Accounting for Unsafe Sex: Interviews With Men Who Have Sex With Men.

Educational campaigns designed to increase awareness of AIDS and safe sex practices have made a remarkable impact.(1) National populations as a whole, and gay men in particular, are now highly knowledgeable about the ways in which HIV can be transmitted (e.g., Jadack, Hyde, & Keller, 1995; Lewis, Malow, & Ireland, 1997; Myers, Godin, Calzavara, Lambert, & Locker, 1993; O'Donnell, San Doval, Vornfett, & O'Donnell, 1994). Many gay men have managed to integrate safe sex into their everyday lives (e.g., Ekstrand & Coates, 1990; Hunt et al., 1993), yet this change remains incomplete (e.g., Gruer & Ssembatya-Lule, 1993; Hunt et al., 1993; Kelly et al., 1995; Meyer & Dean, 1995; Offir, Fisher, Williams, & Fisher, 1993; Perkins, Leserman, Murphy, & Evans, 1993; Peterson et al., 1992). Whereas some gay men occasionally relax the safety standards that usually govern their sexual activities, others have never adopted safe sex to a great degree. Because the prevalence of HIV-infection in the Western world remains higher among men who have sex with men than among any other segment of the population, unsafe male-to-male sex continues to pose a major public-health challenge.

Much of the research on safe sex education has focused on the association between knowledge about AIDS and sexual practices (see Joffe, 1996; Kippax, Connell, Dowsett, & Crawford, 1993). In the initial phase of the epidemic, primary concerns included tailoring educational programs to specific high-risk populations (e.g., gay men, injection-drug users), and identifying those who were undereducated about AIDS. A central assumption of this approach was that information would shift attitudes, which, in turn, would lead to changes in behavior. As Kippax et al. (1993) remark,

   The dominant model of health education that has been adopted by many AIDS
   researchers, particularly in the United States, is a refinement of what
   might be called the KAP (or KAB) model--knowledge, attitudes, practices (or
   behavior). The KAP model is a linear one, which initially assumed that
   knowledge shapes or determines attitudes which, in turn, shape or determine
   behavior. (p. 5)

As levels of knowledge and awareness about AIDS rose rapidly in many countries throughout the 1980s, it became increasingly clear that rates of safe sex practice often failed to keep up. For example, in one study (Fisher, Fisher, Williams, & Malloy, 1994), a KAP model accounted for 35% of the variance in gay men's AIDS-preventive behavior. Although this association is relatively strong in social-science research, it should not go unnoticed that almost two thirds of the variance could not be attributed to individual differences in knowledge of HIV transmission. Indeed, as people in the European Union, North America, and Australia became increasingly knowledgeable about the basic facts of how HIV is spread, the attention of researchers turned toward the gap between knowledge and behavior (e.g., Catania, Kegeles, & Coates, 1990; Waddell, 1992), which tends to be particularly large for heterosexuals (Maticka-Tyndale, 1992; Maxwell, Bastani, & Yan, 1995; Maxwell & Boyle, 1995).

Accordingly, issues of compliance and relapse in terms of conformity to safe sex norms came to dominate the agendas of many researchers, who asked why people "who know better" still have unsafe sex (e.g., Adib, Joseph, Ostrow, & James, 1991; Kelly et al., 1991; McCusker, Stoddard, McDonald, Zapka, & Mayer, 1992; see Hart, Boulton, Fitzpatrick, McLean, & Dawson, 1992, for a review and critique). This approach attempts to identify factors that compromise rationality, the assumption being that rationality would win out if these factors were eliminated (Hart et al., 1992). Perhaps the factor most commonly identified as an impediment to safe sex is substance abuse (Ekstrand & Coates, 1990; Kalichman et al., 1994; Leigh, 1990; Lewis & Ross, 1995; Meyer & Dean, 1995; Paul, Stall, & Davis, 1993; Penkower et al., 1991; Peterson et al., 1992; Siegel, Mesagno, Chen, & Christ, 1989; Waddell, 1992). Nonetheless, results from studies conducted in the United Kingdom (Weatherburn et al., 1993), the United States (Temple & Leigh, 1992), Australia (Gold, Skinner, & Ross, 1994), Belgium (Bolton, Vincke, Mak, & Dennehy 1992), and Canada (Myers et al., 1992) raise questions about the reliability of this finding.

The "rational man" approach also tends to equate all homosexual practices with high levels of risk, presuming that the simple intervention of the condom suffices as the technique for HIV prevention in all circumstances. Prevention strategies that advocate using a condom every time are consistent with this approach. We know, however, that many male-to-male sex practices carry no risk of transmitting HIV (e.g., mutual masturbation), whereas others have negligible (e.g., anilingus) or low (e.g., receptive fellatio) risk (McClure & Grubb, 1999). Even unprotected anal intercourse carries no risk of transmitting HIV if neither participant is infected with HIV. Traditional KAP and rational-man models also fail to provide a clear understanding of how AIDS messages interact with the meanings and reasoning that homosexually active men use in understanding and expressing their sexuality (Adam, 1992; Donovan, Mearns, McEwan, & Sugden, 1994; Dowsett, 1996; Flowers, Sheeran, Beail, & Smith, 1997; Joffe, 1996; King, 1993; Kippax et al., 1993; Parker & Carballo, 1990). In fact, such models--in combination with discourse about "gaps," "relapse," and so on--tend to gloss over the ways in which standards of practice for safe sex are often provisional and inconsistent.

In the present study, we interviewed homosexually active men about their sexuality. A primary goal was to obtain detailed accounts of recent incidents of unsafe sex. We predicted that levels of knowledge of HIV and AIDS would be uniformly high, and that unsafe sex would be associated with relatively complex issues, such as participants' sexual and affective preferences, the nature of the relationship between partners, ambiguities about what is truly safe, and the influence of dominant cultural values and discourse. In other words, we did not expect that unsafe sex would be best attributed to gaps in knowledge, or to complicating factors that block the progression from knowledge to attitudes to practices. Indeed, gay communities have often displayed a rich expertise around AIDS and HIV, and the history of AIDS has seen important clashes between competing forms of expertise (Altman, 1994). Some of the most successful safe sex campaigns have built on existing community infrastructures and know-how (see Dowsett, 1990; Watney, 1989). We hoped, then, that by obtaining a detailed account of men's understandings of unsafe sex, we could inform the development of future prevention efforts.

We assumed that men who had been sexually active with other men in the recent past would have first-hand familiarity with safe sex practices and the discourse surrounding such practices. By conducting in-depth, semi-structured interviews with a sample of these men, we hoped to uncover affective and emotional preferences that influence sexual decision making. Our qualitative approach enabled us to highlight the complexities men encounter in the real world of sexual interaction. Understanding such complexities is crucial if education campaigns are to engage with men's existing sexual knowledge and practices. Compared to standard quantitative approaches, qualitative research methods may provide a deeper understanding of perceptions of AIDS messages and the reasoning processes by which people exempt themselves from these messages (Donovan et al., 1994; Maticka-Tyndale, 1992). To anticipate the results somewhat, our findings imply that it is not simply the absence of information that leads men to unsafe sexual practices, but also the presence of particular preferences, feelings, and worldviews. Thus, our main argument is that the next generation of safe-sex education programs will need to engage with the complex considerations that guide men's sexual decision-making (see also Flowers, Sheeran, et al., 1997).

Previous empirical and theoretical studies with gay and bisexual men point to a range of understandings about sexual practices that appear to influence how men make decisions around safe sex. Studies from different countries have identified several issues that recur, including the association of safe sex with emotional and physical distance, distrust, and thoughts of death (Prieur, 1990); the demonstration of trust in couples (Connell, Davis, & Dowsett, 1993); the meaning of semen and insemination (Odets, 1994); the perceived riskiness of ejaculation in oral sex (Lowy & Ross, 1994); the transgressive value of homosexuality (Pollak, 1988); and the rhetoric of drive, lapse, and inadvertence (Boulton, McLean, Fitzpatrick, & Hart, 1995).

