C 2004) Archives of Sexual Behavior, Vol. 33, No. 6, December 2004, pp. 585–591 (


Sex With Men Before Coming-Out: Relation to Sexual Activity and Sexual Risk-Taking Behavior Roger K. Schindhelm, M.D.,1,2 and Harm J. Hospers, Ph.D.3,4 Received February 26, 2003; revision received February 16, 2004; accepted March 24, 2004

This study examined the difference in sequence of coming-out and first same-sex experience in relation to risk-taking behavior in a sample of Dutch gay and bisexual men. A questionnaire assessed age of disclosure (coming-out) and age of first same-sex experience, and information on sexual history, sexual relationships, and sexual behavior. It was found that 68% of respondents engaged in their first same-sex experience before coming-out. This proportion increased with age. Men who had sex with men before coming-out reported more lifetime sex partners and more casual sex partners in the past 6 months than men in which this sequence was reversed. In addition, among this group a higher proportion of men reported STDs, engaging in anal intercourse with casual partners in the previous 6 months, and unprotected anal sex with casual sex partners in the recent past, compared to men who came out before having their first same-sex experience. The extent to which external factors (such as tolerance toward homosexuals) or internal factors (such as personality factors) can account for the difference in sexual behaviors in general and sexual risk-taking behavior in particular could be subject of further study. KEY WORDS: gay and bisexual men; coming-out; sexual experience; sexual risk-taking behavior.

Most models distinguish distinct stages that occur in a certain order. The process of homosexual identity formation starts with a period of having vague feelings of attraction to people of the same sex, followed by having sex with other boys or men. The process continues with self-labeling as a homosexual, followed by disclosure of sexual identity to others (coming-out), becoming involved in a same-sex romantic relationship and exploring gay venues (also referred to as “coming-in”), and ending with the integration and consolidation of a homosexual identity. However, recent studies have shown that developmental sequences can vary between individuals and that a considerable number of gay men went through their self-identification and coming-out process before having sexual experience with other men (Dub´e, 2000; Dub´e & Savin-Williams, 1999; Maguen, Floyd, Bakeman, & Armistad, 2002; Savin-Williams & Diamond, 2000). Dub´e (2000) studied a sample of 166 American gay and bisexual men. Participants were divided into two groups according to the sequence through which the men identified their same-sex attractions. They were coded as either sex-centered (labeling as homosexual after first same-sex experience) or identity-centered (labeling

INTRODUCTION Several “coming-out” models of sexual identity development of sexual minority men have been proposed (e.g., Cass, 1979, 1984; Coleman, 1982; Martin, 1991; Rotheram-Borus & Fernandez, 1995; Troiden, 1979). Most coming-out models presume that awareness of same-sex attraction emerges in early childhood, i.e., before the age of 10 (Herdt & McClintock, 2000), and that the awareness and exploration of same-sex attraction takes place during adolescence and eventually results in an integration of a homosexual identity into personality.

1 Faculty

of Medicine, Maastricht University, Maastricht, The Netherlands. 2 Present address: Department of Endocrinology, VU University Medical Center, Amsterdam, The Netherlands. 3 Department of Experimental Psychology, Maastricht University, Maastricht, The Netherlands. 4 To whom correspondence should be addressed at Department of Experimental Psychology, Maastricht University, POB 616, 6200 MD Maastricht, The Netherlands; e-mail: [email protected].

