threat is being misclassified as gay (Bosson et al., 2005; Bosson, Taylor, & Prewitt-Freilino, 2006). Because heterosexuality is a strong prescription for men (Herek, 1989; Kimmel, 2004), being called “gay” is often deemed insulting to heterosexual men (Burn, 2000; Thurlow, 2001). Indeed, compared with women, men report more sexual prejudice (e.g., Herek, 2002), likely because they strive to differentiate themselves from gay men to protect their masculinity (Carnaghi, Maass, & Fasoli, 2011; Falomir-Pichastor & Mugny, 2009). Consistent with this view, men under masculinity threat showed more derogation of a feminine (but not a masculine) gay man, compared with unthreatened counterparts (Glick, Gangl, Gibb, Klumpner, & Weinberg, 2007; see also Talley & Bettencourt, 2008). Therefore, Experiment 2 used rejection “for being gay” as another instantiation of masculinity threat.
The Role of Emotion Regulation
Why do some men respond poorly to insults and female rejection, while others remain unaffected (Malamuth et al., 1995)? We theorized that responses might be related to men’s emotional processing. Although social rejection often increases negative affect and emotional distress (Leary, Twenge, & Quinlivan, 2006), this is not always the case. For some people, rejection causes emotional numbness (Blackhart, Nelson, Knowles, & Baumeister, 2009), referred to as a “shock” reaction (Twenge, Catanese, & Baumeister, 2003; Werner, Kerschreiter, Kindermann, & Duschek, 2013). Others may deliberately attempt to inhibit emotional distress post rejection (Larsen, 2000), spurred by sensitivity to physiological cues corresponding to negative affect (interoceptive sensitivity; Werner, Duschek, Mattern, & Schandry, 2009). If men high on body shame are low on interoceptive sensitivity (i.e., are “out of touch” with their bodies), they may be less able to self-regulate when they are upset, compared with men low on body shame. Therefore, we anticipated that only men high on body shame who reacted negatively to female rejection would respond with sexual aggression, reasoning that this response would be most likely for participants who felt significant emotional distress that they were unable to regulate.
Overview and Hypotheses
In Experiment 1, men were rejected by a woman as a partner for a dating study for being unattractive. In Experiment 2, men were rejected due to “being gay” by either a male or female confederate. Compared with (not rejected) control men, we expected both types of rejection to elicit more sexual aggression from men high on body shame but only when the perpetrator was female (not male) and only if they were unable to down-regulate negative affect.
Across experiments, we measured body shame using the same measure employed by the pilot research. However, we used two measures of sexual aggression to determine the generality of our results. In Experiment 1, we adopted the rape proclivity index used in the pilot research. In Experiment 2, we relied on a rape behavioral analogue (RBA; Rudman & Mescher, 2012). Finally, we examined the discriminant validity of body shame by including other shame-related measures in each experiment.
Experiment 1
To model masculinity threat based on romantic rejection, an attractive woman rejected male participants as dating partners on the basis of finding them unattractive after viewing their photo. In the control condition, her decision was unavailable “due to a computer failure.” For rejected men, we predicted that body shame and negative affect together would elicit sexual aggression. For men in the control group, this pattern should not appear because there was no threat to their masculinity.
To measure sexual aggression, we used a rape proclivity index that correlates with men’s sexual arousal when viewing depictions of rape (Malamuth, Haber, & Feshbach, 1980; Malamuth, Heim, & Feshbach, 1980) and with men’s body shame (see Table 1). To test the discriminate validity of body shame, we included two measures of generalized shame proneness (negative self-evaluation and withdrawal in response to shame-inducing situations; Cohen, Wolf, Panter, & Insko, 2011) and an author-designed index of sexual shame.
Method
Participants
Participants (N = 127) were heterosexual men who completed the experiment in exchange for credit toward their Introductory Psychology research requirement. Six participants were excluded due to missing data and/or sexual orientation. Of the final sample, 63 (50%) were White, 45 (35%) Asian, 8 (6%) Latino, 5 (4%) Black, and 6 (5%) selected another option.
