Abstract
The present study examined whether men view gender-atypical (i.e., feminine) psychological disorders as threats to their gender status. Men and women (N = 355) rated their expectations of gender status loss, feelings of distress, and help-seeking intentions in response to 10 different stereotypically masculine and feminine psychological disorders. Men as compared to women expected greater gender status loss for, and reported more distress to, gender-atypical versus gender-typical disorders. Expectations of gender status loss partially mediated the link between participant gender and distress at the thought of gender-atypical disorders. These findings suggest that feminine disorders pose more powerful gender status threats for men than masculine disorders do and that men’s expectations of gender status loss for feminine disorders drive their negative reactions to these mental illnesses. The discussion emphasizes the importance of considering the gender-typicality of disorders, and the implications of these findings for clinical interventions.
When attempting to explain men’s, relative to women’s, low rates of help-seeking for mental illness, theorists note that male gender role norms of toughness, emotional stoicism, and self-reliance are incompatible with admissions of mental illness and expressions of emotional distress (e.g., Addis & Mahalik, 2003; Brownhill, Wilhelm, Barclay, & Schmied, 2005). When men do seek help, they less often than women receive diagnoses for internalizing disorders (characterized by mood disturbance, sadness, and anxiety), but they more often than women receive diagnoses of externalizing disorders (characterized by low impulse control; Kessler, Chiu, Demler, & Walters, 2005; Needham & Hill, 2010; Rosenfield, Vertefuille, & Mcalpine, 2000; Seedat et al., 2009). These gender differences in help-seeking and symptom presentation may ultimately reflect the pressures that men experience to behave consistently within the male gender role.
The current work examines how men’s reactions to certain types of mental illness diagnoses reflect their concerns about the implications these diagnoses have for their gender status. Proposed here is that men perceive gender-atypical (i.e., stereotypically feminine) mental illness diagnoses as greater threats to their gender status than gender-typical (i.e., stereotypically masculine) diagnoses. The present study examines whether men’s tendency to view gender-atypical diagnoses as gender threats partially accounts for their negative reactions to these disorders. Women should expect relatively little loss of gender status for gender-atypical (i.e., stereotypically masculine) versus gender-typical (i.e., stereotypically feminine) disorders.
These predictions derive from precarious manhood theory, which describes manhood as an elusive and tenuous social status harder to earn and easier to lose than womanhood (e.g., Bosson & Vandello, 2011; Vandello & Bosson, 2013; Vandello, Bosson, Cohen, Burnaford, & Weaver, 2008). Whereas both manhood and womanhood can be conceptualized as social statuses whose legitimacy depends to some degree on social recognition from others, “real man” status requires the passage of social achievements and proofs (e.g., landing a high-status job, exhibiting bravery) not required for “real woman” status. Unlike manhood, womanhood is viewed as resulting primarily from the passage of physical developmental milestones (e.g., reaching puberty; Vandello et al.). Because manhood requires social achievements, a failure to demonstrate behaviors socially prescribed for men can threaten men’s gender status in others’ eyes. Consequently, men (but not women) respond to gender threats—such as negative feedback about their masculinity, or gender role violations—with public demonstrations of active, agentic behaviors and avoidance of stereotypically feminine behaviors (e.g., Bosson & Michniewicz, 2013; Vandello & Bosson).
When men and boys fail to uphold masculine gender role norms, they usually suffer harsher interpersonal penalties than do women and girls who violate feminine gender role norms. For example, parents, peers, and perceivers generally offer more punishment and more negative evaluations of male than female gender role violators (e.g., Feinman, 1981, 1984; G. D. Levy, Taylor, & Gelman, 1995; McCreary, 1994; Sirin, McCreary, & Mahalik, 2004). Not surprisingly, male gender role violators anticipate negative evaluations from others for exhibiting feminine behaviors or otherwise falling short of masculine behavioral ideals (e.g., Bosson, Prewitt-Freilino, & Taylor, 2005; Michniewicz, Vandello, & Bosson, 2014).
The present investigation extends precarious manhood theory to the domain of mental illness and ask whether men anticipate a loss of gender status in others’ eyes for gender-atypical mental illness diagnoses. According to this theory, any seemingly feminine emotion, behavior, or interest can challenge a man’s gender status. Thus, if gender-atypical disorders communicate such gender-atypical qualities, men should expect these disorders to raise questions about their manhood. Given their relatively assured gender status, women should have relatively less concern than men that gender-atypical disorders will threaten their womanhood.
