Abstract
Objectives. We investigated whether high gender role conflict (GRC; internal conflict with traditional gender-role stereotypes and an individual’s perceived need to comply with these roles) is associated with psychological distress and HIV-related risk behaviors in a sample of African American men who have sex with men and women (MSMW).
Methods. We analyzed baseline data collected from questionnaires completed by 400 MSMW participating in the Men of African American Legacy Empowering Self project in Los Angeles, California, in 2007 to 2010 for associations between participants’ GRC and experiences of poor mental health and HIV risk outcomes.
Results. MSMW who reported higher levels of GRC than other participants also reported more psychological distress, lower self-esteem, greater internalized homophobia, less HIV knowledge, lower risk reduction skills, less disclosure of same-sex behaviors to others, and more unprotected vaginal or anal intercourse with female partners.
Conclusions. Future research should consider how high GRC affects African American MSMW’s lives and identify specific approaches to help alleviate the psychological distress and other negative behavioral outcomes associated with internal conflict caused by rigid gender role socialization.
Despite the heavy toll the HIV/AIDS epidemic has exerted on the lives of African American men who have sex with men (MSM), conspicuously few behavioral interventions have been developed to reduce HIV transmission in this population.1,2 Behavioral HIV prevention interventions designed specifically for men of any race/ethnicity who have sex with both men and women (MSMW) are practically nonexistent. In response to a 2005 report of HIV prevalence as high as 46% among African American MSM,3 the Centers for Disease Control and Prevention and other funding agencies began making resources available to adapt existing HIV prevention interventions, developed originally for gay White men, and to design and evaluate theory-driven HIV interventions to address the specific prevention needs of gay-, bisexual-, and heterosexual-identified African American MSM.
The Men of African American Legacy Empowering Self (MAALES) project, conducted in Los Angeles, California, is an example of a collaborative academic–community effort to develop and rigorously test a culturally appropriate behavioral HIV intervention for African American MSMW.4 The critical thinking and cultural affirmation model, one of the behavioral theories underlying the MAALES small-group intervention, posits that a healthy integration of racial/ethnic pride, gender role, and sexual identity will allow African American MSMW to operate from a culturally rooted foundation to make healthier choices in their lives and relationships with sexual partners. This model recognizes that concerns regarding the preservation of African American masculinity underlie much of the conflict and discourse regarding male homosexuality and bisexuality in African American communities. A unique component of the MAALES intervention—aimed at reducing HIV acquisition and transmission risk behaviors—is its reliance on a culturally congruent framework that bolsters racial and cultural pride while reducing HIV stigma and gender role conflict (GRC).4
GRC refers to negative consequences that can result from individuals' socialization regarding the roles ascribed to their gender and the need to comply with these roles. According to O’Neil, such consequences for men can stem from “negative critiques of self or others when conforming to, deviating from, or violating stereotypic gender role norms of masculinity ideology.”5(p363) Previous research on how males are socialized into traditional gender roles and the negative impacts of high GRC on mental health and other issues was conducted in samples primarily comprising heterosexual men.6–9 With the exception of a recent investigation of the impact of high GRC on condom use in a sample of African American MSM,10 most investigations of the impact of GRC on the lives of MSM have enrolled samples largely consisting of White gay-identified men.11–13
In recent work to develop more effective HIV prevention programs for African American MSMW, Operario et al. conducted qualitative interviews with heterosexual-identified, behaviorally bisexual African American men to better understand the context of MSMW’s sexual behaviors with both male and female partners and the need for secrecy and confidentiality regarding same-sex behaviors.14 The study findings revealed a high level of internal conflict among participants stemming from their desire for male partners in the context of community norms that valued rigid male gender roles, namely, an explicit expectation that “real” men will only partner with women. The investigators also found that participants emphasized the need to conceal their same-sex behaviors from female partners and members of their broader social network because of their perceptions that the African American community viewed homosexuality as a weakness. Previous quantitative research in samples of heterosexual men also found higher levels of GRC to be associated with greater psychological distress.15,16 Similarly, among predominantly gay-identified White MSM, high GRC has been associated with higher levels of depression and anxiety and lower self-esteem.12,17
We expanded on previous work by quantitatively investigating whether high GRC is associated with psychological distress as well as HIV-related risk behaviors in a sample of behaviorally bisexual African American men. Our theoretical rationale was that GRC may be related to African American men’s engagement in HIV risks through 2 mechanisms: indirectly, by influencing poor mental health outcomes such as psychological distress and lower self-esteem, which may reduce protective, safer-sex practices, and directly, because men with greater GRC may view vulnerability (either physical or emotional) as contrary to society’s expectation of masculinity. In other words, men may try to avoid the appearance of physical vulnerability by electing not to use condoms, an act that may be perceived as showing fear about health threats such as HIV. Men with greater GRC may also avoid the appearance of emotional vulnerability demonstrated by their use of condoms to protect their sexual partners’ health. Similarly, men struggling with GRC may see a need to have multiple female partners or to have more insertive anal intercourse with men to reinforce their masculinity.