Materials
Pretesting the confederate photo
A pilot sample (N = 239, 121 men) rated the attractiveness of 21 portraits of college-aged White women posed against neutral backgrounds downloaded from the Internet (gettyimages.com). Participants rated the physical attractiveness of each image (presented randomly) on a scale ranging from 1 (very unattractive) to 10 (very attractive). The female phantom’s image selected was high (but not extreme) on attractiveness, M = 7.28 (SD = 1.80).
Negative affect
To reduce reactivity post rejection, participants were told that their mood would be measured at random intervals during the session; in fact, the measure was always completed first and immediately after the rejection manipulation. Participants indicated the extent to which they felt each emotion “right now” on a scale ranging from 1 (not at all) to 5 (very much). Negative emotions were hurt, insulted, offended, ashamed, angry, disgusted, sad, and hostile. Positive emotions (happy, confident, calm, proud, and amused) were filler items. A principle components factor analysis of this measure at Times 1 and 2 revealed only a positive and a negative factor in each case. Therefore, negative emotions at Time 2 were combined to form the negative affect index (α = .92). High scores reflect more negative affect following the rejection manipulation.
Body shame
Participants completed the body shame subscale from the Objectified Body Consciousness Scale (McKinley & Hyde, 1996). It consists of 13 items anchored on a scale ranging from 1 (strongly disagree) to 7 (strongly agree). Sample items include, “I am ashamed by the size and shape of my buttocks,” “I feel like I must be a bad person when I don’t look as good as I could,” and “Overall, I am comfortable with how my body looks” (reverse scored). Items were averaged so that high scores reflected more body shame (α = .81).
Shame proneness
Participants responded to the Guilt and Shame Proneness Scale (GASP; Cohen et al., 2011) on a scale ranging from 1 (very unlikely) to 7 (very likely). The subscales of primary interest, shameful self-evaluation and shame-withdrawal, each consist of four shame-inducing scenarios. An example of the self-evaluation subscale is,
You rip an article out of a journal in the library and take it with you. Your teacher discovers what you did and tells the librarian and your entire class. What is the likelihood that this would make you would feel like a bad person?
An example of the shame-withdrawal subscale is, “You take office supplies home for personal use and are caught by your boss. What is the likelihood that this would lead you to quit your job?” Items were averaged so that high scores reflected more shameful self-evaluation (α = .60) or shame-withdrawal (α = .58). 2
Sexual shame
To provide additional discriminant validity for body shame, participants completed eight author-designed items on a scale ranging from (1) strongly disagree to (5) strongly agree. Sample items include “I worry that sexually, I am not like other men,” and “I sometimes feel ashamed of my own sexual inclinations.” Items were averaged so that high scores reflected greater sexual shame (α = .79).
Rape proclivity
Participants completed Malamuth’s (1989) Attraction to Sexual Aggression Scale, anchored at 1 (very unlikely) to 5 (very likely). Participants indicated whether, if they could be assured of never being known or punished, they were aroused by, attracted to, or would be likely to commit rape. Three other items replaced “rape” with “force a sexual partner to do something they did not want to do.” These six items were averaged to form the rape proclivity index (α = .84).
Procedure
Participants were recruited for a study investigating “the factors that build effective teamwork,” in which they believed they would compete with a partner over a networked computer in teamwork building tasks, with the most effective teams earning a chance to win US$100. Upon arriving at the lab, the experimenter took a digital photo of the participant, ostensibly for upload to his networked partner, and escorted participants to a private cubicle with a computer. The experimenter provided basic instructions before starting the program and leaving the participant alone. After consenting to the study, participants completed a personality profile, ostensibly to be uploaded along with their digital photo to their networked partner. Profile items included age, gender, college major, and questions such as, “If you could have a super power, what would it be?” and “What would you rather be doing right now”? After “uploading” this information, the computer program explained that the participant had been “randomly assigned” to the condition in which his partner (in fact, a phantom female confederate) would choose whether or not to perform the teamwork tasks with him. Participants were shown the digital photo of the attractive confederate and a brief profile describing her as 19 years old, female, and majoring in psychology.