Men may avoid admission of both stereotypically masculine and feminine mental illness, because the implied vulnerability can threaten manhood status. In support of this logic, work on masculinity and mental health suggests that many aspects of mental health diagnosis and treatment, such as admitting emotional distress, seeking help, and the emotional openness required for successful psychotherapy, contradict male gender role norms of emotional restrictiveness, toughness, and self-reliance (Addis & Mahalik, 2003; Brownhill et al., 2005; Mahalik, Good, & Englar-Carlson, 2003; O’Neil, 2008). Thus, men who adhere more closely to traditional male gender role norms also report more negative attitudes about seeking help from mental health professionals (Berger, Levant, McMillan, Kelleher, & Sellers, 2005; Levant & Richmond, 2007; Mahalik, Burns, & Syzdek, 2007). Further, endorsement of traditional masculine ideologies predicts men’s tendency to self-stigmatize, or evaluate themselves negatively for having psychological problems that require professional assistance (Levant et al., 2013; Vogel, Wade, & Haake, 2006).
Some classes of psychological disorders, such as externalizing disorders, coincide with stereotypically masculine qualities and male role norms such as impulsiveness, action, fighting, and violence. Internalizing disorders coincide with stereotypically feminine qualities and role norms, including sadness, anxiety, passivity, and withdrawal (Kessler et al., 2005; Kramer, Krueger, & Hicks, 2008; Leadbeater, Kuperminc, Blatt, & Hertzog, 1999; F. Levy, Hay, Bennett, & McStephen, 2005). Men’s social pressure to eschew all femininity (Bosson & Michniewicz, 2013; Pleck, 1981, 1995; Thompson, Grisanti, & Pleck, 1985) may lead them to view stereotypically feminine mental health diagnoses as especially aversive. Consistent with this logic, men’s but not women’s self-stigmatization for an eating disorder (a stereotypically feminine disorder) predicts more negative attitudes toward seeking psychological help for the disorder (Hackler, Vogel, & Wade, 2010). These findings coincide with the thesis that men view psychological disorders in general, and perhaps feminine disorders in particular, as challenges to their gender status.
From this perspective, past work on men’s reactions to mental illness has two limitations. No quantitative work directly examines men’s expectations of gender status loss for gender-atypical versus gender-typical disorders. Whereas the existing literature contributes much to the understanding of men’s discomfort surrounding mental illness diagnoses and their negative attitudes toward help-seeking, no quantitative studies have explicitly documented men’s belief that some types of mental illness diagnoses will threaten their status as a “real man” in others’ eyes. Qualitative interview studies report that men view depression as a feminine mental disorder and accordingly expect to be seen as feminine if they express depressive thoughts or emotions to others (Danielsson, Bengs, Samuelsson, & Johansson, 2011; O’Brien, Hart, & Hunt, 2007). Although such interviews provide insight into people’s unique experiences, they lack the systematicity and control of quantitative experiments. A systematic investigation of men’s expectations of gender status loss for stereotypically feminine versus masculine disorders represents an important addition to the literature, in part because men’s concerns about manhood loss can have implications for their functioning. To illustrate, in a nationally representative sample of involuntarily unemployed U.S. adults, endorsement of the belief that “others saw me as less of a man [woman] when I lost my job” was a significant predictor of depression, anxiety, and low self-esteem among unemployed men but not among unemployed women (K. S. Michniewicz, J. A. Vandello, & J. K. Bosson, unpublished data).
Second, studies on men and mental health often focus primarily on only one particular disorder (e.g., depression, eating disorders, posttraumatic stress disorder; Jakupcak, Osborne, Michael, Cook, & McFall, 2006; Magovcevic & Addis, 2008; Petrie, Greenleaf, Reel, & Carter, 2008), or they broadly measure men’s attitudes regarding “psychological disorders” as an entire category (Berger et al., 2005; Hammer, Vogel, & Heimerdinger-Edwards, 2013; McKelley & Rochlen, 2010; Vogel, Wade, & Hackler, 2007). In contrast to these approaches, the present study systematically varied the stereotypical masculinity versus femininity of a set of 10 specific disorders to allow direct comparisons between people’s reactions to these classes of disorders. These masculine and feminine disorders were carefully piloted to ensure people viewed them as comparably gendered and familiar. This allows us to address two key questions that currently remain unanswered: (1) Do men expect more gender status loss for gender-atypical than gender-typical disorders? (2) Do expectations of gender status loss account for men’s relatively more negative reactions to gender-atypical versus gender-typical disorders?