585 C 2004 Springer Science+Business Media, Inc. 0004-0002/04/1200-0585/0


586 as homosexual before first same-sex experience). The participants were further divided into three age cohorts: the first cohort consisted of men between 16 and 21 years of age, the second cohort consisted of men between 22 and 29 years of age, and the third cohort consisted of men between 30 and 39 years of age. The aim of Dub´e’s (2000) study was to test whether the difference in sequence predicted the later adjustment to sexual identity development and relationship involvement. Fifty-eight percent of the men in Dub´e’s sample developed through the sex-centered sequence. This proportion increased with age: most men of the older cohorts developed through a sex-centered sequence, whereas the majority of the younger men developed through an identity-centered sequence. Dub´e (2000) found that the sequence was unrelated to the extent to which participants had disclosed their sexual identity to others. Men who followed a sex-centered sequence compared to men who followed an identity-centered sequence reported a higher level of internalized homophobia, measured by attitudes toward homosexuality in general and toward other gay men. Dub´e speculated that this internalized homophobia may be the result of projecting negative feelings about their homosexuality onto self-identified men with whom they have sex. Sequence was unrelated to the number of same-sex steady relationships. Finally, sex-centered men reported more lifetime sex partners than identity-centered men. Dub´e (2000) suggested that developmental sequence may also predict high-risk sexual behavior (i.e., unprotected anal intercourse) and that sex-centered men might be at greater risk for STD and HIV infection. The aim of this study was to elaborate on Dub´e’s findings using data from a sample of Dutch men who have sex with men (MSM). First, we investigated the extent to which Dutch MSM have followed a sex-centered sequence or an identity-centered sequence. As it is well known that that the Dutch have the most liberal attitudes toward homosexuality and that homosexuality is well accepted (Numerical & Spatial Data Services, 2004), one would expect that the proportion of men who followed the identity-centered sequence would be substantially higher than Dub´e found is his U.S. sample. Second, we investigated whether Dub´e’s findings regarding sexual history variables are replicated in the Dutch sample. Finally, we tested the hypothesis that sex-centered MSM show higher levels of sexual risk behavior than identity-centered MSM. METHOD Participants Participants were 961 MSM (M age, 37 years; range, 15–80) who were part of a cross-sectional survey of

Schindhelm and Hospers Dutch MSM. This survey was set up to collect data through anonymous questionnaires on demographics, sexual history, recent sexual behavior with steady and casual partners, and acculturation into the gay subculture. The results of the survey allowed Dutch public health departments to develop HIV/AIDS prevention activities for MSM. In 1999, 2,442 questionnaires were distributed at meeting places for MSM (such as gay bars, cruising areas, gay saunas, gay dance parties, etc.), of which 1,336 were returned (Hogeweg & Hospers, 2000). Men who had their first sexual experience with men at the same age as their coming-out (N = 336), men who did not have sexual experience with men and who had not yet come out (N = 2), and men who had missing values on one or more of the classification variables (N = 37) were excluded from the analyses. The analyses were conducted on the remaining 961 participants.

Measures The questionnaire asked for demographic information, such as age and educational level (lower: high school or less; medium: high school plus vocational training; or higher: college or university training). The participants were further asked whether they ever told others that they were sexually attracted to men (coming-out) and, if so, how old they were when they first told someone that they were sexually attracted to men. Furthermore, participants were asked whether they ever had sex with men and, if so, at what age they had their first male sexual experience. Participants were asked to select the identity label that best described their current sexual identity on a 5-point scale, ranging from completely heterosexual to completely homosexual. Self-acceptance of homosexuality was measured with seven items (e.g., “It’s easy for me to talk to others about my homosexuality”). The items were scored on a 5-point scale, ranging from completely disagree to completely agree. Cronbach’s α for this measure was .79. For the analyses, the mean of the seven items was computed. Disclosure was measured by asking participants to indicate to which of seven clusters of family members and acquaintances they had disclosed their sexual identity (mother, father, brothers and sisters, majority of other family members, majority of colleagues, majority of heterosexual friends, general practitioner). For each subject, the total number of disclosure clusters was computed. Participants were asked a number of questions related to sexual history, sexual relationships, and sexual behavior. First, participants were asked to estimate the total number of lifetime male sex partners. Rather than using actual numbers of lifetime male sex partners, which

Coming-Out and Risk-Taking Behavior

587

Table I. Means and Proportions for General Characteristics of the Identity-Centered and Sex-Centered Groups Developmental sequence Identity-centered

Age (in years) Number of years sexually active with men Sexual identity Exclusively homosexual Not exclusively homosexual Educational level Lower Medium Higher Age first sex with a man (in years) Age coming-out (in years) Self-acceptance Disclosure