Consistent with observations in other populations, we hypothesized that high GRC contributes to increased psychological distress and reduced self-esteem in African American MSMW. In addition, we hypothesized that high GRC contributes to increased homophobia by heightening concerns over the implications of intercourse with other men for masculine status. These factors, in turn, may influence HIV risk by reducing African American MSMW’s motivation to acquire HIV knowledge and skills and to practice HIV preventive behaviors, thus increasing the likelihood that they will engage in risky behaviors such as drug use prior to intercourse with other men. We used baseline data collected from MAALES project participants to examine whether African American MSMW with higher levels of GRC than other participants also reported (1) higher levels of psychological distress and poorer self-concept, (2) greater secrecy regarding MSM behaviors, and (3) higher frequencies of engagement in HIV risk behaviors.
We analyzed questionnaire data collected between August 2007 and October 2010 for the first 400 enrollees in the MAALES project. Eligibility criteria were self-identification as African American or Black, age 18 years or older, sexual intercourse with both men and women in the past 24 months, and unprotected vaginal or anal intercourse in the past 3 months. MAALES participants were recruited throughout Los Angeles County by the following routes: referred by a friend or participant (41%); saw a posted flyer (17%); recruited by a study outreach worker (12%); referred by a local agency (10%); saw or heard an Internet, newspaper, or radio advertisement (10%); and saw a bus or bus stop advertisement (10%). Once men were recruited, they were screened for eligibility either in field settings or over the telephone. If eligible to participate, individuals met with study staff to complete the institutional review board–approved informed consent process at the main study office, at Charles Drew University (67%); 1 of 3 collaborating community agencies (23%); or the participant’s home or other location (10%). Participants then completed baseline questionnaires with audio computer-assisted self-interview technology and received $20 compensation for their time.
Our questionnaire collected quantitative data through a variety of established, modified, and newly created measures. For this analysis, we used measures of sociodemographic and sexual orientation characteristics; alcohol and drug use behaviors prior to sexual intercourse in the past 90 days; sexual behaviors with male, female, and transgender partners in the past 90 days; self-efficacy, skills, and prevalence of condom use; disclosure of same-sex behaviors and HIV status; HIV/AIDS knowledge; psychological distress in the past 7 days; self-esteem; internalized homophobia; and GRC. We derived modifications of some scales from our formative research with community advisers to address issues particular to African American MSMW. We then cognitively tested the modified scales in a sample of African American MSMW (n = 20).
We administered a modified version of the 37-item Gender Role Conflict Scale (GRCS) originally developed by O’Neil et al.18,19 We selected the GRCS because it is a validated scale that has been administered in a variety of populations, including homosexual and African American men. The GRCS measures participants’ view of gender role stereotypes across 4 main dimensions: (1) success, power, and competition (e.g., “I worry about failing and how it affects my status as a man”); (2) difficulty expressing emotions or having others express them (e.g., “I have difficulty expressing my emotional needs to my partner”); (3) difficulty showing or observing affection between men (e.g., “I am sometimes hesitant to show my affection to men because of how others might perceive me”); and (4) conflict between work and family relations (e.g., “My needs to work or study keep me from my family or leisure more than I would like”).
Internal consistency reliability of the 4 multi-item scales ranged from 0.75 to 0.85 in a sample of 527 mostly White, male undergraduate college students.19 A recent investigation that used the GRCS in a sample of mostly White, gay, and bisexual males12 obtained α coefficients ranging from 0.87 to 0.91 for the first 3 dimensions of the scale. Concern over the construct validity of the measure for conflict between work and family relations20,21 led Szymanski et al. to omit this dimension from their analyses of GRC.12 Similarly, we omitted the fourth dimension of the scale because of formative work that questioned its validity for a sample of low-income, largely unemployed African American MSM.
