The Prevalence of Sexual Aggression and Victimization Among Homosexual Men.

Sexual aggression of men toward women has received widespread attention in psychological research for many years (e.g. Allison & Wrightsman, 1993; Craig, 1990; Koss et al., 1994; Schwartz & DeKeseredy, 1997). Sexual aggression includes a range of forced sexual activities, such as rape (unwanted sexual intercourse through the use or threat of physical violence), unwanted sexual intercourse through the use of verbal pressure, and other unwanted sexual acts, such as touching and kissing (cf. Abbey, McAuslan, & Ross, 1998). There is ample evidence to show that sexual aggression is a serious problem in heterosexual relationships. For example, Koss, Gidycz, and Wisniewski (1987) reported that 15% of women in a large college sample had been forced to engage in unwanted sexual intercourse since the age of 14. Over 4% of their male respondents indicated they had forced a woman into sexual intercourse since the age of 14. Other more recent studies provide similar rates of experiencing or performing sexual aggression (e.g., Abbey et al., 1998; Abbey, Ross, McDuffie, & McAuslan, 1996; Himelein, 1995; Tyler, Hoyt, & Whitbeck, 1998).

In contrast, the problem of sexual aggression against male victims has been largely ignored in sexual aggression research, and evidence on the prevalence and determinants of male sexual victimization is limited. This may be due in part to the fact that legal definitions of rape have traditionally been restricted to female victims, and changes in legal regulations are slow to be implemented (King, 1992; Scarce, 1997). Until 1997, the German penal code defined rape as the use or threat of violence against a woman to force her into extramarital sexual intercourse. In a revision which came into effect in July 1997, a broader definition was adopted which removed the restriction to female victims and nonmarital intercourse. The new legal definition includes the use or threat of violence or the exploitation of the victim's incapacitated state to force another person into sexual activities. Forced sexual activities qualify as rape if the act involves penetration of the victim's body, which is no longer restricted to vaginal intercourse.

Given that men feature as the main group of perpetrators of sexual aggression, homosexual men are at greater risk of being forced into unwanted contacts because they seek sexual contacts with a male partner. Evidence comparing the risk of sexual victimization for homosexual and heterosexual men is scarce, but it suggests that homosexual men are particularly vulnerable to sexual victimization (e.g., Stermac, Sheridan, Davidson, & Dunn, 1996). Hodge and Canter (1998) looked at both heterosexual and homosexual victims of male sex offenders and found that assaults on homosexual victims tended to be more violent than those on heterosexual victims. Further studies explored the prevalence of sexual victimization in homosexual men without including a heterosexual comparison group. In a sample of 930 homosexually active men, Hickson et al. (1994) reported that 27.6% had been subjected to unwanted sexual contact at some point in their lives. In a much smaller sample of 34 homosexual college students, Waterman, Dawson, and Bologna (1989) identified a prevalence rate of 12.1%. These studies suggest that even though prevalence rates may be lower than for female victims, sexual aggression is a problem among homosexual men that needs to be addressed. This is especially true since the traumatizing effect of sexual assault on a male victim was shown to be no less serious than on female victims, often following the pattern of the "rape trauma syndrome" identified for female victims of sexual assault (Anderson, 1982; Coxell & King, 1996; Scarce, 1997; Turner, 1992).

However, the current body of evidence on sexual aggression among homosexual men is limited not only by the small overall number of studies, but by additional methodological problems:

1. In the majority of studies, information is collected from small samples of self-identified victims, often contacted via crisis intervention agencies or similar services (e.g., Kaufman, Divasto, Jackson, Voorhees, & Christy, 1980; Mezey & King, 1989). These studies do not speak to the prevalence of sexual aggression in the population of homosexual men at large.

2. Broad questions, (e.g., "Have you ever been forced to have sex against your will?") are often used to elicit reports of sexual victimization (Duncan, 1990). Few studies provide detailed information about (a) the type of coercion used, (b)the sexual activity to which the victim was subjected, and (c) the nature of the victim-perpetrator relationship.

3. Researchers have focused almost exclusively on the victim, not the perpetrator. When information is reported about offender characteristics, it does not come from offenders themselves, but is extracted from victim reports (e.g. Hodge & Canter, 1998). We located only one study (Waterman et al., 1989) which questioned homosexual men directly about their perpetration of sexually aggressive acts. Given that victims may not always be correctly informed about the characteristics of their assailants (e.g. the assailant's sexual orientation) or may have distorted impressions about the assailant because of the trauma of the assault, this is an unfortunate limitation. The lack of perpetrator surveys also stands in stark contrast to research on heterosexual aggression where the collection of parallel data sets from female victims and male perpetrators is a standard procedure (e.g., Koss et al., 1987; Koss & Oros, 1982; Tyler et al., 1998).

The present study was designed to address these limitations and add to our current knowledge of sexual aggression among homosexual men. We addressed prior methodological limitations by examining sexual aggression among a large sample of homosexual men. The first aim was to establish prevalence rates for sexual aggression among homosexual men involving different types of coercive strategies (physical force, exploitation of the victim's incapacitated state, and verbal pressure) and different sexual acts (touch, masturbation, oral sex, and anal sex). In addition, we explored the possibility that the prevalence of sexual aggression and victimization varies as a function of the victim-perpetrator relationship, such as (ex-)partners, friends/acquaintances, and unknown men. The Sexual Experiences Survey (SES), an instrument with documented psychometric properties (Koss & Gidycz, 1985; Koss et al., 1987; Koss & Oros, 1982) was used in a revised form to measure prevalence rates. The advantage of the SES is that it explicates the meaning of "unwanted sexual contacts" as referring to situations where people are made to engage in sexual activities against their will through the use or threat of physical force (e.g. hitting, holding down, injuring), the exploitation of the other person's inability to offer resistance (e.g. after heavy drinking), or the use of verbal pressure (e.g. threat to end the relationship). Revisions were made to adapt the SES to the assessment of sexual aggression and victimization among homosexual men.