In one study (Gaies, Sacco, & Becker, 1995), American gay men were asked to imagine a sexual encounter with a very attractive partner. Although respondents reported thinking about the personal consequences of HIV infection, they also considered whether insisting on using condoms might ruin the "golden opportunity," whether the promise of pleasure justified some risk, and how important it was for them to have sex with the imagined partner. In an Australian study (Gold et al., 1994), HIV-positive gay men rationalized unsafe sex in the following terms: being in a negative mood state, having nothing to lose, taking advantage of a sexual opportunity, disliking condoms, resolving to withdraw without ejaculation, and demonstrating confidence in themselves and their partners. By contrast, HIV-negative men rationalized unsafe sex by believing they could infer a partner's HIV-status, being in a long-term relationship, wanting time out from having to worry about AIDS, disliking condoms, and being in a negative mood state. These findings are comparable to those from other reports (Kelly et al., 1991; Levine & Siegel, 1992; Offir et al., 1993). In each case, gay men justified unsafe sex in terms of a highly refined calculus of risk, or excused it by mentioning factors such as drunkenness, sexual passion, emotional needs, or partner coercion. Denying one's vulnerability to AIDS is not solely the province of gay men; similar factors were identified in a sample of the general population from the Netherlands (Sandfort & van Zessen, 1992).

In sum, the goals of the present study were threefold: (a) to examine the influence of sexual and affective preferences on gay and bisexual men's sexual behavior, (b) to investigate the ways these men define their own sexual activities as safe or unsafe, and (c) to improve our understanding of the reasoning processes involved when determining whether to practice safe sex.

METHOD

Participants

A convenience sample of 102 men who have sex with men was recruited to be as diverse as possible in terms of their cultural background, sexual identity, relationship status, age, social class, and educational level. The men lived in the Ottawa (n = 26), Toronto (n = 37), and Windsor (n = 37) areas, or in other communities in the Southern Ontario region of Canada (n = 2). Recruitment strategies were similar to those used previously to study homosexually active men (e.g., Carballo-Dieguez, Remien, Dolezal, & Wagner, 1999; Gold et al., 1994; Kalichman et al., 1994; Myers et al., 1992; Peterson et al., 1992; Waddell, 1992). These included advertising in the gay press, making appeals at meetings of gay organizations, and distributing leaflets at gay bars in the three target cities. Additional participants were recruited through the personal networks of the researchers and research assistants.

The participants ranged in age from 19 to 72 years (M = 35.2 years). Most of the men identified their sexual orientation as gay or homosexual (n = 86), with the remainder identifying as bisexual (n = 10), heterosexual (n = 2), or other (e.g., "bi-curious," n = 4). A majority of the participants had tested negative for HIV (n = 81); the remaining men were either HIV-positive (n = 11) or had not been tested (n = 10). The sample was relatively well-educated: 12 men had a graduate degree and 46 had completed college or university. Of the remaining men, 30 had some university or college education and 6 had finished high school; another 8 did not graduate from high school. Participants' incomes varied greatly: 20 of the men earned more than $50,000 per year, whereas 30 earned less than $10,000. The remaining participants had incomes ranging from $10,000 to $20,000 (n = 21), or from $20,000 to $50,000 (n = 31). (Incomes are reported in Canadian dollars: $1 US = $1.45 Canadian.)

English was the first language of 74 men, 15 first spoke French, and 13 reported that their mother tongue was a language other than English or French. When asked which language they usually spoke, 90 spoke English, 11 spoke French, and 1 spoke Spanish. First-named responses to a question about ethnicity included British (n = 36), French (n = 17), other European (n = 18), Canadian (n = 11), African and Afro-Caribbean (n = 5), Latin American (n = 4), Asian (n = 4), aboriginal (n = 2), Middle Eastern (n = 2), and Jewish (n = 2). When asked about their relationship status, the vast majority of the sample chose one of three categories: single or not dating (n = 46), dating a man (n = 26), or living with a male partner (n = 24). Other participants reported that they were involved in a dating relationship with a woman (n = 3), living with a female partner (n = 1), living with a male partner and dating a woman (n = 1), or dating both a man and a woman (n = 1).

On average, the interview took approximately 90 minutes. Participants received token remuneration ($20 Canadian) for their time and trouble.

Procedure

After being briefed on the overall objectives of the study and providing their written consent to participate (including consent to audiotape the interview), the participants completed a short questionnaire in addition to being interviewed. The questionnaire consisted of demographic categories, a checklist of sexual activities practiced with regular and casual male partners, and a knowledge measure that required participants to rate 11 male-to-male sex acts in terms of the likelihood that each act could result in transmission of HIV. Detailed analyses of these quantitative data will be reported elsewhere. The interviews were semi-structured, with pre-established "probes" used to ensure that we obtained detailed information about recent and memorable sexual experiences, particularly those that participants considered to be unsafe. Many of the probes were designed to elicit factors that make sex particularly satisfying, based on the assumption that satisfaction of desire would exert a strong influence on decision making during sexual interactions. The present report is based primarily on responses to the probe: "Describe the last time you had unsafe sex with a man."

Most of the interviews were conducted by research assistants who were graduate students in clinical or applied psychology programs and had research interests in human sexuality. A bilingual assistant interviewed French-speaking

participants. Approximately 10% of the interviews were conducted by the authors.

Analytic Strategy

Tapes of the interviews were transcribed and organized using qualitative research software (QSR NUD-IST), which allowed us to identify all comments related to instances of unsafe sex. The analyses involved identifying recurring themes the men used in their accounts and organizing informative quotes around these themes. We assumed that people in their everyday activities deploy rich ways of knowing that are integrally related to their behaviors (see Smith, 1987). Our intent was to highlight connections between unsafe sex on the one hand, and men's understanding of their sexuality and of HIV/AIDS discourse on the other hand.

The analyses followed a modified version of grounded-theory principles. Our analytical strategy was to capture emerging themes from the interviews rather than to code the data into a preexisting scheme. However, in contrast to the classical grounded-theory approach of Glaser and Strauss (1967), we do not claim that our initial reading of the transcripts was naive. Our own experiences as gay men and AIDS activists as well as our prior theoretical orientations (as discussed in the introduction) obviously shaped our approach to the data.

The analytic process involved six steps:

   1. In a preliminary reading of the transcripts, we identified key themes
   emerging from participants' reflections on their self-defined unsafe sexual
   activities.

   2. Quotes relevant to the themes were grouped together.

   3. The thematic categories were refined, merged, or subdivided based on
   associations and overlap among quotes. For example, two quotes that seemed
   initially to be making the same point might seem quite different when
   placed side-by-side.

   4. Representative quotes from each of the refined thematic groups were
   selected and connected according to the narrative of the paper.

   5. Every interview was read in detail and categorized according to the
   presence or absence of a particular theme.

   6. The thematic structure was further refined by the authors until a
   consensus was reached.

RESULTS AND DISCUSSION

Whenever participants are quoted in the qualitative analyses that follow, demographic characteristics (sexual identity, age, HIV-status, and occupation) are reported in parentheses after each quote. These represent self-reported categories. In other words, previously married men with children are classified as "gay" rather than "bisexual" if that was the participant's choice. Similarly, men who provided nonstandard descriptions of their sexual identity (e.g., omnisexual) are classified accordingly. While cultural differences are important considerations in any social-science research (e.g., Diaz, 1998; Landrine, 1995; Myrick, 1999), no cultural group among our non-White participants was represented by more than five men. (In general, Canadian society is multicultural without having large minority categories such as African American or Latino.) Accordingly, our analyses of the interviews do not address the issue of cultural differences in safe sex practices.(2)

Overview

Most of the men reported some sort of sexual activity they defined as unsafe(3) during the last 5 years. Eleven men stated categorically that they never had unsafe sex; these claims were consistent between the interview and the checklist. Another 19 men reported no unsafe sexual acts over the past 5 years. Many of these men had to reach back to the early or mid-1980s--when widespread awareness of the AIDS epidemic was just emerging--to identify the last time they had unsafe sex. For example, one man remarked, "I had unsafe sex probably from 1968 to 1986. The idea of safe sex was not in my vocabulary or my understanding" (Gay, 41, HIV-positive, businessman). Another man noted, "Well it depends on the definition. If you mean fucking or getting fucked without a condom, then it must have been with Ted. That was more than 15 or 16 years ago" (Gay, 59, HIV-negative, retired). This leaves 72 participants who reported at least one incident of unsafe sex over the last 5 years, a number that is relatively high by epidemiological standards. The second quote ("it depends on the definition") makes it clear, however, that some participants were aware of discrepancies in various definitions of safe and unsafe. Moreover, despite a consensus in our sample that unprotected anal sex is risky, other sexual activities considered to be unsafe proved to cover a wide variety of practices.