These questions are addressed here by directly examining men’s and women’s expectations of gender status loss, distress, and willingness to seek help for gender-typical and gender-atypical disorders. If men are motivated to avoid the appearance of femininity, then men should experience more negative reactions (distress, aversion to help-seeking) to psychological disorders to the extent that they anticipate gender status loss for these disorders. The focus on distress in response to disorders reflects an effort to capture a broadly defined, affectively aversive reaction. In past studies, men reacted to gender-threatening experiences with more negative affect (e.g., self-conscious discomfort, anxiety) than women. For example, following (false) negative feedback about their performance on a measure of masculine gender identity, men completed more anxiety-related words in a word-completion task, and expressed greater discomfort with the thought of others learning about their test feedback, than women did (Caswell, Bosson, Vandello, & Sellers, 2014; Vandello et al., 2008). These findings indicate that threats to their gender status elicit stressful thoughts and feelings for men, as well as a desire to keep one’s gender role violations hidden (see also Rudman & Fairchild, 2004). This work implies that men should report greater distress at the thought of gender-atypical as compared to gender-typical mental illness diagnoses, and should express less willingness to seek assistance from others in dealing with gender-atypical versus gender-typical diagnoses. These reactions (distress and resistance to help-seeking) should follow from men’s anticipated gender status loss for gender-atypical diagnoses. Women should not react especially negatively to gender-atypical disorders, as women’s gender status is relatively more stable than men’s, and women are not penalized as much as men for gender role violations.
Overview and Hypotheses
Given the precarious nature of men’s gender status, and the interpersonal punishments that many men anticipate for violating male gender role norms, the present investigation explores the possibility that receiving a gender-atypical diagnosis implies losing one’s manhood. This logic was tested here by asking men and women to imagine having diagnoses of five stereotypically masculine and five stereotypically feminine mental illnesses and then indicate how much gender status they would lose in other people’s eyes, how bothered they would feel, and how much help they would seek. Prior to imagining the disorders, approximately half of participants underwent a gender threat manipulation, whereas the remaining participants did not experience this threat. This threat may heighten men’s expectations of gender status loss for gender-atypical disorders, by reminding them of the precariousness of their gender status.
Data were collected from college students and nonstudents to achieve a more broadly representative sample. Given work suggesting demographic and background differences in people’s attitudes toward psychological disorders and help-seeking (e.g., Eisenberg, Golberstein, & Gollust, 2007; Golberstein, Eisenberg, & Gollust, 2008; Mackenzie, Gekoski, & Knox, 2006; Wang et al., 2005), the current study assessed sample (students vs. nonstudents), familiarity with both specific mental illnesses and mental illness more broadly, and race/ethnicity for use as potential covariates. Also measured was participants’ estimates of the likelihood of suffering from each disorder for use as potential covariates.
Men, relative to women, should theoretically report more gender status loss and, accordingly, stronger feelings of distress and lower help-seeking intentions when considering diagnoses of gender-atypical disorders relative to gender-typical ones. The current work also explores whether a gender status threat would exaggerate this difference. Theoretically, gender status loss drives men’s negative reactions to gender-atypical disorders. Here, this possibility was tested using a process model in which participant gender predicts distress and resistance to help-seeking, for gender-atypical disorders, through gender status loss (Hayes, 2013).
Method
Piloting the Disorders
To select a set of gendered disorders, 72 undergraduates received extra for rating 19 psychological disorders along dimensions of familiarity and gender stereotypicality. Ratings of familiarity with each disorder and its main symptoms were made on 4-point scales where 1 = not at all familiar, 2 = somewhat familiar, 3 = moderately familiar, and 4 = very familiar. Participants rated stereotypicality on scales of 1 (men much more likely to have this) to 5 (women much more likely to have this), with a midpoint of 3 (men and women equally likely to have this). The final set of 10 disorders were familiar (received average familiarity ratings above a “2”) and clearly associated with one gender or the other (received average stereotypicality ratings that differed significantly from the scale midpoint of “3”). Masculine disorders included antisocial personality disorder, narcissistic personality disorder, drug addiction, alcoholism, and attention-deficit hyperactivity disorder; Mfamiliar = 3.05, t(71) = 13.53, p < .001; Mstereotypical = 2.52, t(71) = −10.86, p < .001. Feminine disorders included anorexia nervosa, bulimia nervosa, major depressive disorder, generalized anxiety disorder, and binge eating disorder; Mfamiliar = 3.14, t(71) = 13.89, p < .001; Mstereotypical = 3.98, t(71) = 17.36, p < .001.