While participants waited for their partner’s response to their photo and personality profile, they completed the first negative affect measure and filler questionnaires. Control participants were then told that the computer malfunctioned and could not connect them with their partner. Participants in the rejection condition were told they were rejected by their partner and given the following feedback:
I heard about this study from my roommate. She said it was actually about dating, after the test she had to hang out with the guy and answer a bunch of questions about attraction. Looking at this photo, I’m really not attracted to this guy. He’s not my type at all and I don’t want to have to go out with him. I’d rather do the other study for the points.
Participants were then enrolled in a “second study” to complete their obligation; a second consent form described it as “investigating how relationships have evolved with technology.” Participants then completed the second negative affect measure, the GASP, and measures of body shame, sexual shame, and sexual aggression (in that order). The program randomly assigned participants to condition and administered the items for each measure in random order. After indicating their sexual orientation for exclusion purposes, they were thoroughly debriefed and compensated. 3
Results and Discussion
Preliminary Analyses
Table 2 presents the descriptive statistics and correlations among the focal variables as a function of rejection. Providing convergent validity, body shame was positively correlated with shame-withdrawal and sexual shame for both groups. For rejected men, body shame was positively related to Time 2 negative affect and also rape proclivity, replicating our observation from pilot research. Furthermore, for rejected men only, rape proclivity was positively related to Time 2 negative affect, shame-withdrawal, and sexual shame. Thus it will be important to distinguish the effects of body shame from those of shame-withdrawal and sexual shame.
Descriptive Statistics and Correlations by Rejection Condition (Experiment 1).
Note. Means with a, b subscripts significantly differ by condition. T2 = Time 2 (post-rejection).
n = 65 men.
n = 62 men.
p < .05. **p < .01. ***p < .001.
Also seen in Table 2, rejected men scored higher on Time 2 negative affect than control men, t(125) = 4.52, p < .001, d = .75. Not shown in Table 2, the difference between Time 1 and Time 2 negative affect was also greater for rejected men (M = 0.47, SD = 0.65) than control men (M = 0.06, SD = 0.30), t(125) = 4.56, p < .001, d = .76. Thus the manipulation had the intended effect. No significant differences emerged on Time 1 negative affect, body shame, sexual shame, shameful self-evaluation, or rape proclivity, all ts(125) < 1.35, ps > .17. Unexpectedly, control men scored higher on shame-withdrawal (M = 2.59, SD = 0.98) than rejected men (M = 2.22, SD = 0.88), t(125) = 2.21, p = .03, d = .39. That is, control men were more likely to respond to shameful situations by withdrawing than rejected men.
Body Shame, Response to Rejection, and Sexual Aggression
We expected men high on body shame to retaliate with sexual aggression to the extent that being rejected by an attractive woman upset them. To test this, we mean-centered predictor variables (Aiken & West, 1991) and hierarchically regressed rape proclivity on rejection condition (coded 1 = rejected, −1 = control), body shame, Time 2 negative affect (Step 1), and all interaction terms (Step 2). There was an unsurprising positive main effect of negative affect, β = 0.21, p = .03, and an unexpected negative main effect for condition, β = −0.20, p = .03, which further analyses suggested was due to a suppressor effect. 4 There was also a Rejection × Body Shame interaction, β = −0.73, p = .02. Simple effects showed that body shame and rape proclivity were positively related in the rejection condition, r(63) = .32, p = .01, but not in the control condition, r(60) = .04, p = .74 (see Table 2).