Participants' responses indicated that they did not practice unsafe sex out of ignorance in any simple sense. Rather, they offered detailed accounts of their experiences of unsafe sex in which they assessed the safety of particular activities while explaining their preferences and choices. Five recurring themes emerged from the interviews that highlighted specific obstacles to practicing safe sex as well as the issues our participants considered in sexual encounters. First, sexual decision-making in couple relationships was often complicated by the men's HIV status and by issues of monogamy, trust, and intimacy. Second, unsafe sexual activity was sometimes described as being inadvertent or involuntary. Third, negative self-images and moods appeared to be predictive of a willingness to have unsafe sex. Fourth, by reading signs that a particular partner was "likely" to be HIV-negative, strategies of intuiting safety often led to unsafe sex. Finally, unclear boundaries between safe and unsafe activities complicated the decision-making process. The remainder of the present report is organized around these five themes.

Couple Relationships

Twenty-three participants reported having unsafe sex in the context of a continuing relationship. Most of these involved some form of agreement (either tacit or verbalized) between two HIV-negative men. Twelve of the 23 had arrived at an explicit agreement with their partner about permitting unsafe sex inside the relationship. Following Kippax et al. (1993), we consider these to be instances of negotiated safety. Most couples agreed on a safe-sex-only policy with casual partners; a few agreed to refrain from having other sexual partners altogether. These kinds of agreement about sex with others outside the relationship allow two HIV-negative men to have unprotected yet safe sex with each other. One participant remarked:

   We chose about two years ago, I'm guessing, that sex between the two of us
   would be unprotected. It was something of a long agonizing decision.
   Especially with all the friends who told us, "You can't possibly do that."
   To answer your question, it came to be that way by mutual agreement and a
   longstanding agreement as well, the agreement being that sex between us is
   unprotected, but we absolutely trust each other to have protected sex with
   anyone else. We also know we're both HIV-negative. (Gay, 24, HIV-negative,
   clerk)

Others made a point of incorporating HIV-testing into these agreements. For example:

   We initiated the relationship with the feeling that we be only safe until
   we both got tested. And we were tested again Tuesday.... I got my results
   back and they were negative for the second time. So then we had unprotected
   sex for the first time on Tuesday. That was different. It was very special.
   (Gay, 31, HIV-negative, clerk)

Explicit discussion of outside relationships combined with knowledge of both partners' HIV status can create a situation in which unprotected sex within a relationship is completely safe. The question of trust is always an issue, however, because the mutual safety of the partners depends on consistent adherence to the terms of the arrangement (i.e., no unsafe encounters with other partners).

Negotiating safety can be particularly complex when partners do not share the same HIV status. The notion of an on-going and shared future is disrupted when one of the partners has a life-threatening disease that can be transmitted to the other partner (Adam & Sears, 1996). Nonetheless, people tend to be relatively unconcerned about contracting HIV (Pilkington, Kern, & Indest, 1994) or germs in general (Nemeroff, 1995) from a close partner. One of our HIV-positive participants described a pattern of inconsistent condom use with an HIV-negative partner:

   He is not HIV-positive despite our histories. I am, and we don't use any
   form of condom, protection.... We do discuss it from time to time because
   generally I guess I am more the bottom than the top in the relationship,
   but I like to fuck too, and every once in a while he wants to get fucked,
   and I do. Usually I don't cum, if I am fucking him, if I want to cum then I
   usually will use a condom. But the number of times that we have fucked and
   cum without a condom is more frequent than not. (Gay, 33, HIV-positive,
   unemployed)

Hence, condom use is one of a series of interconnected considerations that guide sexual practices within a relationship. The last quote also highlights a distinction many participants made between the insertive and receptive roles in anal sex, and between anal sex with and without ejaculation. Both of these issues are discussed further below.

Other men noted that unsafe sex was more likely to occur within an ongoing relationship because it was viewed as a means of expressing or maintaining a feeling of intimacy or romance. In practice, the process of negotiating safety can be complicated by underlying and often implicit presumptions about monogamy and fidelity. As Sobo (1995) noted in her study of African American women, the discourse of monogamy may work to prevent the adoption of safe sex practices. Similar findings were reported among a sample of women over 30 years of age from the United Kingdom (Maxwell & Boyle, 1995). The monogamy script was less predominant among the men we interviewed than among these samples of women. In relationships in which it arose, however, it sometimes served as an impediment to practicing safe sex. Unsafe sex appears to be read as a primary sign of the special trust that partners in a couple have for each other (Afifi, 1999; Ames, Atchinson, & Rose, 1995; Bartos, 1994; Bartos, McLeod, & Nott, 1993; Flowers, Smith, Sheeran, & Beail, 1997; Hospers, Molenaar, & Kok, 1994; Silvestre, Lyter, Valdiserri, Huggins, & Rinaldo, 1989). Such symbolism can inhibit partners from adopting protection, which can be interpreted as an accusation of infidelity. Indeed, one participant explained the lack of condom use in his relationship this way: "No, he didn't want to. For him, if we practiced safe sex, he would be admitting to going outside [the relationship]" (Gay, 48, HIV-negative, unemployed; translated from French). Seven of the men we interviewed spoke of their relationships in these terms; an additional four spoke with some regret of having been caught up in a set of assumptions that led them to practice unprotected sex with a man who was later found to be having unsafe sex with other partners. One participant spoke directly about the trade-offs he experienced in thinking about introducing condoms into his relationship:

   Once I suspected he was being unfaithful to me, it was very difficult to
   say, "OK, from now on I want you to wear a condom," without mining what was
   at that point a really shaky relationship. And I wanted him. There was
   nothing worse than being gay and gray.... And since I'm 54 years old, quite
   frankly I wanted him around. (Bisexual, 54, HIV-negative, city employee)

Negotiations about safety may be precluded when questions of trust are embedded in a romantic discourse. For example, when sexual monogamy is equated with trust, insistence on using condoms can symbolize distrust. Obviously, a great deal of trust is required in the present historical context for gay-male couples to forsake using condoms, particularly when traditional concepts of romance make no provision for the confession of sexual activity outside of a relationship. Moreover, because trust can be subject to deterioration, desire and romance can provide unwarranted reassurance about presumptions of monogamy.

In general, our participants' responses confirm that reconciling safe sex with the emotional dynamics of a relationship is too complicated to be portrayed simply as a gap in knowledge or information. Moreover, instances of negotiated safety serve as an important reminder that one cannot simply categorize all men who report having unprotected anal intercourse (on a questionnaire, in an interview, etc.) as having engaged in unsafe sex. It is noteworthy that unprotected male-female sex in the context of a committed relationship (marriage or otherwise) is rarely considered unsafe.

Inadvertent or Involuntary Unsafe Sex.

Sixteen participants described instances of unintended unsafe sexual encounters, or cases in which such encounters were clearly the exception rather than the norm. In these cases, unsafe sex was understood in retrospect to be an accident. Two men attributed unsafe sex to intense passion that began during semi-conscious sleep. One of these remarked:

   We got to doing our thing, but before either one of us climaxed, we took a
   break and had a little snooze.... And I decided, I guess, in my semi-sleep
   state that I was going to pull off the condom that I had on. Ten or 15
   minutes later, we were back to getting screwed again. We never stopped the
   second time to put back on the condom.... The second time around it was
   like a dog in heat. My mind was thinking it. My mind knew that I shouldn't
   have been having sex with this guy not wearing a condom but it was not
   enough to control the heat instincts. (Gay, 38, HIV-negative, musician)

Twelve men described a surprise moment of unsafe intercourse in the midst of sexual activity that involved verbal and nonverbal gestures. For example, one man said:

   I just blurted out that I'm HIV-negative and the next thing I know he's
   sitting on it. I mean I haven't done that for a long time. Two years. It
   was good. I mean, no fluids were shared between us. I mean we did have
   unsafe sex a little bit.... Then he put the condom on and we had safe sex
   so it wasn't totally unsafe. (Gay, 24, HIV-negative, waiter)

These incidents were described as momentary, nonorgasmic breaks in sexual encounters that also included the use of condoms.

Five participants described unsafe sex in the context of first-time sexual encounters. Initial sexual experiences are often unplanned, taking a spontaneous and covert form because of the potential dangers of discovery. Social taboos against talking about sex, particularly in the case of younger people and same-sex encounters, also mean that sexual initiations often take place in a context where there is little knowledge or access to information. One participant described his first gay encounter with a 38-year-old when he was 21:

   It was the first time I ever actually had real intercourse, where I
   actually had gay intercourse. It was just incredible.... I made good
   friends with one guy.... He took me to his place and we had unsafe sex--all
   the way. It was incredible.... There was some kind of trust between us
   already built up, I guess.... I guess he realized that I'd never been with
   anyone else except for a couple of females and they were OK. (Gay, 31,
   HIV-negative, professional)

This participant's account drew on two other narrative lines that entered repeatedly into accounts of unsafe sex: (a) a gay/straight distinction that reproduces the widely-held view that heterosexual sex is a safe practice, and (b) an interpretive process of reading a complex set of physical and character signs to determine a partner's HIV status. Both of these themes are treated more extensively below.