Participants and Design
Participants were 200 college students (140 women and 60 men) who completed an online survey in exchange for credit toward a course requirement and 155 nonstudents (79 women and 76 men) recruited from Amazon’s Mechanical Turk (MTurk; www.mturk.com), a website where subscribers can complete posted tasks for monetary compensation. MTurk workers received $0.25 to complete all measures and were required to reside in the United States and be at least 18 years of age. The data of five people who did not complete the gender threat task and seven people who failed an attention check were deleted, leaving a final sample of 343 respondents (212 women, 131 men). Respondents ranged in age from 18 to 63 (Mdn = 22 years) and described themselves as White (64.1%), Black (12.5%), Latino/Latina (9.6%), Asian (7.0%), Arabic (2.6%), and Native American (0.6%); the remaining 3.5% declined to indicate their race/ethnicity.
Participants were randomly assigned to threat condition in a 2 (Participant gender: men, women) × 2 (Gender threat: threat, no threat) × 2 (Gender typicality of disorders: typical, atypical) design with repeated measures on the last factor. Note that “gender-typical” disorders were operationalized as masculine disorders for men and as feminine disorders for women; “gender-atypical” disorders were operationalized as feminine disorders for men and as masculine disorders for women.
Procedure
College student participants were recruited through the Psychology Department’s web-based participant pool. Nonstudents were recruited through MTurk. All interested respondents followed a link to the study at the Surveygizmo website (www.Surveygizmo.com). Following informed consent, participants completed the threat manipulation and then imagined themselves having each of the 10 mental illnesses. Participants then completed measures of gender status loss, distress, help-seeking, attention checks, and demographics.
Threat manipulation
Following Weaver, Vandello, and Bosson (2013), participants were randomly assigned to recall and list either 10 (threat condition) or 2 (no threat condition) behaviors that they performed recently that met the cultural definition of what a “real man” or a “real woman” should be like. This manipulation is based on the tendency for people to draw inferences about their own standing on a given dimension from the ease or difficulty with which they can recall behaviors relevant to this dimension (Schwarz, 1998). When recall is easy (i.e., 2 behaviors are requested) people should infer that they are high in “real man” [“real woman”] status, and when recall is difficult (i.e., 10 behaviors are requested), people should infer that they are low in “real man” [“real woman”] status. After listing their behaviors, participants responded to two items on scales of 1 (not at all) to 9 (very much): “How difficult was it to do the behavior listing task?” and “How closely do you think that you match your culture’s standards for the ideal man [woman]?”
Gender status loss (GSL)
Participants were asked to “imagine that you were diagnosed with each of the disorders below. How would this shape how other people evaluated you? For each disorder, please rate the extent to which other people would see you as ‘less of a man [woman]’ if they learned that you were diagnosed with that disorder.” The 10 disorders then appeared in a random order, and participants rated each one on scales of 1 (people definitely would not see me as “less of a man [woman]”) to 7 (people definitely would see me as “less of a man [woman]”). These ratings were averaged to create separate masculine (α = .84) and feminine (α = .89) indices.
Bothered
Participants indicated how bothered they would be to get diagnosed with each of the 10 disorders. The word “bothered” was used rather than “distressed” to reflect that an aim was to assess negative emotion using gender-neutral terminology. In past studies (see Vandello & Bosson, 2013, for a review), male participants were reluctant to endorse terms such as “distressed” or “anxious,” presumably because admitting to such emotions violates male role norms. Ratings were made on scales of 1 (not at all bothered) to 7 (very bothered). These were averaged to create masculine (α = .75) and feminine (α = .82) composites.
Help-seeking (HS)
Written instructions asked participants to “imagine what it would be like if you thought that you might be suffering from the disorders below. Would you be likely to seek help?” They then used scales of 1 (definitely would not seek help) to 7 (definitely would seek help) to rate the likelihood that they would seek help first from a mental health professional, and then from a close friend or relative, if they thought they might be suffering from each of the 10 disorders. Help-seeking ratings were internally consistent for both masculine (α = .88) and feminine (α = .90) disorders, so they were averaged to create composites.
Likelihood
Participants indicated the likelihood that they would suffer from each of the 10 disorders in their lifetime. Ratings were made on scales of 1 (not at all likely) to 7 (very likely), and they were averaged to create masculine (α = .83) and feminine (α = .79) composites.
Familiarity
Participants rated their familiarity with each of the 10 disorders on scales of 1 (not at all familiar) to 4 (very familiar), and these were averaged to create masculine (α = .79) and feminine (α = .86) composites.