Sex-centered

M

SD

M

SD

F or χ 2

p

33.3 10.2

10.4 9.0

39.2 22.4

11.6 12.7

56.1 229.2

<.001 <.001

78.8% 21.2%

74.2% 25.8%

12.8% 26.6% 60.7% 22.8 20.1 4.1 5.7

23.8% 35.8% 40.4% 16.7 24.0 4.0 5.1

typically result in a highly skewed distribution, we used a 7-category system that has been shown in previous research (Hogeweg & Hospers, 2000) to result in a normal distribution (categories: 1, 2–5, 6–10, 11–50, 51–100, 101–501, 500 and over). Second, participants were asked if they ever had engaged in anal sex, if they ever were tested for HIV, and if they ever had contracted a sexually transmitted disease (excluding HIV and pubic louse). Third, participants were asked to indicate whether they were involved in a steady relationship with a male at assessment. Fourth, information about sexual behavior was gathered by asking the occurrence of anal sex with casual partners in the 6 months preceding assessment, and the use of condoms for anal sex with casual partners. Subject Characteristics Twenty percent had a lower educational level, 33% a medium educational level, and 47% a high educational level. Men who had their coming-out before their first sexual experience with men (N = 286) and men who had their coming-out but no sexual experience with men (N = 23) comprised the identity-centered group (IC), totaling 309 participants (32% of the sample). Men who had their first sexual experience with men before coming-out (N = 596) and men who had sexual experience with men but had not come out (N = 56) comprised the sex-centered group (SC), totaling 652 participants (68% of the sample). The majority of participants (75.7%) labeled themselves as exclusively homosexual, 17.1% as more homosexual than heterosexual, 5.5% as bisexual, and 1.8% as more heterosexual than homosexual.

7.3 6.7 0.7 1.5

2.4

5.9 7.9 0.9 2.3

Lower vs. medium 2.1 Medium vs. higher 19.5 179.9 54.4 3.1 15.6

ns

ns <.001 <.001 <.001 ns <.001

RESULTS General Characteristics Table I summarizes the results described in this paragraph. Participants in the SC group were significantly older compared to participants in the IC group (39.2 vs. 33.3 years). Of the men aged 30 and younger, 52.2% had followed the sex-centered sequence, compared to 68.4% of the men aged 31–40 years, and 80.8% of the men older than 40 years. Men in the SC group had been sexually active with men significantly longer than men from the IC group (22.4 vs. 10.2 years). The proportion of exclusively homosexual men did not differ significantly between the groups (IC, 78.8%; SC, 74.2%). The IC group comprised significantly more men with a higher educational level than the SC group (60.7% vs. 40.4%). However, educational level did not reach significance in any of the analyses that follow. The mean age of participants’ first same-sex experience was 18.6 years and the mean age of participants’ coming-out was 22.6 years. The men in the SC group had their first same-sex experience at 16.7 years whereas their mean coming-out age was 24.0 years. For the men in the IC group, these ages were 22.8 and 20.1 years, respectively. Mean self-acceptance was 4.0 (range, 1– 5), and did not differ between the SC and the IC group. Sixty-one percent of participants had disclosed their sexual identity to six or all of the seven clusters of family members and acquaintances. The mean number of disclosure clusters was higher in the IC group than in the SC group (5.7 vs. 5.1). However, if men in the SC group were excluded who had not (yet) come out and thus

588 had 0 disclosure clusters, the IC group and SC group were comparable with respect to disclosure (5.7 vs. 5.6).