In some cases, men attributed their involvement in unsafe sex directly to willful acts on the part of their partners. For example, three participants cited instances of engaging in receptive anal sex believing that their partner was wearing a condom only to discover later that he was not. One of these men made the following observation: "Where the unsafe sex came in is he was originally wearing a condom and half way through the sex he took it off and continued and he was in a top position" (Gay, 39, HIV-positive, travel agent).

Two participants attributed unsafe sex to sexual assault. In one of these cases, a young man (who was a heterosexual, fundamentalist Christian at the time) had a first-time sexual experience with a man that he characterizes as a "rape"; nevertheless, he sought out this same man for his second and third experiences as well. He noted:

   I was only 16 at the time. So we went back to his apartment [from an
   apartment building gym] and he just started completely working my
   sensations, kissing me for the first time, touching me. It was the first
   time I had ever gotten a blow job. He pinned me down and raped me, and then
   after he got done and got off inside my ass--and I guess he must have hit
   my prostate because I got off without him actually touching me, or me
   touching myself.... And that was interesting and there were two other times
   while I was visiting my girlfriend during that week that I went back to the
   same guy. It was just the spontaneity of the thing. (Gay, 21, HIV-negative,
   waiter)

The other interviewee described an nonconsensual act of anal sex that occurred at a warehouse party for leather men. Only one participant described an act of accidental unsafe sex due to a condom breaking.

When reflecting upon their participation in unsafe sex, some of our participants viewed their involvement as involuntary, inadvertent, or otherwise beyond their control. Regardless of whether these encounters were "truly" accidental, in most cases the men were referring to relatively isolated incidents rather than regular practices.

Negative Self-Images and Moods

Seventeen participants explained unsafe sexual experiences in terms of a heightened orientation to risk-taking that was often associated with depression or a negative mood state. Although risk-taking can be an important part of sexual pleasure, it can cloud or come into direct conflict with considerations of safety. Five participants noted explicitly that they were using sex as a way of escaping from negative feelings. For example, one man referred to sex as "animal" and "wild" but later talked about his need to relieve feelings of depression:

   This was the first time in my life where I, like, with somebody had really
   let go and we were really just animal about each other. It was just
   wild.... [Unsafe sex has happened] a couple of times in the bathhouse
   because I've felt very lonely and very sad, really needed affection, and
   allowed myself to open myself up so much so that I just went beyond my
   boundaries and my own limitations. (Sexual, 37, HIV-negative, artist).

This quote highlights the gulf between a sense of responsibility that is connected with safe sex practices, and feelings of abandon that can accompany sex when it is used as a means of escape. Other researchers have also identified a positive association between unsafe sex and using sex as an escape from everyday life (Pollak, 1988), and between sexual risk-taking and depression (Hospers et al., 1994).

Unsafe sex may become particularly likely when drugs and alcohol are used to heighten a sense of abandon. One participant noted:

   About that time last July when I had that [unsafe encounter] and the
   sensation was just incredibly much--incredibly richer and all that and--but
   I know at the same time that if I were not even sober, but less blitzed
   than I was, then I would not have given myself into that. It would have
   felt far too self-conscious and wrong and all of that. Apart from times
   when I have essentially chosen to disable my judgment (laughs).... I am not
   able in my experience to start having--to start fucking unsafely. (Gay, 31,
   HIV-negative, student)

It may be simplistic, however, to identify drug or alcohol use as a cause of unsafe sex. Physiological effects of drugs require interpretation, and drug vernacular provides a rhetoric of exculpation. For example, a cross-cultural study of alcohol consumption found that people become intoxicated in culturally expected ways (McAndrew & Edgerton, 1969). Indeed, the respondent quoted above was candid about having "chosen to disable [his] judgment" rather than blaming drug use for his actions. Another man made it clear that he did not think that drug use influenced his ability to behave rationally:

   Any time I use drugs they just make me more friendly, happier, mellower. I
   get sort of a stronger surge from them to do something more aggressive,
   that's true, but I don't think it hinders my rational decision making. I
   think the only thing drugs do to me is inhibit all my inhibitions. (Gay,
   33, HIV-negative, researcher)

The notion that sexual activity involves surrender to the "heat of the moment" is part of a broader understanding of sex as ecstatic (Bartos et al., 1993; Diaz, 1997). This perspective requires that sexual activity be outside of the ordinary, a break from everyday obligations and commitments. One participant (cited above) mentioned that "the heat instincts" during a particular encounter were stronger than his will to use a condom. Another participant viewed it this way: "I guess I simply enjoy sex and the pleasure and the excitement of the moment should be more important than [safety] to me" (Gay, 68, HIV-negative, professional).

Such escapes from the mundane to the ecstatic can be shaded with self-negation, representing a flight from oneself. On the basis of observations from his practice in clinical psychology, Odets (1995) reported that unsafe sex was linked to covert or overt self-destructive intent. Five of our participants discussed sexual experiences in which risk-taking was difficult to distinguish from overt self-destructiveness. One described his participation in unsafe sex as a self-destructive quest at one point, and as a fatalistic resignation to destiny at another:

   I am one of those stupid, even ridiculous people who favors safe sex, who
   promotes safe sex, teaches safe sex to everybody, who always has a condom
   on himself, but who never uses it, for two reasons.... The first is that
   for quite a while I exposed myself to the dangers of unprotected encounters
   in order to try to catch AIDS. It was a disguised form of suicide. The
   second reason was that I considered that it was one of those things that
   happens, an occupational hazard. If I was going to get it, I would get it.
   If I didn't, I didn't. (Gay, 36, HIV-negative, social services; translated
   from French)

The same participant stated later that he was abused as a child, that he was a recovering alcoholic and addict, and that he believed that getting drunk or high is related to unsafe sex because one loses a sense a danger. For this man, a troubled life resulting in a general state of unhappiness or depression appears to be associated with self-destructive tendencies in general, and with engaging in unsafe sex in particular.

Three respondents who described conservative religious upbringings reported difficulties adopting safe-sex practices that incorporated elements of self-destructive intent. These men were raised to think of nonmarital sexual activity as a transgression. As such, they had difficulty reconciling the idea of safety with that of transgression. Having "fallen into temptation" through expression of their homosexual desire, they found themselves caught up in the "AIDS as punishment" discourse propagated by conservative churches, which, ironically, can become a self-fulfilling prophecy by increasing risk of exposure to HIV through unsafe sex. The following participant traced a period of unsafe sex in his life to his struggle over acknowledging his homosexuality while he was a member of a fundamentalist Christian denomination:

   I developed maybe a fatalistic attitude, and either I'm going to get it or
   I'm not. And if I get it, well, then I'm out of here. Like I'm being
   punished, and you have to remember too I still believed that very, very
   strongly that what I was doing was wrong, and well, if I got it, well, God
   was just punishing me and that was it. (Gay, 41, HIV-negative, nurse)

Another man drew a link between his religious upbringing and his approach to sex as an unconstrained plunge into ecstasy. He contrasted his own sexual behavior with that of his friends, whom he described as having a more self-controlled sexual style. He also drew an explicit link between his conservative religious upbringing and troubles dealing with his sexuality:

   For those individuals who grow up in a strong religious environment, they
   have a problem accepting their sexual lifestyle as being part of their
   lifestyle or their existence. They have a strong religious sense. They
   always want to go, "OK, God's going to punish you. He's going to punish you
   tomorrow for this, whatever you did last night. There's going to be major
   punishment. He'll be like raining brimstone and fire on you." And you have
   a guilty conscious in a strong way. (Gay, 33, HIV-negative, researcher)

Thus, for this man, safe sex may be irrelevant when divine retribution is impending. In some cases, HIV-positive people who have been immersed in conservative religious traditions continue to struggle with this punitive logic after their diagnosis (Adam & Sears, 1996).

In certain situations, then, depression or simply a heightened sense of sexual risk-taking can work against insistence on safe sex. Nonetheless, obstacles in the progression from knowledge to attitudes and behavior often include very rational considerations of the risks and pleasures involved.