For the MAALES intervention, we used formative research to develop 5 new items to measure a dimension we termed masculine presentation (e.g., “Men should never show their feminine side” and “I never want to look or seem weak”). We obtained α coefficients between 0.86 and 0.87 for O’Neil’s first 3 subscales and 0.73 for our newly developed subscale. (Table A, which presents correlations found in our baseline sample for the 4 dimensions in the modified GRCS, is available as a supplement to the online version of this article at http://www.ajph.org). Internal consistency was 0.94 for our modified, 35-item GRCS. Respondents rated each item according to a 6-point Likert scale (1 = strongly disagree to 6 = strongly agree) for a total score 35 to 210. Higher total and average scores indicated greater GRC.
Participants were asked to select 1 of the multiple-choice responses to the question, “As you know, this intervention is for men who have had sex with both men and women. We realize that men in this group have many different ways of identifying their sexuality. What do you consider your sexual orientation or identity to be?” Another section of the questionnaire focused on whether the participants had ever disclosed their same-sex relationships to a range of individuals in their lives and whether they felt it was important to keep their same-sex relationships a secret. In particular, we were interested in whether participants disclosed their same-sex relationships to any female sexual partners, health care providers, or family members.
We assessed psychological distress with the Brief Symptom Inventory 5322 and focused on symptoms of anxiety (6 items), depression (6 items), and somatization (7 items). These 3 scales had internal consistency reliability estimates ranging from 0.86 to 0.88 in our sample. Cronbach’s α for the inventory’s 53-item global severity index was 0.98 in our sample. Respondents rated the frequency of experiencing each symptom according to a 5-point categorical rating scale (0 = not at all to 4 = extremely). We summed the responses and calculated mean scores, which we used as a continuous outcome variable. Higher mean scores corresponded to greater psychological distress.
We also administered the Rosenberg Self-Esteem Scale (10 items; Cronbach’s α = 0.84).23 Respondents rated each item on a 4-point Likert scale (1 = strongly disagree to 4 = strongly agree) with higher mean scores indicating greater self-esteem. We summed the responses and calculated mean scores for use as a continuous outcome variable.
We assessed discomfort with homosexuality or bisexuality with a modified version of the Internalized Homophobia Scale.24 Our measure comprised 10 items and produced a Cronbach’s α of 0.88 for our sample. Respondents rated each item on a 5-point Likert scale (1 = strongly disagree to 5 = strongly agree). We summed responses to each item for total scores for a continuous outcome variable. Higher scores corresponded to higher levels of internalized homophobia.
We assessed participants’ knowledge of risks of HIV transmission with a 16-item scale of true–false statements. We selected 8 items from a 12-item HIV/AIDS knowledge scale previously used with heterosexual respondents.25–27 We created 8 additional items derived from formative research (e.g., “If someone has HIV, they will have a sunken face and a large belly” and “You have to have sex with someone more than once to get HIV/AIDS”). We conducted a factor analysis to identify 10 items composing 1 factor reflective of HIV knowledge in this sample (Cronbach’s α = 0.79). We summed the correct responses to the items for a total score ranging from 0 to 10.
We evaluated participants’ condom use self-efficacy with an 11-item measure that used a 5-point Likert scale to assess confidence in using condoms in a variety of situations (e.g., “I can talk to every new partner about the importance of using condoms”). Cronbach’s α in our sample was 0.92, and summed item responses ranged from 11 to 55.
We assessed condom use skills with a 7-item measure that rated participants’ confidence in applying the proper techniques when using a condom (e.g., “How confident are you that you could squeeze the tip of the condom while rolling it down?”). Respondents rated their condom use skills on a 3-point scale (1 = not at all confident/sure to 3 = very confident/sure). We obtained a Cronbach’s α of 0.93 and obtained total individual scores between 7 and 21.
We assessed participants’ engagement in oral, anal, and vaginal sexual intercourse with male, female, and transgender partners within the past 90 days. These sexual behavior questions also assessed the number of partners of each gender, the number of partners with whom the respondent had vaginal or anal sexual intercourse without using a condom, and whether drugs were used prior to or during intercourse. We assessed whether participants disclosed their same-sex behaviors to female partners, health care providers, and family members. Finally, we examined the frequency with which known HIV-positive participants disclosed their HIV status to sexual partners of all genders within the past 90 days.