These effects were qualified by the predicted Rejection × Body Shame × Negative Affect interaction, β = 0.99, p = .01. Among control men, the Body Shame × Negative Affect interaction was weakly negative, β = −0.25, p = .17, and the main effects were negligible, both ps > .50. However, rejected men showed the expected Body Shame × Negative Affect interaction, β = 0.52, p < .001. Figure 1 shows the results for rejected men ±2 SD from the mean on body shame and negative affect (in black). Simple slopes tests (Aiken & West, 1991) confirmed that negative affect was positively related to rape proclivity for rejected men high on body shame, β = 1.05, t(61) = 6.43, p < .001. In contrast, negative affect was negatively related to rape proclivity for rejected men low on body shame, β = −0.67, t(61) = 2.76, p = .008. Thus only men high on body shame were more sexually aggressive to the extent that they were upset by the female confederate’s rejection.
Predicted rape proclivity scores for men ±2 SD from the M on body shame and post-rejection negative affect for men who were rejected for a dating study by an attractive woman (black squares) and for control men (gray circles)—Experiment 1.
Figure 1 also shows the predicted means for control men in gray. The control means are descriptively lower than those for the focal group (rejected men high on body shame and negative affect), with one exception: Control men high on negative affect but low on body shame scored similarly to the focal group. Nonetheless, body shame and negative affect did not significantly affect rape proclivity among control men as it did for rejected men. The pattern of results was quite different (reflecting a negative interaction of body shame and negative affect) but because this effect was weak, p = .17, we refrain from interpreting it.
The (nonorthogonal) simple slopes tests decomposing the interaction the other way showed a comparable pattern for rejected men. Body shame was positively related to rape proclivity for rejected men high on negative affect, β = 1.37, t(61) = 5.28, p < .001, but negatively related to rape proclivity for rejected men low on negative affect, β = −1.05, t(61) = 3.91, p < .001. In concert, our results suggest that high body shame and negative affect catalyze sexual aggression, but only for rejected men (i.e., men under masculinity threat).
Discriminant Validity for Body Shame
To examine body shame’s discriminant validity, we mean-centered all predictor variables. First, we hierarchically regressed rape proclivity on rejection condition, shame-withdrawal, Time 2 negative affect (Step 1), and all interaction terms (Step 2). Results showed only a weak effect for negative affect, β = 0.19, p = .06, with no remaining effects, all ps > .10. The three-way interaction term was weak, β = 0.27, p = .43. Second, we replaced shame-withdrawal with shameful self-evaluation. Results showed only a main effect for negative affect, β = 0.24, p = .01, with no remaining effects, all ps > .21. The three-way interaction term was weak, β = −0.26, p = .43. Thus generalized shame proneness, with or without negative affect, was not a factor in men’s rape proclivity.
Third, we regressed sexual shame in the same regression analysis and found a significant Rejection × Sexual Shame × Negative Affect interaction, β = 0.77, p = .04. Among control men, the Sexual Shame × Negative Affect interaction was nonsignificant, β = −0.20, p = .37, as were the main effects, both ps > .28. However, rejected men showed a main effect for negative affect, β = 0.44, p < .001, qualified by a significant Sexual Shame × Negative Affect interaction, β = 0.42, p < .001. Simple slopes tests revealed a positive relationship between negative affect and rape proclivity for rejected men high on sexual shame, β = 1.24, t(61) = 5.81, p < .001. In contrast, post-rejection negative affect was inversely related to rape proclivity for men low on sexual shame, β = −0.54, t(61) = 2.14, p = .03. Thus, as with body shame, men high on sexual shame were more sexually aggressive to the extent they were upset by the confederate’s rejection of them.
Would body shame contribute unique variance to rape proclivity after controlling for sexual shame? To find out, we hierarchically regressed rape proclivity on condition, body shame, sexual shame, Time 2 negative affect (Step 1), and all interaction terms (Step 2). Beyond the known main effect for negative affect, β = 0.20, p = .04, the only significant findings were the known effects involving body shame: the Rejection × Body Shame interaction remained significant, β = −0.73, p = .04, as did the Rejection × Body Shame × Negative Affect interaction, β = 1.04, p = .04. By comparison, the Rejection × Sexual Shame × Negative Affect interaction was weak, β = 0.23, p = .62, as were the remaining effects, all ps > .21. Thus after controlling for the effects of body shame, sexual shame was unrelated to rape proclivity. These findings bolster the unique significance of men’s body shame as a factor in sexual aggression.