Attention check and demographics
Embedded randomly within the scales above was a single item that read “Please select ‘Moderately Disagree’ for this item.” This served as an attention check. Seven responses were deleted from analyses for failing this check. After completing the measures described above, as well as several other scales not relevant to the current study, participants provide some basic demographic information, including their gender and race/ethnicity. Participants were asked whether (a) they personally, (b) any of their family members, and (c) any of their friends had ever been diagnosed with a mental disorder. All of these questions were answered using a yes/no format, and the sum across all three responses (no = 0, yes = 1) constituted a rough index of “general familiarity with mental illness” that ranged from 0 to 3.
Results
Correlation Analyses and Covariate Selection
Tables 1 and 2 contain correlations among all of the variables (the potential covariates, two independent variables, and six dependent variables) and descriptive statistics. Note that sample was uncorrelated with familiarity ratings of the specific sets of gender-typical and gender-atypical disorders examined here. This suggests that although the 10 disorders were originally piloted in a college student sample, noncollege participants found them equally familiar.
Correlations Among All Study Variables.
Note: Race/ethnicity was coded 0 = non-White, 1 = White; sample was coded 0 = student, 1 = nonstudent; gender was coded 0 = women, 1 = men. GT = gender-typical; GA = gender-atypical.
p < .05, **p < .01.
Descriptive Statistics for All Study Variables.
Note: Race/ethnicity was coded 0 = non-White, 1 = White; Sample was coded 0 = student, 1 = nonstudent. GT = Gender-Typical; GA = Gender-Atypical.
Recommendations by Porter and Raudenbush (1987) determined which variables to include as covariates in primary analyses. Specifically, covariates should (a) correlate relatively strongly with one or more of the dependent measures; (b) not correlate strongly with one another; (c) not differ strongly as a function of the independent variables; and (d) not interact with the independent variables. Note that none of the potential covariates met the criterion of correlating strongly with the dependent variables. That said, sample and likelihood of suffering from gender-typical disorders each correlated with several of the dependent measures (rs from –.12 to –.30, ps < .04) and did not correlate strongly with either of the independent variables or with each another. To test whether these two potential covariates interacted with the independent variables, analyses consisted of a series of multiple regression analyses using a Bonferroni adjustment to control for the large number of tests conducted. Setting α = .004 (.05/12 = .004), neither of the potential covariates interacted with either threat condition or participant gender in predicting any of the six dependent variables, all ps > .01. Sample and likelihood of gender-typical disorders were therefore included as covariates in all analyses. Note, however, that all results remain virtually identical (i.e., no conclusions change) when excluding these covariates.
Manipulation Checks
Participants’ responses to the two manipulation check items were submitted to separate 2 (Participant gender: men, women) × 2 (Gender threat: threat, no threat) between-participants analyses of covariance (ANCOVAs), controlling for the two covariates. The ANCOVA on the difficulty item produced a main effect of threat condition, F(1, 335) = 20.62, p < .001, ηp2 = .058, such that threatened respondents (M = 5.55, SE = 0.20) rated the behavior listing task as more difficult than nonthreatened respondents (M = 4.27, SE = 0.20). No other effects approached significance; all Fs < 1.54, ps > .21. Unexpectedly, however, the ANCOVA on the item assessing how well participants matched the cultural ideal for their gender produced no effects, Fs < 1. People in the threat condition did not decrease respondents’ perceptions of themselves as “ideal” men and women, making it unlikely that the gender threat manipulation had its intended effect.
Primary Analyses
Gender status loss
The two GSL composites were submitted to a 2 (Participant gender: men, women) × 2 (Gender threat: threat, no threat) × 2 (Gender typicality: typical, atypical) ANCOVA with repeated measures on the last factor. A gender-by-typicality interaction, F(1, 335) = 14.05, p < .001, ηp2 = .040, qualified main effects of gender, F(1, 335) = 6.68, p < .02, ηp2 = .020, and typicality, F(1, 335) = 13.74, p < .001, ηp2 = .039. Threat condition did produce any main or interactive effects, and no other effects were significant, all Fs < 1. Figure 1 depicts the means associated with the gender-by-typicality interaction. Both women and men anticipated more gender status loss for gender-atypical than gender-typical disorders (Fs = 56.63 and 114.99, ps < .001, ηp2s = .145 and .256, respectively), although disorder typicality played a much bigger role in men’s gender status loss concerns than women’s. Whereas men and women expected similar gender status loss for gender-typical disorders, F < 1, men anticipated more gender status loss than women for gender-atypical disorders, F(1, 335) = 15.73, p < .001, ηp2 = .045.
Gender status loss as a function of participant gender and gender typicality of disorders. Response scales ranged from 1 to 7.