We categorized men as having high GRC when their total GRCS score exceeded the median cutpoint (GRCS > 125 for our sample). We calculated prevalence odds ratios and 95% confidence intervals to assess whether selected sociodemographic characteristics were associated with higher odds of high GRC in our sample. We also used analysis of variance to assess whether increasing quartiles of GRCS scores similarly demonstrated associations with these sociodemographic characteristics.
We evaluated the associations between higher GRCS scores and psychological measures, HIV knowledge, and condom use self-efficacy and skills with the Spearman rank correlation for continuous outcomes. We used the Mann–Whitney rank sum test to evaluate associations between higher GRCS scores and dichotomous HIV risk characteristics. Finally, we investigated whether secrecy about MSM behaviors, which may suggest adherence to rigid male gender roles, was related to self-reported sexual orientation and GRC among our participants.
Sociodemographic characteristics of participants are shown in Table 1. Mean age at baseline was 42.5 years (range = 18–83 years). Twenty-five percent of the men had attained postsecondary education, and 77% were currently unemployed or unable to work. More than half of the sample earned less than $1000 per month, and 38% reported that they did not have a regular place to live within the past 12 months. History of incarceration was high (77%). Most of our sample reported that they were bisexual (61%), followed by heterosexual (14%), gay or homosexual (12%), or down low (7%). A high proportion of the sample reported being HIV positive (49%). Eight percent had never been tested for HIV.
TABLE 1— Sociodemographic Characteristics of 400 Men Who Have Sex With Men and Women: Men of African American Legacy Empowering Self Project, Los Angeles, CA, 2007–2010
| Characteristic | No. (%) |
| Age, y | |
| 18–29 | 60 (15) |
| 30–39 | 70 (18) |
| 40–49 | 177 (44) |
| ≥ 50 | 93 (23) |
| Education | |
| < high school | 66 (16) |
| High school diploma/GED | 231 (58) |
| Associate degree | 74 (18) |
| ≥ college degree | 29 (7) |
| Employment | |
| Full time | 21 (5) |
| Part time/occasional | 55 (14) |
| Unemployed | 181 (45) |
| Retired | 14 (3) |
| Unable to work | 128 (32) |
| Individual annual income, $ | |
| <12 000 | 224 (56) |
| 12 000–23 999 | 90 (22) |
| 24 000–47 999 | 47 (12) |
| ≥ 48 000 | 35 (9) |
| Living situation in past 12 mo | |
| Did not always have regular place to live | 154 (38) |
| Always had regular place to live | 246 (62) |
| Currently living with a sexual partner | |
| No | 342 (86) |
| Yes, male | 18 (5) |
| Yes, female | 36 (9) |
| Yes, transgender | 4 (1) |
| Ever incarcerated | |
| Yes | 306 (77) |
| No | 93 (23) |
| Sexual orientation | |
| Heterosexual/straight | 56 (14) |
| Bisexual | 243 (61) |
| Gay | 36 (9) |
| Same-gender loving | 6 (2) |
| Down low | 29 (7) |
| Homosexual | 11 (3) |
| Other/none of the above | 19 (5) |
| HIV status | |
| Positive | 195 (49) |
| Negative | 162 (41) |
| Inconclusive/indeterminate/refused | 13 (3) |
| Never tested | 30 (8) |
Note. GED = general equivalency diploma. Percentages may not total 100% because of rounding.
MAALES participants scored an average of 124.7 total points on the modified, 35-item GRCS (range = 41–210). On a scale of 1 to 6, participants’ mean subscale scores at baseline were (1) success, power, and competition, 3.79 (SD = 0.89); (2) restrictive emotionality, 3.39 (SD = 1.09); (3) restrictive affectionate behavior between men, 3.27 (SD = 1.05); and (4) masculine presentation, 3.74 (SD = 1.05).
Table 2 displays the prevalence and odds of high GRCS scores (median split) by participants’ sociodemographic characteristics. Higher odds of GRC were associated with younger age; lower educational attainment; higher annual income; unstable living situation; identification as bisexual or heterosexual, straight, down low, or other (rather than gay–homosexual); and being HIV negative or having unknown status. In our analysis of variance, we observed the same associations between higher quartiles of GRCS scores and these sociodemographic characteristics, except for age and educational attainment.