In summary, Experiment 1 modeled a situation in which an attractive woman rejects a man for not being “her type” and he retaliates with sexual aggression. Our findings supported our expectation that female rejecters are at risk of retaliation from men high on body shame who feel threatened by the rejection. Experiment 1 ruled out generalized shame proneness as a competing moderator, and although sexual shame moderated exactly as body shame did, after controlling for body shame, it no longer did (whereas the effects for body shame remained significant). Thus Experiment 1 suggests that body shame may be an important precursor for male sexual aggression. We sought to conceptually replicate and extend these findings in Experiment 2.
Experiment 2
One limitation of Experiment 1 concerns the scope of our findings. Believing a woman rejected them after she viewed their photo may have heightened men’s appearance concerns (although we note that there was no main effect of rejection on body shame; see Table 2). Moreover, for obvious reasons, the confederate was always female, but she was also attractive. In Experiment 2, no photos were displayed, of either participants or confederates. We also varied the confederate’s gender and used sexual misclassification as the masculinity threat (e.g., Bosson et al., 2005; Thurlow, 2001). As a dependent measure, we employed a RBA, which obliges men to choose between violent and sexually violent pictures to repeatedly expose women to them. Men who implicitly dehumanized women as animals (Rudman & Mescher, 2012) or evaluated rape favorably (Widman & Olson, 2011) also scored high on the RBA. As in Experiment 1, we predicted that when a female confederate rejected men as a team partner for “being gay,” men high on body shame would respond with sexual aggression to the extent that they were upset by the rejection. To provide discriminant validity, we did not expect this pattern for men provoked by a male perpetrator (see also Anderson & Anderson, 2008; Malamuth, 1988).
Method
Participants
Men prescreened to be heterosexual (N = 214) completed the experiment in exchange for credit toward an Introductory Psychology research requirement. Of these, 104 (49%) were White, 67 were Asian (31%), 17 (8%) Latino, 12 (6%) Black, and 14 (6%) selected another option.
Materials
Participants completed the same negative affect and body shame measures from Experiment 1. Reliabilities were comparable (αs = .91 and .82 for Time 2 negative affect and body shame, respectively). We included the sexual shame measure from Experiment 1 as a further test of body shame’s discriminant validity (α = .73).
RBA
The RBA (Rudman & Mescher, 2012) obliged men over 17 trials to choose between two images, one depicting rape or sexual harassment, the other male-on-male aggression. Stimuli included both classical paintings and contemporary photos or advertisements.
5
The cover story for the RBA was as follows:
For this last part of the study we need you to help us select pictures for an upcoming study with women. In this study we will show women a subset of these pictures many times to test their perceptions. For this next task, we will show you two pictures and we would like you to pick the one you think we should use in the women’s study. Pick the one you think should be shown to women many times.
Responses were scored so that 0 = violent, 1 = sexually violent, and summed to form the RBA (α = .82; possible range: 0-17).
Procedure
Participants were recruited using Experiment 1’s cover story and instructions, which were identical with the exception that no photos were taken. After “uploading” the same personality profile used in Experiment 1, the computer program provided participants with a brief profile of their networked partner, who was described as either female or male, 19 years old, and a psychology major. After completing the first negative affect measure and filler questionnaires, control participants were informed of a technical failure, as in Experiment 1.
6
Rejected participants were provided with the following feedback:
I don’t think we have anything in common and won’t be a good team. It would be a waste of time to work on an experiment together if we can’t win the money. I’d rather work with someone else, or complete a different study, than work with this guy on teamwork tasks. Looking at his profile, I get the impression he is gay. We won’t work well together if he likes men.