Intuiting Safety

Twelve participants described attempts to read implicit signs of a partner's HIV status as a method of ascertaining safety, a process that has been observed among young heterosexual men and women in the U.S. (Williams et al., 1992) and Canada (Maticka-Tyndale, 1992), and among gay men in studies from several countries (Ames et al., 1995; Aveline, 1995; Gold et al., 1994). This strategy involves a shift in emphasis from the sexual act to the sexual partner--or from the universal to the particular--by evaluating a partner's HIV status from various cues. In the following excerpt, the participant reflects on the likelihood that his partner was HIV-positive and on the risk of the particular act. Specifically, he intuits a sense of trust and safety from his partner's youth and by adopting the insertive role in anal sex:

   About a month and a half ago, I fucked a guy without using a condom so that
   is unsafe sex by general standards. I don't feel it was unsafe because I
   don't feel that I am HIV-positive and I don't think he is.

   How do you make that judgment?

   Not in very good ways. Just because he is young and relatively
   inexperienced and just in the fact that I am fucking him. (Gay, 30,
   HIV-negative, service industry).

At times, HIV-negative status may be inferred from connections with heterosexuality. For example, the next participant expressed a certain embarrassment in confronting his fantasy logic, although his candor allowed a glimpse into sexual decision making that was not unusual:

   The guy fit within a certain kind of fantasy of mine that made me want to,
   and also I felt this--stupid reason--because he fit into this certain
   criterion, I felt he was a safer candidate to have unsafe sex with.... He
   gave the illusion of being straight--his haircut, his body type, his
   attitude, what he would or wouldn't do. (Gay, 39, HIV-negative, artist)

Another participant (quoted earlier) noted that because his sexual history included only "a couple of females," his partner had intuited he was HIV-negative. Thus, the idea that heterosexuals are safe while homosexual men are not safe can be adopted by gay men as well as the heterosexual mainstream (Boulton et al., 1995). This notion sets up a self-negating dynamic, such that unsafe sex is justified simply because it is (or seems) heterosexual, which ultimately assures that it is not safe.

Each of the six participants who reported recent sexual activity with women as well as men adopted some form of this perspective. For example, one man reported that he had unsafe sex with women but not with men:

   With guys, no [unsafe sex], with women, yes. I've had unsafe sex with women
   like so many times. A few years ago I had no idea I could get a girl
   pregnant. It was really stupid at the time. (Omnisexual, 19, HIV-negative,
   odd jobs)

Boundaries Separating Safe from Unsafe

Safe-sex decision making becomes particularly complex when prevention messages are contradictory or when the epidemiology itself is unclear. Gay men often have a rich unofficial knowledge of HIV-associated risks, which was developed and shared through largely informal processes. Pioneering efforts in safe-sex education generally wove together this unofficial knowledge with the best scientific knowledge of the time (see Watney, 1989). Areas of ambiguity for lay people are often direct reflections of areas of contention in "official" sources (e.g., epidemiology, medicine, or public health; see Kippax et al., 1990; Myers et al., 1993). Indeed, awareness of gray areas in the demarcation of safe from unsafe may be better interpreted as a sign of sophistication rather than one of ignorance.

Boundaries between what participants considered to be safe and unsafe were particularly ambiguous in four instances: unprotected sex in couple relationships, being the insertive partner in anal sex, anal intercourse without ejaculation, and fellatio with ejaculation. The ambiguity of unprotected sex between couples who are both HIV-negative (discussed above) stems from problems over how practices inside the relationship relate to outside activities. Remarks about unprotected anal sex from two participants cited above indicate that the receptive role was viewed as considerably more risky than the insertive role. Whereas one man noted that the sex was less risky "just in the fact that I am fucking him," another noted his heightened risk because his partner "was in a top position." These lay-person observations are particularly interesting when one considers that virtually no safe-sex education program makes a distinction in risk between insertive and receptive roles. Although a recent epidemiological report (McClure & Grubb, 1999) acknowledges that more cases have been attributed to the receptive role, it also classifies both roles identically as high risk. On the one hand, homosexually active men may have greater levels of expertise than they are "supposed" to have, with some--perhaps most--being aware of fine nuances in distinctions on the basis of risk. On the other hand, knowledge that the insertive role carries reduced levels of risk may provide little comfort when the reduction remains unspecified. In the worst case, men may place themselves at risk by judging the insertive role to be less risky than it actually is.

The practice of anal intercourse without ejaculation raised additional questions for our participants. Although they recognized that it is risky--as do the experts (McClure & Grubb, 1999)--some felt that the actual level of risk was ambiguous. The ambiguity here appeared to depend primarily on the success of the insertive partner in withdrawing before ejaculation. Five interviewees described unsafe sexual encounters that were interrupted before orgasm. One (cited above) noted that "no fluids were shared" as a way of explaining that his risk was reduced from what it would have been otherwise. Another man described his encounter with a police officer:

   We started getting into it and he fucked me but he didn't come but I mean
   afterwards I was completely freaked out because I thought, "Oh my god, oh
   my god, what did I just do?" When you get caught up in the heat of the
   moment and before, and before I knew he was going to come, I kind of pushed
   him off because I knew that was going to happen, and then he ended up
   coming on my chest which is safe enough. (Gay, 31, HIV-negative,
   professional)

Participants appeared to be using a common-sense assumption that "less semen means less risk" in order to modify their assessment of the risks of unprotected anal sex. The assumption in these instances may be providing false assurances, however, because of the "impossibility of determining how much preejaculatory fluid has been deposited in the rectum" and the "efficiency of transmission through penile-anal intercourse" (McClure & Grubb, 1999, p. 28).

Seven participants ruminated over the risks of oral sex, Kippax et al. (1990) found this to be an area of ambiguity among Australian gay men in research conducted as early as 1987. This particular ambiguity is likely exacerbated by national differences in education and prevention programs. Whereas official AIDS-education sources in the U.S. have maintained that unprotected oral sex poses risks similar to unprotected anal sex, Canadian authorities classify insertive oral sex as negligible risk, receptive oral sex as low risk, and insertive and receptive anal sex as high risk (McClure & Grubb, 1999). All of the men we interviewed lived in areas that are proximate to the U.S., with regular access to American media (TV, radio, magazines). Moreover, many of the men from Windsor were likely to be regular participants in Detroit gay life (only the Detroit River separates Windsor from Detroit), with frequent exposure to American safe-sex messages (e.g., pamphlets at bars).

Nonetheless, our interviews revealed a widespread consensus among Canadian gay and bisexual men that fellatio without orgasm need not be abandoned to qualify for safe sex. It is interesting to note, however, that official Canadian guidelines (McClure & Grubb, 1999) do not distinguish between fellatio with or without orgasm. For example, a recent poster from the AIDS Committee of Toronto refers to the "ABCs of oral sex and HIV: A--It, B--is, C--low risk," with no reference to ejaculation. The question of swallowing semen during oral sex remains fraught, however, and of concern to gay men around the world (Hospers et al., 1994). Like all of our participants who worried about oral sex, the next man expressed concern about ingestion of semen:

   In a bathhouse someone ejaculated in my mouth but generally I try to avoid
   that, but when that happens, I spit immediately. But in that case, the
   position was such that he came so deeply in my mouth that I practically
   swallowed it all.... I was more angry with myself perhaps for not having
   taken the precaution of saying to him, "Don't come in my mouth." I should
   have said that to him. (Gay/formerly married, 53, HIV-negative, civil
   servant; translated from French)

Another man made the following observation:

   So I think I ended up having oral sex with four or five guys, but I don't
   remember exactly, I don't know if I swallowed or not....

   You would consider swallowing to be unsafe?

   To an extent, yeah, it is, for me, I don't do it often. It's just when
   there's a lot more [semen] involved, you're sure [HIV-infection] would
   happen. (Gay, 33, HIV-negative, researcher)

Some men established practices that made them feel more secure about oral sex. For example, a few participants described routines that combined not swallowing with other hygienic practices. One remarked:

   I didn't swallow it, if that is what you mean by safe. It was as far as I
   am concerned. Well, the fact that I did not swallow it, and I rinsed right
   after. I went over to the sink, and washed my hands and rinsed my mouth,
   and left. (Gay, 31, HIV-positive, cabinet maker)

Again, men appear to rely on a combination of official knowledge (oral sex is low risk, at least in Canada) and unofficial knowledge (less semen = less risk) in devising such routines, particularly when the scientific literature is ambiguous. This ambiguity is illustrated in the Canadian guidelines, which classify receptive fellatio with or without taking semen in the mouth as low risk, yet specify that the risk "can be reduced by avoiding ejaculation of semen in the mouth" (McClure & Grubb, 1999, p. 24).