TABLE 2— Prevalence and Prevalence Odds Ratios of High Gender Role Conflict Scores by Sociodemographic Characteristics Among 400 Men Who Have Sex With Men and Women: Men of African American Legacy Empowering Self Project, Los Angeles, CA, 2007–2010
| Gender Role Conflict Scorea | |||
| Characteristic | High, % | Low, % | PORb (95% CI) |
| Age, y | |||
| 18–29 | 68 | 32 | 2.3 (1.3, 4.3) |
| 30–39 | 43 | 57 | 0.8 (0.5, 1.4) |
| 40–49 | 48 | 52 | 1.0 (Ref) |
| ≥ 50 | 48 | 52 | 1.0 (0.6, 1.7) |
| Education | |||
| < high school | 56 | 44 | 2.1 (1.1, 4.1) |
| High school diploma/GED | 53 | 47 | 1.9 (1.1, 3.2) |
| Associate degree | 38 | 62 | 1.0 (Ref) |
| ≥ college degree | 45 | 55 | 1.3 (0.6, 3.2) |
| Employment | |||
| Full time | 43 | 57 | 1.0 (Ref) |
| Part time/occasional | 53 | 47 | 1.5 (0.5, 4.1) |
| Unemployed | 51 | 49 | 1.4 (0.6, 3.4) |
| Retired | 50 | 50 | 1.3 (0.3, 5.2) |
| Unable to work | 49 | 51 | 1.3 (0.5, 3.2) |
| Individual annual income, $ | |||
| <12 000 | 45 | 55 | 1.0 (Ref) |
| 12 000–23 999 | 56 | 44 | 1.5 (0.9, 2.5) |
| 24 000–47 999 | 64 | 36 | 2.1 (1.1, 4.1) |
| ≥ 48 000 | 51 | 49 | 1.3 (0.6, 2.6) |
| Unstable living situation in past 12 mo | |||
| Yes | 56 | 44 | 1.5 (1.0, 2.3) |
| No | 46 | 54 | 1.0 (Ref) |
| Currently living with a sexual partner | |||
| No | 49 | 51 | 0.8 (0.3, 2.0) |
| Yes, male | 56 | 44 | 1.0 (Ref) |
| Yes, female | 61 | 39 | 1.3 (0.4, 4.0) |
| Yes, transgender | 50 | 50 | 0.8 (0.1, 7.0) |
| Ever incarcerated | |||
| Yes | 51 | 49 | 1.1 (0.7, 1.7) |
| No | 48 | 52 | 1.0 (Ref) |
| Sexual orientation | |||
| Heterosexual/straight/down low/other | 63 | 37 | 3.0 (1.5, 5.9) |
| Bisexual | 48 | 52 | 1.7 (0.9, 3.1) |
| Gay/same-gender loving/homosexual | 36 | 64 | 1.0 (Ref) |
| HIV status | |||
| Positive | 44 | 56 | 0.6 (0.4, 0.9) |
| Negative | 58 | 42 | 1.0 (Ref) |
| Never tested/unknown | 48 | 52 | 0.7 (0.4, 1.4) |
Note. CI = confidence interval; GED = general equivalency diploma; POR = prevalence odds ratio.
a Gender role conflict scores of selected sociodemographic characteristics were based on the median split.
b Odds of high versus low.
Table 3 shows associations between high GRCS scores and psychological, HIV knowledge, and condom use efficacy outcome variables. We observed associations between high GRCS scores and more psychological distress, internalized homophobia, and low self-esteem in our sample. We also observed lower average HIV knowledge, condom use self-efficacy, and condom use skills among participants with higher GRCS scores.
TABLE 3— Associations Between High Gender Role Conflict Scale Scores and Selected Psychological Outcomes and HIV Knowledge/Condom Use Variables Measured at Baseline Among 400 Men Who Have Sex With Men and Women: Men of African American Legacy Empowering Self Project, Los Angeles, CA, 2007–2010
| Gender Role Conflict Scale | ||
| Variable | ra | P |
| Psychological measures (continuous scale) | ||
| Brief Symptom Inventory (53 items) | ||
| Anxiety (6 items) | 0.25 | <.001 |
| Depression (6 items) | 0.20 | <.001 |
| Somatization (7 items) | 0.19 | .0002 |
| Global severity index (53 items) | 0.26 | <.001 |
| Rosenberg Self-Esteem Scaleb (10 items) | −0.32 | <.001 |
| Internalized homophobiac (10 items) | 0.45 | <.001 |
| HIV knowledge/condom use (continuous scale) | ||
| HIV knowledged (10 items) | −0.15 | .002 |
| Condom use self-efficacye (11 items) | −0.20 | <.001 |
| Condom use skillsf (7 items) | −0.26 | <.001 |
a Spearman rank correlation.
b Higher scores = higher self-esteem.
c Higher scores = greater homophobia.
d μ = 8.8; SD = 1.9; higher scores = greater knowledge.
e μ = 41; SD = 11.1; higher scores = higher efficacy.
f μ = 17.9; SD = 3.6; higher scores = higher confidence.