As in Experiment 1, participants were then enrolled in a “second study” to complete their obligation. After providing consent, they completed the second negative affect measure and measures of body shame, sexual shame, and the RBA (in that order). Participants were then fully debriefed and compensated.
Results and Discussion
Preliminary Analyses
Table 3 shows the descriptive statistics and correlations among the focal variables as a function of rejection condition. Replicating Experiment 1, body shame was positively correlated with Time 2 negative affect, sexual shame, and RBA scores for men in the rejected condition. Among controls, body shame was only correlated with Time 2 negative affect.
Descriptive Statistics and Correlations by Condition (Experiment 2).
Note. Means with a, b subscripts significantly differ by condition. T2 = Time 2 (post-rejection).
n = 109 men.
n = 106 men.
p < .05. **p < .01. ***p < .001.
Table 3 also shows that rejected men scored higher on Time 2 negative affect than control men, t(212) = 5.62, p < .001, d = .71, as in Experiment 1. Moreover, the difference between Time 1 and Time 2 negative affect was again greater for rejected men (M = 0.60, SD = 0.95) than for control men (M = 0.12, SD = 0.43), t(212) = 4.82, p < .001, d = .63. Results of a 2 (confederate gender) × 2 (rejection condition) ANOVA confirmed main effects of condition for Time 2 negative affect and the negative affect difference score, both Fs(1, 210) > 4.27, ps < .001. Thus regardless of confederate gender, men rejected for “being gay” were more upset than controls. This same ANOVA for the remaining variables (Time 1 negative affect, body shame, sexual shame, and the RBA) produced no significant results, all Fs(1, 210) < 2.95, ps > .09.
Body Shame, Response to Rejection, and Sexual Aggression
To conceptually replicate Experiment 1, men high on body shame in the female confederate condition should be more willing to expose women to sexually offensive materials to the extent that they were upset by the rejection. Using mean-centered predictor variables, we first sought to establish that female (not male) rejection was necessary to provoke men into sexual aggression. To preserve statistical power, we conducted a focused test by hierarchically regressing the RBA on condition (coded 1 = rejected, −1 = control), confederate gender (coded −1 = male, 1 = female), body shame, Time 2 negative affect (Step 1), and the focal four-way interaction term (Step 2). Results showed a main effect for body shame, β = 0.20, p = .01, and also the expected four-way interaction, β = 0.17, p = .01. To specifically test our hypotheses, we then hierarchically regressed the RBA on condition, body shame, negative affect (Step 1), and all interaction terms (Step 2) separately for men in the male [female] confederate condition.
For men in the male confederate condition, results showed no significant effects (all ps > .08), and the Rejection × Body Shame × Negative Affect interaction was negligible, β = 0.16, p = .64. For men in the female confederate condition, there was a positive main effect for body shame, β = 0.21, p = .04, but on replicating Experiment 1, the predicted Rejection × Body Shame × Negative Affect interaction was also significant, β = 0.57, p = .01. For men rejected by a woman for “being gay,” there was a positive main effect for body shame, β = 0.32, p = .03, qualified by the expected Body Shame × Negative Affect interaction, β = 0.32, p = .04. Figure 2 shows the results for these men ±2 SD from the mean on body shame and negative affect in the female confederate condition (in black). Simple slopes tests were weaker, but in the same direction as for Experiment 1: Negative affect was positively related to the RBA for rejected men high on body shame, β = 0.44, t(52) = 1.72, p = .09, but negatively related to the RBA for rejected men low on body shame, β = −0.49, t(52) = 1.77, p = .08. The interaction confirms that the two slopes significantly differed, replicating Experiment 1.
Predicted RBA scores for men ±2 SD from the M on body shame and post-rejection negative affect who were rejected for being gay by a female confederate (black squares) and for control men in the female confederate condition (gray circles)—Experiment 2.
