Exploring the Relationship Between Gender Nonconformity and Mental Health among Black South African Gay and Bisexual Men

Studies in Western countries have consistently demonstrated that, as a consequence of more frequent discrimination, gender nonconforming gay and bisexual men experience more mental distress than gender conforming gay and bisexual men (D’Augelli, Grossman, & Starks, 2006; Grossman, D’Augelli, Salter, & Hubbard, 2005; Henning-Stout, James, & Macintosh, 2000; Landolt, Bartholomew, Saffrey, Oram, & Perlman, 2004; Ploderl & Fartacek, 2009; Sandfort, Melendez, & Diaz, 2007; Skidmore, Linsenmeier, & Bailey, 2006). These relationships have never been studied in low- and middle-income countries, even though gender nonconformity (GNC) is evident in expressions of same-sex sexuality in such countries. We explored whether GNC among gay Black South African men was associated with depression and if this association was mediated or moderated by discrimination.

Further, we explored the potential protective effect of being “out” and gay community involvement on the relationship between GNC and depression. Being out and gay community involvement may suppress the potentially negative effects of GNC on mental health. Gender nonconforming men are more likely to be open concerning their sexuality than gender conforming men (Toomey, Ryan, Diaz, Card, & Russell, 2010), while concealment may induce more physical and psychological distress (Cole, Kemeny, Taylor, & Visscher, 1996). In addition, gay community involvement may protect gay men from some of the mental health distress associated with being gender nonconforming (Friedman, Koeske, Silvestre, Korr, & Sites, 2006).

GNC is pervasive in the expression of same-sex sexuality in South Africa. The expression of same-sex sexuality appears to reproduce a binary notion of masculinity and femininity among gay men much more so than in developed countries (Murray & Roscoe, 2001). Some Black gay men display feminine behaviors and participate in traditionally feminine occupations while others display more masculine identities. For instance, within some Black same-sex relationships, there are “wives” and “husbands” with clearly defined gender boundaries (Reid, 2005).

The analysis reported here was based on a subsample of 353 Black gay-identified or bisexual men from a total sample of 2010 self-identified lesbian, gay, bisexual, and transgender persons who participated in a community-based survey conducted in three provinces in South Africa. Participants were recruited through LGBT organizations, support groups and counseling centers, friendship networks, at the annual Gay and Lesbian Pride March, and via the Internet. For additional details pertaining to the sampling procedures, see Sandfort, Nel, Rich, Reddy, and Yi (2008).

Table 1 shows the descriptive statistics for the sample. Men ranged in age from 16 to 49 years. Gender conforming men had a higher mean age (t = −3.04, p <.01). There were no significant differences in education, employment or sexual activity between gender conforming and nonconforming men. The means and SDs for the main study variables are shown in Table 2.

Table 1.

Demographic characteristics of gender nonconforming and conforming Black gay and bisexual men

	Nonconforming Men	Conforming Men
Mean age (in years) (SD)*	(n = 141)	(n = 197)
	23.6 (5.5)	25.6 (5.6)
Education (%)	(n = 147)	(n = 193)
Less than grade 12	21.1	17.6
Grade 12	27.2	31.1
Certificate	11.6	12.4
Diploma	20.4	18.7
Degree/Post graduate degree	19.8	19.8
Employment (%)	(n = 148)	(n = 200)
Employed	26.4	34.5
Unemployed	18.9	24.5
Student/Scholar	47.3	32.5
Self employed/Other	7.4	8.5
Sexual Activitya (%)	(n = 150)	(n = 191)
Single (not sexually active)	21.3	18.3
Single (sexually active)	34.0	32.5
Committed relationship (casual sex)	26.7	28.3
Committed relationship (no casual sex)	18.0	20.4

Note. Chi-square statistics were used to examine differences between nonconforming and conforming men.

^aOther category for sexual activity was less that .5% of the sample for both nonconforming and conforming men.

^*p < .05

Table 2.