A final quote from a participant who was confused about the risk of anilingus is particularly illuminating: "I don't know what the official stand is on anilingus, you know, rimming and that" (Gay, 34, HIV-negative, teacher). Canadian authorities consider anilingus to have negligible risk for transmitting HIV (McClure & Grubb, 1999). Regardless, it is the mention of an "official stand" that is interesting here. The term combines the idea of authority with a sense that expertise exists at a level distinct from practical, everyday solutions. Indeed, in at least four different sexual contexts (sex within a committed relationship, insertive anal sex, anal sex without ejaculation, and oral sex with ejaculation), participants appear to incorporate official knowledge with unofficial knowledge to arrive at a practical solution. In some cases, the resulting solution may be overly safe. In others, it may not be safe enough.

CONCLUSIONS AND IMPLICATIONS

We interviewed homosexually active men in order to examine the factors and considerations that influence whether men engage in unsafe sexual practices. Interviewees' responses confirmed that factors other than lack of knowledge play a central role in deciding to have unsafe sex and rationalizing about it afterward. Indeed, our participants exhibited levels of AIDS knowledge that appeared to be remarkably sophisticated and to extend well beyond the information provided in most HIV-prevention campaigns. Our findings suggest that HIV-prevention programs might benefit from engaging with the rich understandings of sexual risk and safety that many men have already developed. Such an approach would need to address directly the issues at the forefront of the minds of gay and bisexual men, such as avoiding ejaculation during anal and oral sex, the distinction between the insertive and receptive roles in anal sex, and whether condoms are needed in couple relationships.

Although safe-sex campaigns often convey universal messages ("Use a condom every time"), such a straightforward and simplistic approach may not fit well with the emotional complexities that influence the sexual practices of men with regular partners. Many couples have worked out unique and creative ways to make sex safe in the context of an ongoing relationship. Such strategies may be specific to a particular couple or may remain problematic to varying degrees. Nonetheless, for prevention programs to be effective, they may need to address these individual strategies head-on, highlighting what works and what does not. In short, for gay male HIV-prevention programs to have continuing efficacy and relevance, they may be compelled to engage with the issues that are at the forefront of the population they intend to serve. In Australia (see Kippax et al., 1993; Rotello, 1995) and Canada (Maxwell, 1997), specific education campaigns have been designed around the negotiation of safety in couple relationships. These campaigns include discussion of abandoning condoms between two HIV-negative men while making explicit the need for clear communication and trust between partners.

Some of our participants attempted to make their sexual lives compatible with the desire to communicate trust in couple relationships. Such attempts met with varying degrees of success. The degree to which condom use can be equated with the message "I don't trust you" appears to be directly associated with the degree to which the abandonment of condoms is interpreted as a sign of fidelity. This logic was conveyed on billboards in the streets of Geneva during the 12th World AIDS Conference which read (roughly translating from French), "Not faithful in bed, faithful to condoms." The implicit logic of this advertisement may be self-negating because it identifies safe sex as a practice of morally dubious people. Male homosexual couples may be less susceptible to this logic than their heterosexual counterparts, however, because they are less likely to treat monogamy as an essential sign of love or trustworthiness.

Our data suggest that safe-sex education programs may benefit by addressing the ambiguous risks of particular sex acts. Several of our participants' accounts took the form of retrospective inquiries about the safety of a particular sexual interaction. In these cases, participants attempted to apply scientific knowledge of what constitutes "safety" to their own experiences and were sometimes left in doubt in the end. Other participants relied on a variety of discourses about determining safety that, at times, intersected and diverged from the discourses propounded by public-health authorities and community-based AIDS organizations. Many of these reports of unsafe sex revealed the instability and permeability of the boundaries that separate safe from unsafe. Moreover, participants' responses highlighted the limitations of prevention strategies that ignore the emotional and communicative value of sexual expression. Unsafe practices are not so much a question of a lack of knowledge or the disablement of reason, but the result of a complex consideration of scientific and nonscientific factors that will need to be addressed if HIV-transmission is to be affected.

In some cases, ill-advised prevention strategies that have little relevance to gay men may actually end up promoting unsafe sex, albeit inadvertently. For example, the "know your partner" advice propagated by public-health authorities---especially common in the early days of the epidemic--allows people to exempt themselves from the need for safe sex by "reading signs" of their partner's putative safety (Ames et al., 1995). Indeed, our data show that some homosexually-active men attempt to intuit their partner's HIV status on the basis of unreliable signs (e.g., age and heterosexuality). A related study of U.S. college students (Misovich, Fisher, & Fisher, 1996), reported that it was common for students to believe that knowing a partner well made safe sex unnecessary. Moreover, students who held this belief were significantly less likely than other students to adopt safe-sex practices. The differences between our sample and the one examined by Misovich et al. (i.e., Canadian vs. American, homosexually-active men vs. heterosexual men and women, a wide range of ages with a mean of 35 vs. a narrow range with a mean of 20), as well as the difference between methods (qualitative interviews vs. quantitative questionnaires) imply that this finding may generalize widely. In other words, "know your partner" strategies may be relatively counter-productive across populations.

The relatively sophisticated and nuanced levels of knowledge exhibited by our participants raise additional questions about the efficacy of prevention programs that appear to assume widespread ignorance among their target populations. From our perspective, health-education programs should attempt to engage with existing cultures and stocks of knowledge rather than intervening with a vacuum of "ignorance." Some American programs are particularly curious in this regard, having deviated markedly throughout the history of the epidemic from those in Canada, Australia, and Europe regarding the risks of oral sex. For example, Gay Men's Health Crisis (the largest and oldest AIDS organization in the U.S.) currently classifies oral sex as high-risk behavior, albeit with a caveat noting that oral sex is less risky than anal sex (Gay Men's Health Crisis, 1999a). They also note that "the biggest risk from oral sex is getting cum in the mouth. The only way to be totally safe is to lick only the shaft or to use a condom" (Gay Men's Health Crisis, 1999b). When American gay men are asked about sex with casual partners, however, most endorse the use of condoms for anal sex but not for oral sex (e.g., Herek & Glunt, 1995). In other words, the behavior and attitudes of American gay and bisexual men are similar to those of our Canadian participants. Moreover, many American researchers who study the sexual behavior of gay and bisexual men classify unsafe sex simply as "unprotected anal sex" (e.g., Kelly et al., 1991, 1995; Peterson et al., 1992), which implies that any act of oral sex is relatively safe. In short, education programs that are inconsistent with the views of experts and the opinions and practices of their target audience may need to be rethought.

Participants' responses highlighted the dangers of the bedrock semiotic binary of our culture, which equates homosexuality with sexual risk and heterosexuality with safety. This equation influences gay and bisexual men in their decisions about whether to adopt or to drop safe sex. Men who are deemed to be "straight" according to superficial signs are sometimes treated as if they pose no risk for HIV infection. Thus, cultural identification of AIDS with homosexuality contributes to increasing unsafe sex among men, apart from generating mass complacency among heterosexual men and women.

For some men who have sex with men, personal crises--whether extraneous to sexuality (such as job loss or depression) or intrinsic (coming out under difficult circumstances)--appear to be associated with unsafe sex and other risky activities. In particular, the coming-out processes of married men and conservative, religious men may involve considerable personal turmoil that is associated with unsafe practices. Although conventional public-health strategies typically hold individuals responsible for preventing HIV transmission, our interviews showed the harmful effects on some men of homophobic ideologies that can lead to personal conflicts about using condoms during sex.

When describing instances of unsafe sex, none of our respondents mentioned rationales that have attracted a good deal of media attention, political debate, and research funding, namely (a) that new drug combinations have made gay men complacent about safe sex, (b) that gay men are succumbing to an ideology of "bare backing" as a result of AIDS fatigue or as a backlash against safe sex, or (c) that the unreliability of condoms jeopardizes safe sex. Some of these perceived problems may be more conjecture than fact. Although much less attention is being targeted toward the complex problems associated with the emotional aspects of sexual relationships and the sometimes ambiguous boundaries between safe and unsafe activity, our data imply that these types of problems influence the decision-making processes that surround safe sex.

Ultimately, gay and bisexual men need to have the opportunity to think carefully about the issues raised in these findings and those reported in similar studies. Such a process could be facilitated by prevention programs that deal in a straightforward manner with issues that are of greatest concern to homosexually active men. When populations are relatively well informed about HIV and AIDS, a shift in emphasis from knowledge of AIDS to the complexities of sexual practice may help people navigate through the snares and pitfalls that lead to exposure to HIV. It is particularly important to abandon the assumption that a single "correct" strategy can deal with multiple difficult issues, which are often idiosyncratic and context-specific. As it stands now, men who have sex with men sometimes exempt themselves from risk categories on the basis of moralistic or otherwise ill-advised judgements about appropriate sexual comportment. Ironically, those who exempt themselves from risk often engage in practices that place them at risk. Exposure of these discursive tangles that organize everyday life merits a place in HIV-prevention programs.