Figure 2 also shows the predicted means for control men in the female confederate condition (in gray), who showed only a main effect for Time 2 negative affect, β = 0.37, p = .01. As can be seen, all predicted means were descriptively lower than the focal group’s mean (rejected men high on body shame and negative affect). A check on the predicted means for men in the male confederate condition (who showed no significant effects) also revealed low RBA scores (range: 1.30-1.56).
Taken together, our findings confirm that high body shame and negative affect together promote sexual aggression, but only for men rejected by a woman (not a man) for “being gay.”
Discriminant Validity for Body Shame
For men rejected by a woman, it was possible that sexual shame would echo the effects of body shame, as it did in Experiment 1. We therefore hierarchically regressed the RBA on rejection condition, sexual shame, Time 2 negative affect (Step 1), and all interaction terms (Step 2) for men with a female confederate. Results were negligible for sexual shame, all ps > .10. Instead, there was the known main effect for negative affect, β = 0.23, p = .03. Thus the effect of negative affect on men rejected by a woman “for being gay” was specific to men high on body shame, not men high on sexual shame. For men with a male confederate, results of this same analysis were negligible, all ps > .21.
In summary, Experiment 2 conceptually replicated Experiment 1 by showing that men high on body shame who reacted negatively when a female confederate delivered a masculinity threat responded with more sexual aggression, compared with their counterparts low on negative affect. Thus men high on body shame do not need to be rejected for their appearance (as in Experiment 1) to retaliate against women with sexual aggression. By comparison, men with low body shame did not show this pattern; instead, they were somewhat less likely to retaliate against women if the rejection upset them (and significantly less likely in Experiment 1). Furthermore, being rejected by a male confederate had no effect on men’s RBA scores, which were generally low. Thus body shame and negative affect together catalyzed sexual aggression on the part of men rejected by a woman for “being gay.”
General Discussion
Two experiments uniquely demonstrated the role of male body shame in sexual aggression for men who responded negatively to masculinity threats. In Experiment 1, men rejected for a dating study by a desirable woman were willing to commit sexual assault to the extent that they were high on body shame and upset by the rejection. In Experiment 2, men high on body shame who were upset when rejected by a woman for “being gay” were likely to repeatedly expose future women to pictures of sexual assault. As in past research, men retaliated against women when their masculinity was threatened (Dall’Ara & Maass, 2000; Maass et al., 2003; Richardson et al., 1984), but unique to the present research, this effect was specific to men high on body shame and post-rejection negative affect. These findings support the role of body shame in male sexual aggression and further suggest that this linkage can be exacerbated by men’s reactions to masculinity threats.
Our findings may seem counterintuitive, given that prominent rape theories have posited that men high on masculinity or self-esteem pose the greatest risk to women (Baumeister et al., 2002). Moreover, men who are large or strong tend to also be aggressive (DeWall, Bushman, Giancola, & Webster, 2010; Sell, Tooby, & Cosmides, 2009). For this reason, Bosson and Vandello (2011) speculated that muscular or “big” men might be especially likely to aggressively restore masculinity when under threat. This may well be the case, but our findings suggest that body-ashamed men may be particularly vulnerable to masculinity threats from women, which promotes sexual aggression, and thus deserve more attention from theorists.
Both experiments relied on men reacting with negative affect following a masculinity threat, but why do some men react negatively to rejection while others do not? If body shame causes men to feel “out of touch” with their bodies, they may be compromised in their ability to regulate negative affect. In the present research, men low on body shame were less likely to sexually aggress to the extent that they were upset by female rejection. This suggests that men with a healthy orientation to their bodies may be more sensitive to physiological cues that promote self-regulation, compared with men high on body shame (Werner et al., 2009).