Means and SDs for depression, school discrimination, general discrimination, gay community involvement, and outness

	Nonconforming M (SD)	Conforming M (SD)
Depression	1.75 (.71)	1.69 (.68)
School discrimination	2.30 (.88)	1.85 (.81)
General discrimination	1.19 (.34)	1.15 (.38)
Gay community involvement	2.59 (.80)	2.30 (.77)
Outness	2.85 (.97)	2.54 (.98)

Note. Absolute ranges: Depression, 1–5; School discrimination, 1–3; General discrimination, 1–4; Gay community involvement, 1–5; Outness, 1–3

Gender non-conformity was assessed utilizing a single item: “How do you present yourself?” Participants either selected “Feminine way,” “No preference,” or “Masculine way.” About 43% of the men reported to present themselves in a feminine way, 13% had no preference, and 45% presented themselves in a masculine way. This variable was recoded to include individuals who selected “no preference” into the “masculine” category. Thus, the gender nonconforming variable had two categories: 1 = feminine and 2 = masculine. Depression was assessed using a 5-item Likert scale (α = .83). Example items were “I have trouble getting to sleep or staying awake” and “I have suicidal thoughts.” School discrimination was assessed with a 3-item Likert scale (α = .73). The survey question asked, “When you were at school, did you experience any of the following things because of your sexual orientation?” Example items include “Verbal abuse and harassment” and “Negative jokes about LGB individuals.” General discrimination was assessed on a 4-item Likert scale (α = .77). The question asked, “Have you personally experienced any of the following because of your sexual orientation in the last 24 months?” Example items were “Physical abuse” and “Experience property attacks.” Gay community involvement was assessed using a 5-item Likert scale (α = .79). Example items included “How frequently do you socialize at the homes of other LGBT friends” and “How frequently do you socialize at LGBT restaurants?” Level of outness was assessed using a 3-item Likert scale (α = .83). Participants were asked about how many of their friends, family, and coworkers they were out to.

Because all continuous variables were normally distributed, we conducted parametric statistical tests. Initial analyses examined the relationship among study variables. Contrary to our expectation, there was no significant relationship between GNC and depression. Therefore, we did not conduct a mediation analysis, as initially planned, because not all assumptions for mediation were met (Baron & Kenny, 1986). We used linear regression analyses to test if (1) discrimination moderated the relationship between GNC and depression, (2) gay community involvement moderated the relationship between GNC and depression, and (3) level of outness moderated the relationship between GNC and depression.

Table 3 shows the bivariate correlations for the independent and dependent variables. Both school discrimination and general discrimination were associated with depression (r = .27, p < .01; r = .17, p < .01, respectively). In addition, school discrimination and general discrimination were correlated with GNC (r = −.25, p < .01; r = −.12, p < .01, respectively). The association between GNC and depression was not significant.

Table 3.

Correlation matrix for gender nonconformity, depression, discrimination, outness, and gay community involvement

Variable	1	2	3	4	5
1. Gender nonconformity	-
2. Depression	−.03	-
3. School discrimination	−.25**	.26**	-
4. General discrimination	−.12*	.16**	.37**	-
5. Gay community involvement	−.17**	.06	.30**	.29**	-
6. Outness	−.15**	.07	.22**	.24**	.47**

Note. Spearman rank correlation coefficient reported for correlations between gender nonconformity, depression, school discrimination, general discrimination, gay community involvement, and level of outness. Pearson correlation coefficient reported to specify the relationship between all other study variables.

^**p < .01;

^*p < .05

Table 4 shows the results for the moderation analyses. Regression Model 1 and regression Model 2 tested the moderating effect of school and general discrimination, respectively. The interaction terms in both analyses were not significant, indicating that school discrimination and general discrimination did not moderate the relationship between GNC and depression. Regression Model 3 shows that the overall model was significant (F = 6.21, p < .001); the interaction term was, however, not significant, indicating that more involvement in the gay community did not protect gender nonconforming men from experiencing increased levels of depression. Model 4 shows that the overall model was significant (F = 6.13, p < .001); however, the interaction term was not significant, indicating that outness was not a resiliency factor that suppressed the effect of GNC on depression.

Table 4.