(1) We are aware of the shift in terminology from safe to safer adopted by some researchers and educators. We chose to use "safe" because we were interested in the construction of safety made by our participants, most of whom referred to "safe" sex.

(2) Preliminary quantitative analyses of the sexual- behaviors and demographics questionnaires revealed that recent participation in unprotected anal sex (i.e., in the 6 months prior to the interview) was not associated with respondents' age, income, occupation, or cultural background.

(3) The interviewers did not define unsafe for the participants. In fact, many participants had definitions that diverged from standards used by AIDS researchers and organizations.

REFERENCES

Adam, B. D. (1992). Sociology and people living with AIDS. In J. Huber & B. E. Schneider (Eds.), The social context of AIDS (pp. 3-18). Newbury Park, CA: Sage

Adam, B. D., & Sears, A. (1996). Experiencing HIV: Personal, family, and work relationships. New York: Columbia University Press.

Adib, S. M., Joseph, J. G., Ostrow, D. G., & James, S. A. (1991). Predictors of relapse in sexual practices among homosexual men. AIDS Education and Prevention, 3, 293-304.

Afifi, W. A. (1999). Harming the ones we love: Relational attachment and perceived consequences as predictors of safe-sex behavior. The Journal of Sex Research, 36, 198-206.

Altman, D. (1994). Power and community: Organizational and cultural responses to AIDS. London: Taylor & Francis.

Ames, L., Atchinson, A., & Rose, D. T. (1995). Love, lust, and fear. Journal of Homosexuality, 30, 53-73.

Aveline, D. T. (1995). A typology of perceived HIV/AIDS risk-reduction strategies used by men who "cruise" other men for anonymous sex. The Journal of Sex Research, 32, 223-234.

Bartos, M. (1994). Community vs population. In P. Aggleton, P. Davies, & G. Hart (Eds.), AIDS: Foundations for the future (pp. 81-96). London: Taylor & Francis.

Bartos, M., McLeod, J., & Nott, P. (1993). Meanings of sex between men. Canberra, Australia: Australian Government.

Bolton, R., Vincke, J., Mak, R., & Dennehy, E. (1992). Alcohol and risky sex: In search of an elusive connection. Medical Anthropology, 14, 323-363.

Boulton, M., McLean, J., Fitzpatrick, R., & Hart, G. (1995). Gay men's accounts of unsafe sex. AIDS Care, 7, 619-630.

Carballo-Dieguez, A., Remien, R. H., Dolezal, C., & Wagner, G. (1999). Reliability of sexual behavior self-reports in male couples of discordant HIV status. The Journal of Sex Research, 36, 152-158.

Catania, J. A., Kegeles, S. M., & Coates, T. J. (1990). Towards an understanding of risk behavior: An AIDS risk reduction model (ARRM). Health Education Quarterly, 17, 53-72.

Connell, R. W., Davis, M. D., & Dowsett, G. W. (1993). A bastard of a life. Australian and New Zealand Journal of Sociology, 29, 112-135.

Diaz, R. M. (1997). Latino gay men and psycho-cultural barriers to AIDS prevention. In M. P. Levine, P. M. Nardi, & J. M. Gagnon (Eds.), In changing times: Gay men and lesbians encounter HIV/AIDS (pp. 222-244). Chicago: University of Chicago Press.

Diaz, R. M. (1998). Latino gay men and HIV: Culture, sexuality, and risk behavior. New York: Routledge.

Donovan, C., Mearns, C., McEwan, R., & Sugden, N. (1994). A review of the HIV-related sexual behavior of gay men and men who have sex with men. AIDS Care, 6, 605-617.

Dowsett, G. W. (1990). Reaching men who have sex with men in Australia: An overview of AIDS education, community intervention and community attachment strategies. Australian Journal of Social Issues, 25, 186-198.

Dowsett, G. W. (1996). Practicing desire: Homosexual sex in the era of AIDS. Stanford, CA: Stanford University Press.

Ekstrand, M. L., & Coates, T J. (1990). Maintenance of safer sexual behaviors and predictors of risky sex: The San Francisco men's health study. American Journal of Public Health, 80, 973-977.

Fisher, J. D., Fisher, W. A., Williams, S. S., & Malloy, T. E. (1994). Empirical tests of an information-motivation-behavioral skills model of AIDS-preventive behavior with gay men and heterosexual university students. Health Psychology, 13, 238-250.

Flowers, P., Sheeran, P., Beail, N., & Smith, J. A. (1997). The role of psychosocial factors in HIV risk-reduction among gay and bisexual men: A quantitative review. Psychology and Health, 12, 197-230.

Flowers, P., Smith, J. A., Sheeran, P., & Beail, N. (1997). Health and romance. British Journal of Health Psychology, 2, 73-86.

Gaies, L. A., Sacco, W. P., & Becker, J. A. (1995). Cognitions of gay and bisexual men in sexual situations: Development of the Sex and AIDS Thought Scale (SATS). AIDS Education and Prevention, 7, 513-522.

Gay Men's Health Crisis (1999a). HIV & AIDS: The basics. Retrieved June 14, 1999 from the World Wide Web: http://www.gmhc.org/stopping /basics.html.

Gay Men's Health Crisis (1999b). Safer sex guidelines for gay men. Retrieved June 14, 1999 from the World Wide Web: http://www.gmhc. org/stopping/disclaimer/gaymen.html.

Glaser, B. G., & Strauss, A. L. (1967). The discovery of grounded theory: Strategies for qualitative research. Chicago: Aldine.

Gold, R. S., Skinner, M. J., & Ross, M. W. (1994). Unprotected anal intercourse in HIV-infected and non-HIV-infected gay men. The Journal of Sex Research, 31, 59-77.

Gruer, L. D., & Ssembatya-Lule, G. (1993). Sexual behavior and use of the condom by men attending gay bars and clubs in Glasgow and Edinburgh. International Journal of STD & AIDS, 4, 95-98.

Hart, G., Boulton, M., Fitzpatrick, R., McLean, J., & Dawson, J. (1992). `Relapse' to unsafe behavior among gay men: A critique of recent behavioural HIV/AIDS research. Sociology of Health and Illness, 14, 216-232.

Herek, G. M., & Glunt, E. K. (1995). Identity and community among gay and bisexual men in the AIDS era: Preliminary findings from the Sacramento men's health study. In G. M. Herek & B. Greene (Eds.), AIDS, identity, and community: The HIV epidemic and lesbians and gay men (pp. 55-84). Thousand Oaks, CA: Sage.

Hospers, H., Molenaar, S., & Kok, G. (1994). Focus group interviews with risk-taking gay men. Patient Education and Counseling, 24, 299-306.

Hunt, A. J., Weatherburn, P., Hickson, F. C. I., Davies, P. M., McManus, T. J., & Coxon, A. P. M. (1993). Changes in condom use by gay men. AIDS Care, 5, 439-448.

Jadack, R. A., Hyde, J. S., & Keller, M. L. (1995). Gender and knowledge about HIV, risky sexual behavior, and safer sex practices. Research in Nursing & Health, 18, 313-324.

Joffe, H. (1996). AIDS research and prevention: A social representational approach. British Journal of Medical Psychology, 69, 169-190.

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.

Kelly, J. A., Kalichman, S. C., Kauth, M. R., Kilgore, H. G., Hood, H. V., Campos, P. E., Rao, S. M., Brasfield, T. L., & St. Lawrence, J. S. (1991). Situational factors associated with AIDS risk behavior lapses and coping strategies used by gay men who successfully avoid lapses. American Journal of Public Health, 81, 1335-1338.

Kelly, J. A., Sikkema, K. J., Winett, R. A., Solomon, L. J., Roffman, R. A., Heckman, T. G., Stevenson, L. Y., Perry, M. J., Norman, A. D., & Desiderato, L. J. (1995). Factors predicting continued high-risk behavior among gay men in small cities: Psychological, behavioral, and demographic characteristics related to unsafe sex. Journal of Consulting and Clinical Psychology, 63, 101-107.

King, E. (1993). Safety in numbers: Safer sex and gay men. New York: Routledge.

Kippax, S., Connell, R. W., Dowsett, G. W., & Crawford, J. (1993). Sustaining safe sex: Gay communities respond to AIDS. London: Falmer.