Prior research has targeted hostility toward women as a factor in male-on-female aggression (Anderson & Anderson, 2008), but what makes men hostile toward women? We have uniquely identified male body shame, which is conceptually unrelated to some previously identified factors (e.g., trait aggression and acceptance of interpersonal violence), while negatively related to others (e.g., high self-esteem and dominance; Baumeister et al., 2002). But why would being ashamed of their body predispose men to sexual aggression? Our findings are consistent with our reasoning that such men may be under chronic masculinity threat, making them more sensitive to acute instances. They are also in line with Malamuth et al.’s (1995) speculation that insecure men may engage in sexual aggression because they resent women’s sexual power. Future research is necessary to untangle these two plausible explanations.
Limitations and Future Directions
Because laboratory settings are limited in their ability to investigate sexual aggression, measures like the RBA may be among the best on offer (Rudman & Mescher, 2012). Nonetheless, there is a sobering gap between actual rape and our attempt to model the consequences of social situations that threaten men’s masculinity and place women at risk. Whether body shame is a catalyst for sexual assault among actual perpetrators is a question for future research.
We focused on men’s body shame because pilot research revealed unexpected linkages between it and men’s willingness to report sexual aggression (e.g., rape proclivity, likelihood to sexually harass, and street harassment). However, in our experiments, body shame was only associated with sexual aggression among rejected men, not controls (see Tables 2 and 3). These results are consistent with our argument that body-ashamed men react aggressively to masculinity threats because this threat is relatively chronic for them, but they raise the question of why control men did not also show this correlation (as in the pilot work). Perhaps not being rejected (because the computer failed) served to affirm men’s masculinity by warding off the threat of rejection, or perhaps they imagined they would have been accepted had the equipment worked.
Although two experiments confirmed the role that body image plays in sexual aggression for men whose masculinity is threatened by a woman, it is left to future research to determine why this occurs. Our efforts to mediate the link between body shame and rape proclivity using the variables described in preliminary research (e.g., those shown in Table 1 but also the PAQ masculinity subscale, selected items drawn from the femininity subscale, state self-esteem, and precarious manhood beliefs) were unsuccessful. We have suggested interoceptive sensitivity (Werner et al., 2009) as one possible mediator. Future research might also include acute measures of masculinity threat, female rejection sensitivity, and resentment of women’s sexual power as possible candidates (Malamuth et al., 1995). To our knowledge, these measures would have to be designed and validated in advance for this purpose.
The age of our participants was appropriate because young men are especially likely to sexually aggress against women (Barbaree, Hudson, & Seto, 1993; Freeman, 2007). Nonetheless, older men should be investigated to examine the generality of our findings. Young men may be more critical about their bodies, even if in better physical shape than older men. It may be the case that body image declines in importance as men mature. Whether difficulty in obtaining physical ideals affects older men’s body shame and sensitivity to masculinity threats is an empirical question.
Finally, it is important to address whether sexual aggression serves an adaptive or maladaptive function for men under masculinity threat, whether they are high on body shame or not. Future research should examine this question.
Conclusion
Our hypothesis that men high on body shame who react negatively to masculinity threats would respond with sexual aggression was supported by the present findings. The findings are consistent with men’s admission that female rejection often precipitates sexual assault (McKibben et al., 1994), surveys suggesting that men aggress against relationship partners to the extent that their body image is poor (Shelton & Liljequist, 2002), and research showing that men under masculinity threats aggress against female confederates (e.g., Dall’Ara & Maass, 2000; Maass et al., 2003; Richardson et al., 1984). We extended these findings to include sexual aggression that may target women as a group, but only for men high on body shame who react negatively to female rejection. Thus men’s body shame may be a key component in male-on-female sexual violence. Considerable research has shed light on the issues that trouble women’s experiences with their bodies (Moradi & Huang, 2008). In light of the present findings, men’s body image should be investigated with the same tenacity, and not merely in contrast to women’s, but as a challenge in its own right.
Footnotes
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This research was partially supported by Grant BCS-1122522 from National Science Foundation to the second author.
Notes
References
Supplementary Material
Please find the following supplemental material available below.
For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.
For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.