Table 4 shows the associations between high GRCS scores and self-reported HIV risk characteristics. Disclosure of participants’ same-sex behaviors to female partners, health care providers, and family members was lower among men with higher than lower GRC. Among participants reporting sexual behaviors in the past 90 days, men with higher GRCS scores reported higher frequencies of unprotected sexual intercourse with women and any vaginal or anal intercourse with women unaware of their MSM behavior. Men with higher GRC also reported a higher prevalence of drug use before or during sexual intercourse with partners of any gender. We observed no association between high GRCS scores and unprotected anal intercourse with men in our sample. This lack of association between high GRCS score and unprotected anal intercourse with male partners applied to all subsets of men in our analysis, regardless of their self-reported sexual orientation (data not shown). In the subsample of HIV-positive men who reported sexual intercourse in the past 90 days, HIV status disclosure with male and female (but not transgender) partners was lower among participants who reported greater GRC.
TABLE 4— Associations Between High Gender Role Conflict Scale Scores and Baseline HIV Risk Characteristics Among 400 Men Who Have Sex With Men and Women: Men of African American Legacy Empowering Self Project, Los Angeles, CA, 2007–2010
| Gender Role Conflict Scale | ||
| HIV Risk Characteristics (Dichotomous) | Mann–Whitney Rank Sum Test | P |
| Disclosure of same-sex behaviors (ever) | ||
| To any female partners | −5.05 | <.001 |
| To any health care providers | 4.71 | <.001 |
| To any family members | −5.53 | <.001 |
| Behaviors in past 90 d | ||
| Unprotected anal intercourse with a mana | −0.08 | .47 |
| Unprotected vaginal/anal intercourse with a womanb | −2.62 | .004 |
| Vaginal/anal intercourse with a woman unaware of MSM behaviorb | −2.30 | .01 |
| Drug use before or during sexual intercoursec | 1.92 | .03 |
| Disclosure to sexual partner of positive HIV status at least half the time prior to first intercourse | ||
| Male partners (n = 189) | 2.11 | .02 |
| Female partners (n = 191) | 2.46 | .007 |
| Transgender partners (n = 90) | −0.43 | .33 |
Note. MSM = men who have sex with men.
a Among 204 men who had anal intercourse with men.
b Among 261 men who had vaginal/anal intercourse with women.
c Among 312 men who had anal/vaginal intercourse with a male, female, or transgender partner.
At screening, all 400 participants reported engaging in sexual intercourse with at least 1 female partner in the past 24 months, and 261 (65%) men reported any oral, vaginal, or anal intercourse with female partners in the past 90 days. The majority of participants (61%) self-identified as bisexual (Table 1). When asked about the importance of keeping their sexual relationships with men a secret, 59% of all participants reported that secrecy was very or somewhat important to them. Secrecy about male sexual relationships was associated with both self-identified sexual orientation and GRCS scores.
Secrecy was very or somewhat important to 40% of the 53 participants identifying as gay or homosexual, 57% of the 243 participants identifying as bisexual, and 72% of the 104 participants identifying as heterosexual (X2 P < .001). Secrecy about male sexual relationships was also more important to participants with higher GRCS scores both overall and within strata of self-reported sexual orientation (Mann–Whitney rank sum test P < .001).
Ours was the first study to our knowledge to examine GRC and its association with mental health and HIV behavioral risk outcomes in a large sample of African American MSMW. We found that African American MSMW who reported higher levels of GRC also reported greater psychological distress, less HIV knowledge, lower condom use efficacy and skills, less disclosure of MSMW sexual behaviors to female partners and others, and higher prevalence of HIV risk behaviors, such as use of drugs prior to sexual intercourse and unprotected vaginal or anal intercourse with female partners, than did respondents with less GRC. Our results indicating an association between high levels of GRC and unprotected sexual behaviors with female partners are similar to those recently reported in a subsample of bisexually active African American men recruited in Atlanta, Georgia.10 Participants who reported higher levels of GRC also reported greater perceived importance of keeping their MSM behaviors a secret, regardless of whether they self-identified as homosexual, bisexual, or heterosexual, than did participants with less GRC.