Multiple regression analyses for variables predicting depression

	Variables	B	Depression
			SE	β
Model 1	Gender nonconformity (GNC)	.34	.19	.24t
	School discrimination (SD)	.42	.13	.53**
	GNC X SD	−.13	.08	−.29
Model 2	Gender nonconformity (GNC)	.28	.25	.20
	General discrimination (GD)	.76	.35	.40*
	GNC X GD	−.27	.20	−.33
Model 3	Gender nonconformity (GNC)	.05	.23	.03
	Gay community involvement (CI)	−.02	.15	−.02
	GNC X CI	−.00	.09	−.01
	School discrimination	.21	.04	.26***
	General discrimination	.14	.10	.07
Model 4	Gender nonconformity (GNC)	.06	.21	.04
	Outness (OT)	.01	.12	.02
	GNC X OT	−.00	.07	−.02
	School discrimination	.20	.04	.25***
	General discrimination	.13	.10	.07

^***p < .001;

^**p < .01;

^*p < .05;

^tp < .10

Contrary to our expectations and to what is usually reported in studies conducted in Western countries, we found that Black gender nonconforming gay and bisexual men did not experience more depression compared to Black gender conforming gay and bisexual men. This is surprising, because we also found that gender nonconforming men experienced more discrimination than their gender conforming counterparts and experiences of discrimination were positively associated with feelings of depression. Based on the latter two findings, it would have seemed logical to expect that gender nonconforming men would have higher levels of depression.

The absence of a relationship between GNC and depression suggests either that the relationship between GNC and mental health as found in Western studies among gay and bisexual men does not exist among Black South African gay and bisexual men or that the relationship was masked by a third variable. It could, for instance, be that, in the South African context, GNC is a much less salient characteristic; this would, however, contradict the increased level of discrimination of gender nonconforming gay and bisexual men found in this study. It could also be that, for some reason, gender nonconforming men are more resilient and consequently less affected by the increased levels of discrimination.

We tested the potential role of two resilience factors: outness and gay community involvement. These factors did, however, not buffer the relationship between GNC and depression. It could be that there are other resiliency factors, which were not assessed in this study. Gender nonconforming men may, for instance, have higher levels of self-esteem compared to the gender conforming men and these feelings of self-esteem might counteract the effects of increased discrimination. Alternatively, gender nonconforming men who experienced more discrimination might, as a consequence, have developed more successful coping mechanisms, attenuating the effect of discrimination on mental health. Such coping practices may have been developed at an early age. Gender nonconforming men may have begun to build and be protected by social networks of friends and family. Rabie and Lesch (2009), for instance, discussed how it was normal for feminine gay Black men in a South African township to visit with the females in the community and participate in traditionally female activities.

There are more complex alternative explanations that take into account the specific situation in South Africa. In the dominant perception among Black persons in South Africa, homosexuality is strongly associated with femininity (Reid, 2005). It is almost as if in the system of hegemonic masculinity, identification with femininity is the major way in which a gay identity can be expressed. Group identification would consequently be stronger among gender nonconforming men than among gender conforming men. The discrimination that gender nonconforming men undergo could, therefore, be experienced as a group-based rejection, resulting in a nurturing and embracing of the excluded identity which subsequently bolsters well-being (Branscombe, Schmitt, & Harvey, 1999; Jetten, Branscombe, Schmitt, & Spears, 2001).

In our reasoning thus far, we assumed that the reported experiences with discrimination only differed between gender nonconforming men and gender conforming men in terms of their frequency. That does not have to be the case. A more elaborate assessment of discrimination, taking into account different kind of actors, more diverse situations, as well as specific time periods in which discrimination occurred, would have helped to explore the extent to which discrimination experiences of gender nonconforming men indeed differ from those of gender conforming men. It is also crucial to understand how discrimination is perceived and to what it is attributed. A limitation of this study relates to how GNC was assessed. Men were asked whether they presented themselves as feminine, masculine, or “no preference.” These preferences do not have to correspond completely with how these men behave and express themselves or with how they are perceived by others. Future studies should make use of scales to assess GNC. Because such scales have only been validated with populations in Western societies, it would be necessary to explore whether they are valid among populations in low- and middle-income countries; GNC might mean different things across cultures (Mair, 2010). Finally, the use of a convenience sample makes it impossible to draw conclusions about the generalizability of the findings.

In summary, the relationship between GNC and mental health in gay and bisexual men might be less universal than what currently is thought. Our findings suggest that it is important to take into account cultural circumstances and prevailing ideas about gender, masculinity, and femininity. Future studies should include more resiliency factors. In addition, studies should examine what GNC means in gay and bisexual men cross culturally.