Kippax, S., Crawford, J., Dowsett, G. W., Bond, G., Sinnott, V., Baxter, D., Berg, R., Connell, R. W., & Watson, L. (1990). Gay men's knowledge of HIV transmission and `safe' sex: A question of accuracy. Australian Journal of Social Issues, 25, 199-219.

Landrine, H. (Ed.) (1995). Bringing cultural diversity to feminist psychology: Theory, research, and practice. Washington, DC: American Psychological Association.

Leigh, B. C. (1990). The relationship of substance use during sex to high-risk sexual behavior. The Journal of Sex Research, 27, 199-213.

Levine, M. P., & Siegel, K. (1992). Unprotected sex: Understanding gay men's participation. In J. Huber & B. E. Schneider (Eds.), The social context of AIDS (pp. 47-71). Newbury Park, CA: Sage.

Lewis, J. E., Malow, R. M., & Ireland, S. J. (1997). HIV/AIDS risk in heterosexual college students: A review of a decade of literature. Journal of American College Health, 45, 147-158.

Lewis, L. A., & Ross, M. W. (1995). A select body: The gay dance party subculture and the HIV/AIDS pandemic. London: Cassell.

Lowy, E., & Ross, M. (1994). It'll never happen to me. AIDS Education and Prevention, 6, 467-482.

Maticka-Tyndale, E. (1992). Social construction of HIV transmission and prevention among heterosexual young adults. Social Problems, 39, 238-252.

Maxwell, A. E., Bastani, R., & Yan, K. X. (1995). AIDS risk behaviors and correlates in teenagers attending sexually transmitted diseases clinics in Los Angeles. Genitourinary Medicine, 71, 82-87.

Maxwell, C., & Boyle, M. (1995). Risky heterosexual practices among women over 30: Gender, power, and long term relationships. AIDS Care, 7, 277-293.

Maxwell, J. (1997). I can relate: Gay men, relationships, condoms. Retrieved June 14, 1999 from the World Wide Web: http://www.actoronto.org/ cl/safersex.nsf.

McAndrew, C., & Edgerton, R. B. (1969). Drunken comportment: A social explanation. Chicago: Aldine.

McClure, C., & Grubb, I. (1999). HIV transmission: Guidelines for assessing risk (3rd ed.). Ottawa, Canada: Canadian AIDS Society.

McCusker, J., Stoddard, A. M., McDonald, M., Zapka, J. G., & Mayer, K. H. (1992). Maintenance of behavioral change in a cohort of homosexually active men. AIDS, 6, 861-868.

Meyer, I. H., & Dean, L. (1995). Patterns of sexual behavior and risk taking among young New York City gay men. AIDS Education and Prevention, 7 (Suppl.), 13-23.

Misovich, S. J., Fisher, J. D., & Fisher, W. A. (1996). The perceived AIDS-preventive utility of knowing one's partner well: A public health dictum and individuals' risky sexual behavior. Canadian Journal of Human Sexuality, 5, 83-90.

Myers, T., Godin, G., Calzavara, L., Lambert, J., & Locker, D. (1993). The Canadian survey of gay and bisexual men and HIV infection: Men's survey. Ottawa, Canada: Canadian AIDS Society.

Myers, T., Rowe, C. J., Tudiver, F. G., Kurtz, R. G., Jackson, E. A., Orr, K. W., & Bullock, S. L. (1992). HIV, substance use and related behavior of gay and bisexual men: An examination of the talking sex project. British Journal of Addiction, 87, 207-214.

Myrick, R. (1999). In the life: Culture-specific HIV communication programs designed for African American men who have sex with men. The Journal of Sex Research, 36, 159-170.

Nemeroff, C. J. (1995). Magical thinking about illness virulence: Conceptions of germs from "safe" and "dangerous" others. Health Psychology, 14, 147-151.

Odets, W. (1994). AIDS education and harm reduction for gay men. AIDS & Public Policy Journal, 9, 3-15.

Odets, W. (1995). In the shadow of the epidemic: Being HIV-negative in the age of AIDS. Durham, NC: Duke University Press.

O'Donnell, L., San Doval, A., Vornfett, R., & O'Donnell, C. R. (1994). STD prevention and challenge of gender and cultural diversity: Knowledge, attitudes, and risk behaviors among Black and Hispanic inner-city STD clinic patients. Sexually Transmitted Diseases, 21, 137-148.

Offir, J. T., Fisher, J. D., Williams, S. S., & Fisher, W. A. (1993). Reasons for inconsistent AIDS-preventive behaviors among gay men. The Journal of Sex Research, 30, 62-69.

Parker, R. G., & Carballo, M. (1990). Qualitative research on homosexual and bisexual behavior relevant to HIV/AIDS. The Journal of Sex Research, 27, 497-525.

Paul, J. P., Stall, R., & Davis, F. (1993). Sexual risk for HIV transmission among gay/bisexual men in substance-abuse treatment. AIDS Education and Prevention, 5,11-24.

Penkower, L., Dew, M. A., Kingsley, L., Becker, J. T., Satz, P., Schaerf, F. W., & Sheridan, K. (1991). Behavioral, health and psychosocial factors and risk for HIV infection among sexually active homosexual men: The multicenter AIDS cohort study. American Journal of Public Health, 81, 194-196.

Perkins, D. O., Leserman, J., Murphy, C., & Evans, D. L. (1993). Psychosocial predictors of high-risk sexual behavior among HIV-negative homosexual men. AIDS Education and Prevention, 5, 141-152.

Peterson, J. L., Coates, T. J., Catania, J. A., Middleton, L., Hilliard, B., & Hearst, N. (1992). High-risk sexual behavior and condom use among gay and bisexual African-American men. American Journal of Public Health, 82, 1490-1494.

Pilkington, C. J., Kern, W., & Indest, D. (1994). Is safer sex necessary with a "safe" partner? Condom use and romantic feelings. The Journal of Sex Research, 31, 203-210.

Pollak, M. (1988). Les homosexuels et le sida: Sociologie d'une epidemie [Gay men and AIDS: Sociology of an epidemic]. Paris: Metailie.

Prieur, A. (1990). Norwegian gay men: Reasons for continued practice of unsafe sex. AIDS Education and Prevention, 2, 109-115.

Rotello, G. (1995, February 21). Beyond condoms. The Advocate, 80.

Sandfort, T. G. M., & van Zessen, G. (1992). Denial as a barrier for HIV prevention within the general population. Journal of Psychology and Human Sexuality, 5, 69-87.

Siegel, K., Mesagno, E P., Chen, J.-Y., & Christ, G. (1989). Factors distinguishing homosexual males practicing risky and safer sex. Social Science and Medicine, 28, 561-569.

Silvestre, A., Lyter, D., Valdiserri, R., Huggins, J., & Rinaldo, C. (1989). Factors related to seroconversion among homo- and bisexual men after attending a risk-reduction educational session. AIDS, 3, 647-650.

Smith, D. E. (1987). The everyday world as problematic: A feminist sociology. Toronto, Canada: University of Toronto Press.

Sobo, E. J. (1995). Choosing unsafe sex: AIDS-risk denial among disadvantaged women. Philadelphia: University of Pennsylvania Press.

Temple, M. T., & Leigh, B. C. (1992). Alcohol consumption and unsafe sexual behavior in discrete events. The Journal of Sex Research, 29, 207-219.

Waddell, C. (1992). Social correlates of unsafe sexual intercourse. Australian and New Zealand Journal of Sociology, 28, 192-207.

Watney, S. (1989). Taking liberties: An introduction. In E. Carter & S. Watney (Eds.), Taking liberties: AIDS and cultural politics (pp. 11-57). London: Serpent's Tail.

Weatherburn, P., Davies, P. M., Hickson, F. C. I., Hunt, A. J., McManus, T. J., & Coxon, A. P. M. (1993). No connection between alcohol use and unsafe sex among gay and bisexual men. AIDS, 7, 115-119.

Williams, S. S., Kimble, D.L., Covell, N. H., Weiss, L. H., Newton, K. J., Fisher, J. D., & Fisher, W. A. (1992). College students use implicit personality theory instead of safer sex. Journal of Applied Social Psychology, 22, 921-933.

Manuscript accepted July 30, 1999

Barry D. Adam and Alan Sears University of Windsor, Canada

E. Glenn Schellenberg University of Toronto, Canada

This research was supported by a grant from the National Health and Research Development Program, a division of Health Canada. Most of the interviews were conducted by Stuart Gibson, Scott Mattson, and Kevin Smith.

Address correspondence to Barry Adam, Department of Sociology and Anthropology, University of Windsor, Windsor, ON, Canada, N9B 3P4; e-mail: [e-mail].