Distress
The two distress composites were submitted to the same ANCOVA model described above. Once again, a gender-by-typicality interaction, F(1, 335) = 35.06, p < .001, ηp2 = .095, qualified a main effect of gender, F(1, 335) = 7.01, p < .01, ηp2 = .020. No other effects emerged, Fs < 2.10, ps > .14. As depicted in Figure 2, women expressed more distress by the thought of gender-typical than gender-atypical disorders, F(1, 335) = 5.78, p < .02, ηp2 = .017, whereas men expressed greater distress instead at the thought of gender-atypical versus gender-typical disorders, F(1, 335) = 32.92, p < .001, ηp2 = .089. Further, women and men were equally bothered by gender-atypical disorders, F < 1, whereas women felt more bothered than men by gender-typical disorders, F(1, 335) = 23.27, p < .001, ηp2 = .065.
Distress in response to diagnoses as a function of participant gender and gender typicality of disorders. Response scales ranged from 1 to 7.
Help-seeking
The two help-seeking composites were submitted to the same ANCOVA model as above. Again, an interaction of gender-by-typicality, F(1, 335) = 27.22, p < .001, ηp2 = .075, qualified a main effect of gender, F(1, 335) = 6.16, p < .02, ηp2 = .018. No other effects emerged, all Fs < 1.31, ps > .25. Contrary to predictions, women reported stronger help-seeking intentions for gender-typical than gender-atypical disorders, F(1, 335) = 16.46, p < .001, ηp2 = .047, whereas men reported stronger help-seeking intentions for gender-atypical than gender-typical disorders, F(1, 335) = 12.50, p < .01, ηp2 = .036 (see Figure 3). Women and men reported similar help-seeking intentions for gender-atypical disorders, F < 1, whereas women reported greater help-seeking intentions than men for gender-typical disorders, F(1, 335) = 16.14, p < .001, ηp2 = .046.
Help-seeking for disorders as a function of participant gender and gender typicality of disorders. Response scales ranged from 1 to 7.
Process model. The indirect effects of gender on the dependent measures (distress and help-seeking) through anticipated gender status loss for gender-atypical disorders was expected. Because the threat manipulation did not yield any main or interactive effects on any dependent measures, analyses collapsed across threat condition when testing for mediation. Given that the pattern for help-seeking was opposite to predictions (men reported stronger help-seeking intentions for gender-atypical than gender-typical disorders), no further consideration was given to this dependent variable.
Testing for indirect effects involved using Hayes’s (2013) PROCESS application to estimate standard errors and confidence intervals from 10,000 bootstrap samples using random sampling with replacement. Participant gender (0 = women, 1 = men) was the independent variable, gender status loss for gender-atypical disorders was the mediator, and distress in response to gender-atypical disorders was the outcome. Sample and likelihood of gender-typical disorders served as covariates. The results of this analysis, displayed in Table 3, indicate significant mediation: Men, relative to women, anticipated more gender status loss for gender-atypical disorders, and gender status loss predicted distress in response to gender-atypical disorders. Although the direct effect of gender on distress was nonsignificant, the indirect effect of gender on distress through gender status loss was significant, as the bootstrapped 95% confidence interval for the standardized indirect effect excluded zero. This indirect effect was further corroborated by a significant Sobel (1986) test, z = 2.22, p < .03. Both sample and likelihood of gender-typical disorders were significant predictors in the final step of the model, ts > 2.06, ps < .04. Nonstudents, and people who viewed gender-typical disorders as more likely, reported significantly less distress at the thought of gender-atypical disorders.
Test Statistics for Process model.
Note: GSL = gender status loss.
Because of expected gender differences in reactions to gender-atypical disorders, the indirect effect reported above should not emerge when considering reactions to gender-typical disorders. Indeed, it did not. As demonstrated in Table 3, men and women did not differ in anticipated gender status loss for gender-typical disorders and, although gender status loss for gender-typical disorders predicted distress in response to such disorders, the bootstrapped 95% confidence interval for the standardized indirect effect contained zero. Similarly, the Sobel (1986) test did not reach significance. Thus, men’s heightened concerns about loss of gender status only drive their distress at the thought of gender-atypical disorders.
Discussion
This study examined whether men view stereotypically feminine as compared to stereotypically masculine psychological disorders as more threatening to their gender status. To do this, the present study systematically varied the gender typicality of familiar psychological disorders to allow comparisons between feminine and masculine disorders, and manipulated gender threat to explore the effect of men’s heightened concerns about their gender status. Data from both college students and a nonstudent sample provided greater generalizability of the current findings.