Sexual History A series of multiple logistic regression analyses were conducted to investigate the relationship between developmental sequence (SC vs. IC) and sexual history variables (see Table II). The number of years that participants were sexually active was simultaneously added as a predictor variable in these analyses.5 With respect to experience with anal sex, it was found that years of sexual activity was not significantly related, but that sequence was 87% of participants in the SC group had ever engaged in anal sex compared to 75% men in the IC group. Compared to the IC group, significantly more men from the SC group ever contracted an STD (38% vs. 18%) and more men had had an HIV test (44% vs. 33%). Also, men who were longer sexually active were more likely to have had an STD and conducted an HIV-test. Of the SC men, 56% reported more than 50 lifetime sex partners compared to 25% of the IC men. The number of lifetime sex partners was also positively related to years of sexual activity. Finally, developmental sequence and years of sexual activity were unrelated to having a steady partner or not.

Recent Sexual Activity With Casual Partners Sexual behavior with casual partners in the 6 months preceding assessment was subjected to multiple logistic regression analyses with developmental sequence and years of sexual activity as predictors (see Table III). These analyses showed no effect of years of sexual activity, but did find significant differences between SC and IC groups. First, a higher percentage of the SC men reported casual partners in the previous 6 months compared to the IC men (80% vs. 61%), reported anal sex with these casual sex partners (44% vs. 26%), and reported unprotected anal sex with these casual sex partners (15% vs. 9%). Men in the SC group reported significantly more casual sex partners in the previous 6 months than men in the IC group (12.9 vs. 7.3, respectively), F (1, 952) = 3.9, p < .05 (controlling for sexual experience). 5 We

considered including both age and sexual experience in these analyses, as either or both could account for differences between the SC and IC groups. The correlation between age and years of sexual activity was .84, which points to multicollinearity of predictors when entered simultaneously. Therefore, years of sexual activity with men was chosen to present in this paper. However, all analyses were also conducted with age as predictor, yielding identical results.

Schindhelm and Hospers DISCUSSION The aim of this study was to elaborate on Dub´e’s (2000) study in a sample of Dutch MSM and to study the difference in developmental sequence (first samesex experience before or after coming-out) with respect to sexual risk-taking behavior. It was found that 68% of men in our sample developed through a sequence in which first same-sex experience preceded comingout. This proportion varied from 52% for men under 30 to 81% for men over 40 years old. Neither sexual identity nor involvement in a same-sex steady relationship was related to sequence. Furthermore, self-acceptance of homosexuality was related to sequence, but this difference could be attributed to the men in the sex-centered group who had not come out. Dub´e (2000) used self-labeling as homosexual to differentiate developmental sequence, whereas actual coming-out was used in the present study. When applying the age restriction that Dub´e used (16–39 years) to the Dutch data, the proportion of MSM that followed a sexcentered sequence was similar. If age of self-labeling had been available in the present study, it is likely that the percentage of identity-centered men would have been considerable higher, because coming-out follows selflabeling as homosexual, and may in fact occur several years after self-labeling (or not at all). These results may reflect the more liberal and tolerant environment for homosexuals in The Netherlands (Numerical & Spatial Data Services, 2004). This is corroborated with the finding that in this study a substantially higher proportion of respondents had disclosed their sexual identity to all or almost all relevant others (61% vs. 44% in Dub´e’s study). Dub´e (2000) also found that sex-centered men reported more lifetime sex partners than identity-centered men. This result was replicated in our study. A major objective of the present study was to empirically test Dub´e’s suggestion that sex-centered men engage in more sexual (risk) behavior than identity-centered men. Indeed, with respect to recent sexual behavior with casual partners, we found that sex-centered men were more likely to report casual sex partners in the previous 6 months, to report anal sex with casual partners, and to report unprotected anal sex with casual partners. Also, sexcentered men reported more casual sex partners than did identity-centered men. Finally, it was found that sexcentered men were more likely to have contracted STDs and to have conducted an HIV test. Unprotected anal sex, especially in combination with a higher number of casual sex partners, is linked with a higher probability for contracting HIV (Eisenberg, 1989). Having had an STD

75% 33% 18% 25% 50%

Experience with anal sex Ever having an HIV-test Ever contracting STD >50 lifetime male sex partners Having steady partner

87% 44% 38% 56% 53%

Sex-centered .72 (.20) .34 (.16) .44 (.19) .61 (.17) .02 (.16)

B (SE) 12.8 4.4 5.7 12.7 0.0

Wald .001 .04 .02 .001 ns

p 2.04 1.40 1.56 1.84 1.02

eB

.15 .14 .27 .39 .05

Spearman’s ρ

Note. The univariate correlation between sexual experience and each of the dependent variables is expressed by Spearman’s ρ.