Our hypothesis that high GRC would be related to unprotected anal intercourse with male partners was not confirmed in our sample of bisexually active men. Malebranche et al. also recently documented this null association between high GRC and unprotected anal intercourse with male partners in their sample of primarily gay-identified African American MSM.10 One explanation for this finding may be that the negotiation of condom use with male partners differs from negotiations with female partners, who may generally have less influence on whether a condom is used. We were not able to adjust for potentially important characteristics, such as GRC, among our participants’ male partners.
Keeping same-sex behaviors a secret was either very or somewhat important to a majority (59%) of MAALES participants at baseline. Men who reported that secrecy about same-sex behaviors was important reported greater GRC than did other participants and tended to identify as bisexual or heterosexual. An investigation of risk behaviors of HIV-positive MSM of different race/ethnicity groups found that, relative to White MSM, African American MSM reported more sexual intercourse with female partners, greater internalized homophobia, less disclosure of HIV status, and less frequent disclosure of same-sex behaviors to others.28 Our investigation extends the work of O’Leary et al. through the identification of GRC as a factor associated with, among other mental health and behavioral risk outcomes, greater internalized homophobia, less disclosure of same-sex behaviors, and less disclosure of HIV status among HIV-positive African American MSMW.
Our data were not obtained from a random sample of African American MSMW. Furthermore, they were collected from men in 1 urban HIV epicenter. Because of differences in the historical and cultural contexts, the associations observed in other parts of the United States might differ substantially. Young African American men were underrepresented in our sample, as were employed and middle- and upper-income men. For example, the median household income for African Americans in Los Angeles County in 2000 was $2819 per month; for 56% of our sample, individual monthly income was less than $1000.
The data were also cross-sectional; hence, we cannot make definitive conclusions about the direction of many of the observed associations. With the analysis of follow-up data, however, we will examine whether hypothesized reductions in GRC attributable to the MAALES intervention result in mental health and behavioral risk improvements for intervention participants.
We used a modified GRCS measure. Our substitution of the GRCS subscale measuring conflict between work and family relations with a novel set of items intended to measure masculine presentation in our specific study population, however, may also be viewed as a benefit. Researchers working with nonmainstream populations are often forced to choose between using established measures that may not resonate with participants or formulating new questions that better characterize the lived experiences and perspectives of less frequently studied populations. Finally, to reduce participant burden in an already lengthy data collection session, we elected not to use more comprehensive, multidimensional measures of masculinity, such as the Male Role Norms Inventory29 or the Conformity to Masculine Norms Inventory.30 In light of our findings derived from the GRCS among African American MSMW, future researchers may consider using additional scales to better detect how male gender role norms may affect both protective and risky health behaviors.
Our data describe baseline associations between high GRC and a host of negative mental health, HIV prevention, knowledge, and behavioral risk outcomes in a population of African American men facing numerous economic and social challenges. Although our findings are important, the ultimate goal of our research is to investigate whether reductions in GRC help to improve mental health and HIV-related outcomes following participation in the MAALES intervention.
Although our sample of younger men (< 30 years) was too small for additional subanalyses, it will be critical to examine the impact of high GRC among young African American MSM in light of their HIV incidence, which continues to be higher than in other MSM subpopulations.31 Thus, future epidemiological and intervention research among African American MSMW should consider additional ways that high GRC affects men’s lives and the specific approaches, at the individual and community levels, that may alleviate psychological distress and reduce the HIV risk behaviors associated with an individual’s internal conflict with prescribed and possibly rigid gender role norms.
Acknowledgments
This study received support from the National Center for Minority Health Disparities (grant 2P20MD000182) and the California HIV/AIDS Research Program (grants AL04-UCLA-840A and AL04-DREW-840B).
We acknowledge our community collaborators Sergio Avina, Cleo Manago, and Kevin Pickett; the input and contributions of Hema Ramamurthi, Jason McCuller, Keith Norris, Martin Shapiro, Martin Lee, and Ron Hays; and our community advisory board, staff, and study participants.
Human Participant Protection
The study protocol was approved by the institutional review boards at Charles R. Drew University of Medicine and Science and the University of California, Los Angeles.