Several interesting findings emerged. Both men and women expected more gender status loss for gender-atypical than for gender-typical disorders, but this pattern was more pronounced among men. Although women expressed more distress than men when considering the entire set of disorders, only men expressed more distress for gender-atypical than for gender-typical diagnoses. These findings suggest not only that feminine disorders pose more powerful gender status threats for men than masculine ones do, but that men more than women feel threatened by gender-atypical disorders. These findings provide the first direct evidence that men associate stereotypically feminine disorders explicitly with a loss of their manhood status, a highly valued social identity.
These results also provide evidence that anticipated gender status loss drives men’s negative reactions to gender-atypical disorders. Consistent with the logic of precarious manhood theory (e.g., Vandello et al., 2008), men’s distress at the thought of gender-atypical diagnoses derives in part from their expectation that others will view them as less of a “real man” for having these disorders. Although the same pattern of results did not emerge for men’s help-seeking intentions, this nonetheless suggests that gender status concerns play a role in men’s affective reactions to feminine disorders. Moving forward, then, efforts to portray disorders as less stereotypically associated with either gender, or less detrimental to one’s manhood status, may encourage men to seek treatment and cope effectively with a variety of mental health diagnoses.
These findings contribute to the general understanding of men’s mental health in several important ways. The fact that men associate gender-atypical disorders with a greater loss of manhood than gender-typical disorders suggests that researchers should consider the gender-typicality of mental illnesses when examining men’s attitudes toward disorders and mental health. Men may indeed react negatively to mental illness in general (Addis & Mahalik, 2003), they react especially negatively to feminine, internalizing disorders such as mood disorders (e.g., depression, anxiety) and eating disorders (e.g., bulimia, binge eating disorder). Importantly, even when past researchers manipulated the gender-typicality of specific disorders (e.g., Wirth & Bodenhausen, 2009), they examined only a small set of disorders. The present study assessed an array of pilot-tested masculine and feminine disorders so as to allow for generalizations across broad classes of gendered disorders.
These findings can potentially inform work on the links between male gender role norms and help-seeking attitudes, intentions, and behaviors (Addis & Mahalik, 2003). One promising mediator of these links is self-stigmatization, or the tendency to internalize public stigma surrounding mental illness (Hammer et al., 2013; Pederson & Vogel, 2007; Vogel et al., 2006; Vogel et al., 2007; Vogel, Heimerdinger-Edwards et al., 2011). Men who endorse traditional masculinity ideologies tend to hold more negative attitudes toward help-seeking, in part, because they stigmatize themselves for having psychological problems that require assistance (Levant et al., 2013). The current findings indicate that that men’s illness-related concerns about being seen as unmanly are linked to the gender-typed characteristics of disorders, meaning that feminine disorders should be generally more stigmatizing than masculine ones. Thus, future research should examine how gender-typicality of disorders moderates the links between masculine ideologies, self-stigmatization, and mental health attitudes.
Another line of research finds that perceivers stigmatize men and women with gender-typical disorders more than those with gender-atypical disorders (Wirth & Bodenhausen, 2009). Male and female targets incurred greater negative affect (anger, disgust, dislike) and less sympathetic affect (sympathy, concern, pity) from perceivers when described as having a gender-typical rather than a gender-atypical disorder. To explain this pattern, Wirth and Bodenhausen proposed that gender-atypical disorders are seen as more “genuine” kinds of mental disturbances than gender-typical ones, and are therefore thought to be relatively uncontrollable. That is, a man’s depression is assumed more uncontrollable than is a man’s alcoholism, whereas the reverse is true for a woman’s disorders. These sets of findings together speak to the potential differences between public and private stigma associated with men’s mental illness. Although men may actually incur greater public stigma for gender-typical than gender-atypical disorders, they may privately stigmatize themselves more for gender-atypical than gender-typical disorders because of the gender role violations inherent in feminine disorders. Because private stigma is more strongly related to help-seeking attitudes than public stigma (Vogel et al., 2007), men with gender-atypical disorders might have more negative attitudes toward help-seeking, regardless of the public stigma they incur from others.
The contrast between the current findings and those reported by Wirth and Bodenhausen (2009) may have implications for the accuracy of men’s expected gender status loss. Although men expect to lose more gender status for having gender-atypical than gender-typical disorders, they receive less stigmatization for atypical than typical disorders. This pattern coincides with research on men’s expectations of gender status loss following unemployment (Michniewicz et al., 2014). Following a hypothetical or real job loss, men (relative to women) expected more loss of gender status than observers offered a hypothetical male target who lost his job. Thus, converging findings suggest that men tend to overestimate the blow to their manhood that will result from gender role violations.