Identity-centered

Dependent variables

Developmental sequence

.01 (.01) .01 (.01) .05 (.01) .07 (.01) .01 (.01)

B (SE)

2.7 4.4 54.4 102.2 1.6

Wald

Sexual experience

ns .04 .001 .001 ns

p

1.01 1.05 1.07 1.00

eB

Table II. Multivariate Logistic Regression on Sexual History Variables, With Developmental Sequence (Identity-Centered vs. Sex-Centered) and Years of Sexual Activity With Men as Predictors

Coming-Out and Risk-Taking Behavior 589

15%

9%

.66 (.25)

.79 (.17) .97 (.17)

B (SE)

7.2

21.3 32.5

Wald

.007

.001 .001

p

1.56

2.20 2.64

eB

.03

.17 .07

Spearman’s ρ

Note. The univariate correlation between sexual experience and each of the dependent variables is expressed by Spearman’s ρ.

80% 44%

61% 26%

Having casual sex partners Having anal sex with casual sex partners Having unprotected anal sex with casual sex partners

Sex-centered

Identity-centered

Dependent variables

Developmental sequence

3.1 2.9 0.4

−.01 (.01)

Wald .01 (.01) −.01 (.01)

B (SE)

Sexual experience

ns

ns ns

p

1.00

1.01 0.99

eB

Table III. Multivariate Logistic Regression on Sexual (Risk) Behavior With Casual Partners in the 6 Months Preceding Assessment, with Developmental Sequence (Identity-Centered vs. Sex-Centered) and Years of Sexual Activity With Men as Predictors

590 Schindhelm and Hospers

Coming-Out and Risk-Taking Behavior is considered to be a marker for unsafe sexual practice and STDs could increase risk for transmitting HIV. The latter relationship has been demonstrated for heterosexual HIV transmission (Clottey & Dallabetta, 1993; Fleming & Wasserheit, 1999), but due to limited adequate studies, this relationship in homosexual transmission is less clear (Bonnell, Weatherburn, & Hickson, 2000). Dub´e (2000) suggested that the higher rates of sexual behavior reported by sex-centered men might be the result of viewing sex as the central element of their sexual identity. One possible explanation that Dub´e provided is that sex-centered men have had lesser access to resources that could help in the homosexual identification process, thus pointing to emancipatory processes. Alternatively, personality factors could also contribute to explaining the findings. It has been suggested that impulsivity and sensation-seeking are related to higher sexual activity as well as to more sexual risk-taking behavior (Kalichman et al., 1994; McCoul & Haslam, 2001; Sheer & Welch Cline, 1995). Scoring high on impulsivity and sensationseeking may also be related to having sex before comingout. To our knowledge, however, this conjecture has never been studied. This study has several limitations. In the first place, the results presented may not be representative for all MSM. The distribution in educational level is different from the general population: higher educated men were overrepresented in the present sample. In the second place, questions regarding recall in age of first samesex experience or age of coming-out are subject to retrospective memory bias (Catania, Gibson, Chitwood, & Coates, 1996; McFarlane & St. Lawrence, 1999). Finally, we had to exclude respondents whose age of comingout and age of first same-sex experience coincided. Asking directly whether coming-out preceded same-sex experience or vice versa would have allowed us to use the full sample. Notwithstanding these limitations, these data suggest that men who had their first same-sex experience before disclosure engage in higher sexual risk-taking behavior. Because the price that is nowadays attached to high-risk-taking behavior is so high, we believe further research into sexual identity process sequences is warranted. REFERENCES Bonnell, C., Weatherburn, P., & Hickson, F. (2000). Sexually transmitted infection as a risk factor for homosexual HIV transmission: A