This work has important clinical implications. For example, clinicians may benefit from knowing that men might resist diagnoses with stereotypically feminine disorders, as the disorders themselves constitute a loss of manhood for many men. Men’s concerns about gender status loss for feminine disorders may also be overblown relative to how others actually view them. Clinical interventions might therefore focus on bringing men’s beliefs in greater alignment with the relatively compassionate attitudes that perceivers hold of men with gender-atypical disorders.
This study contains several limitations worth mentioning. The gender threat (behavior recall task) had no effect on participants’ self-views, suggesting that it failed to raise participants’ gender status concerns. Past studies have typically used more direct, unambiguous means of manipulating gender status threats, such as offering men (bogus) negative feedback on an ostensible test of gender identity (e.g., Caswell et al., 2014; Vandello et al., 2008) or having them perform feminine activities (e.g., styling hair; Bosson et al., 2005). Thus, future research may benefit from using more potent manipulations of gender status threat to examine whether or not such threats exaggerate men’s distancing from gender-atypical disorders. Nonetheless, the present study’s robust finding that men reacted more negatively to gender-atypical than gender-typical mental illnesses is noteworthy. That is, regardless of the presence versus absence of a manipulated gender threat, men perceived feminine disorders, as a class, more threatening to their manhood than masculine ones.
Another limitation of this work is the lack of predicted effects on help-seeking. No support for the hypothesis that expectations of gender status loss would drive men’s resistance to seeking help for gender-atypical diagnoses. Instead, men reported lower help-seeking intentions than women overall and greater intentions to seek help for gender-atypical disorders than for gender-typical disorders. Women conversely reported greater intentions to seek help for gender-typical than gender-atypical disorders. These patterns are consistent with research indicating robust gender differences in help-seeking tendencies in general (Biddle, Gunnell, Sharp, & Donovan, 2004; Blazina & Marks, 2001; Blazina & Watkins, 1996; Deane, Wilson, & Ciarrochi, 2001; Mackenzie et al., 2006), but they contradict the aforementioned logic about the gender-threatening consequences of gender-atypical disorders.
The hypothetical nature of the help-seeking task, which asked participants to imagine their behavioral responses to different disorders, may have affected people’s responses. Participants may have answered questions about their behavioral intentions by relying on their logic about how one ought to act. Because gender-atypical disorders elicited heightened feelings of distress among men, perhaps men relied on these emotional reactions to make estimates of their likely behavior (e.g., more distress equals more help-seeking). Men likely view stereotypically feminine disorders, such as mood and eating disorders, as especially unlikely to befall them. Thus, men may view diagnoses of gender-atypical disorders as highly unusual and therefore as especially likely to require treatment. However, this study addressed this possibility by measuring participants’ perceptions of the likelihood of having both gender-typical and gender-atypical disorders, and by including likelihood of gender-typical disorders as a covariate in all analyses. Rerunning all analyses, the covarying likelihood of gender-atypical disorders produced nearly identical effects to those reported here. This suggests that the reported help-seeking effects cannot be explained entirely by men’s beliefs about the likelihood of experiencing psychological disorders. Nonetheless, participants in this nonclinical sample may have had difficulty gauging their behavioral intentions to seek help for disorders with which they have little firsthand experience. Thus, although the validity of the affective measure of distress seems adequate, because these presumably reflect people’s immediate, gut-level responses to the disorders, the behavioral intention measures may have reflected people’s theories about how they ought to behave. A conclusive test of the effects of gender-typicality of disorders on men’s help-seeking behaviors therefore awaits further testing.
The current findings indicate that men view feminine disorders as more threatening to their gender status than masculine disorders, whereas women make less of a distinction between masculine and feminine disorders. Men’s expectations of gender status loss for gender-atypical disorders predict their feelings of distress at the thought of having these disorders. This work fills a gap in the mental health literature by systematically varying the gender-typicality of disorders and directly examining anticipated gender status loss. By relying on the precarious manhood framework, the present study illuminates why men react more negatively to feminine than masculine disorders, as well as why women appear relatively nonplussed by gender-atypical disorders: Men, relative to women, experience chronic concerns about the loss of their gender status. Whether these concerns are valid or exaggerated, they have real consequences for men’s cognitive and emotional reactions to mental health.
Footnotes
Acknowledgements
We thank Joe Vandello for his helpful comments on an earlier draft of this manuscript.
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) received no financial support for the research, authorship, and/or publication of this article.