591 systematic review of epidemiological studies. International Journal of STD and AIDS, 11, 697–700. Cass, V. C. (1979). Homosexual identity formation: A theoretical model. Journal of Homosexuality, 4, 219–235. Cass, V. C. (1984). Homosexual identity formation: Testing a theoretical model. Journal of Sex Research, 20, 143–167. Catania, J. A., Gibson, D. R., Chitwood, D. D., & Coates, D. R. (1996). Methodological problems in AIDS behavioral research: Influence on measurement error and participation bias in studies of sexual behavior. Psychological Bulletin, 108, 339–362. Clottey, C., & Dallabetta, G. (1993). Sexually transmitted diseases and humane immunodeficiency virus: Epidemiologic synergy? Infectious Disease Clinics of North America, 7, 753–770. Coleman, E. (1982). Developmental stages of the coming out process. In J. Gonsiorek (Ed.), Homosexuality and psychotherapy (pp. 31–44). New York: Hawthorne. Dub´e, E. M. (2000). The role of sexual behavior in the identification process of gay and bisexual males. Journal of Sex Research, 37, 123–132. Dub´e, E. M., & Savin-Williams, R. C. (1999). Sexual identity development among ethnic sexual-minority male youth. Developmental Psychology, 34, 1389–1398. Eisenberg, B. (1989). The number of partners and the probability of HIV infection. Statistics in Medicine, 8, 83–92. Fleming, D. T., & Wasserheit, J. N. (1999). From epidemiological synergy to public health policy and practice: The contribution of other sexually transmitted diseases to sexual transmission of HIV infection. Sexually Transmitted Infections, 76, 3–17. Herdt, G., & McClintock, M. (2000). The magical age of 10. Archives of Sexual Behavior, 29, 587–606. Hogeweg, J., & Hospers, H. J. (2000). Monitoronderzoek 2000 [Monitoring Study 2000]. Maastricht: Maastricht University. Kalichman, S. C., Johnson, J. R., Adair, V., Rompa, D., Multhauf, K., & Kelly, J. A. (1994). Sexual Sensation Seeking: Scale development and predicting AIDS-risk behavior among homosexually active men. Journal of Personality Assessment, 62, 385–397. Maguen, S., Floyd, F. J., Bakeman, R., & Armistad, L. (2002). Developmental milestones and disclosure of sexual orientation among gay, lesbian and bisexual youth. Journal of Applied Developmental Psychology, 23, 219–233. Martin, H. P. (1991). The coming-out process for homosexuals. Hospital and Community Psychiatry, 42, 158–162. McCoul, M. D., & Haslam, N. (2001). Predicting high risk sexual behavior in heterosexual and homosexual men: The roles of impulsivity and sensation seeking. Personality and Individual Differences, 31, 1303–1310. McFarlane, M., & St. Lawrence, J. S. (1999). Adolescents’ recall of sexual behavior: Consistency of self-reports and effect of variations in recall duration. Journal of Adolescent Health, 25, 199–206. Numerical & Spatial Data Services. (2004). World Values Survey 1990. Retrieved on February 1, 2004, from World Values database (http://nds.umdl.umich.edu/cgi/s/sda/hsda?harcWEVS+wevs) Ann Arbor: University of Michigan. Rotheram-Borus, M. J., & Fernandez, M. I. (1995). Sexual orientation and developmental challenges experienced by gay and lesbian youths. Suicide and Life-Threatening Behavior, 25, 26–34. Savin-Wlliams, R. C., & Diamond, L. M. (2000). Sexual identity trajectories among sexual-minority youth: Gender comparisons. Archives of Sexual Behavior, 29, 607–627. Sheer, V. C., & Welch Cline, R. J. (1995). Individual differences in sensation seeking and sexual behavior: Implications for communication and intervention for HIV/AIDS prevention among college students. Health Communication, 7, 205–223. Troiden, R. R. (1979). Becoming homosexual: A model of gay identity acquisition. Psychiatry, 42, 362–373.