A second aim was to elicit information about both victimization by and perpetration of sexually aggressive acts in order to examine the potential overlap of victim and perpetrator roles in the same individual. Unlike studying heterosexual aggression where women are approached as potential victims and men are approached as potential perpetrators, the assignment of respondents to the victim or perpetrator role cannot be made a priori in the study of male-on-male sexual aggression. Respondents may be either victims or perpetrators of sexual aggression, or they may have experienced both roles. Therefore, research instruments need to cover both perspectives (cf. Larimer, Lydum, Anderson, & Turner, 1999, for a similar approach in the heterosexual domain). Especially when the aim is to examine possible risk factors of sexual aggression, it is important to identify the proportion of victims who have also acted as perpetrators, and vice versa, to control for possible confounds due to the overlap of the two roles. Moreover, if perpetration and victimization were operationalized as a "between subjects" variable, asking half of the sample about victimization and the other half about perpetration, the perpetration sample would inevitably include some victimized men. From an ethical point of view, it seems problematic to approach these men as potential perpetrators without giving them the opportunity to report their experiences of victimization.

Examination of each person as potential victim and as perpetrator has the disadvantage of raising the possibility of carry-over effects from one perspective to the other. That is, people who are first asked to report the perpetration of sexually aggressive acts may be subsequently more likely to report victimization experiences as a blame reduction strategy. Conversely, respondents who first identify themselves as victims may be less likely to admit afterwards that they have shown aggressive behavior themselves. To control for this potential source of reporting bias, the present study employed two versions of the sexual aggression measure in which the order of presentation of the victimization and perpetration items was counterbalanced.

Our third objective was to explore two aspects of a homosexual activity: (a) seeking sexual contacts in anonymous settings (such as cruising grounds and public lavatories--"tea rooms") and (b) engaging in sexual contacts in exchange for money. Several large-scale surveys have documented that many homosexual men engage in these practices (e.g., Bochow, 1997; Coxon, 1996). In Bochow's survey of over 3,000 homosexual men in Germany, 44.6% of the respondents reported that more than half, most or all of their sexual contacts over the last 12 months had been anonymous (i.e., were on-off contacts with a previously unknown man whom they would only expect to see again by chance). These sexual contacts involve a higher risk of aggression and victimization: First, there is an implicit understanding between partners that they both came to the respective place in search for sexual contacts. Therefore, rejections of sexual advances may be seen as inappropriate and may be ignored. Second, there is no build-up to the sexual encounter in which sexual intentions and desires can be mutually explored. Third, it is well known from social psychological research on deindividuation (Postmes & Spears, 1998) that anonymity lowers the threshold for antisocial behavior. For these reasons, both victimization and aggression are expected to be more prevalent among men who seek sexual contacts in anonymous settings.

Engaging in sexual contacts in exchange for money also may be a potential risk factor for sexual victimization. Men who pay another person for sexual contacts may feel that this gives them the right to enforce their sexual wishes upon the other. Conversely, men who accept money for sex may find it harder to exert their right of sexual self-determination and reject unwanted sexual activities from the person who paid them. In Bochow's (1997) sample, 9% of respondents indicated that they had paid for sex in the last twelve months, while 3% had accepted money for sex during the same period. Of those 3%, only five men (5%) had accepted money regularly (more than 50 times). For the remaining men, the money did not make a significant contribution to their income. As Bochow suggests, these men engage in a hedonistic lifestyle which includes the readiness to accept money for sex alongside other more liberal sexual attitudes. Thus, we predict that a sizeable proportion of homosexual men report having paid or accepted money for sex and that these behaviors are linked to an increased risk of sexual victimization and aggression.

In sum, the present study seeks (a) to provide data on the prevalence of sexual victimization and aggression from a comparatively large sample, (b) to examine the potential overlap between victim and perpetrator role, and (c) to examine the effect of two sexual activity variables (i.e., having sex in anonymous places and exchanging money for sex) on prevalence rates of sexual aggression and victimization.

METHOD

Participants

The sample consisted of 310 men between the ages of 14 and 35 years (M = 21.79, SD = 3.56).(1) All respondents had experienced sexual contacts with a male partner. Respondents were members of the general public contacted in a variety of settings frequented by homosexual men (cf. below). German nationality was reported by 92.3% respondents, 2.3% were Turkish, and 4.5% were other foreign nationals. In terms of their educational background, 9.4% had completed (or were due to complete) 9 years of formal schooling ("Hauptschulabschlu[Beta]"), 28.1% had completed (or were aiming to complete) 10 years of formal schooling ("Realschulabschlu[Beta]"), and 61.9% had completed (or were aiming to complete) 13 years of formal education (university entrance qualification). The educational background of the present sample is similar to that reported by Bochow (1997).

Instruments

Respondents were asked to complete an anonymous questionnaire that included measures of sexual aggression and victimization, sexual experience, and demographic variables. Sexual aggression and victimization were measured with a revised version of the "Sexual Experiences Survey" (SES: Koss et al., 1987; Koss & Oros, 1982) which we named the Homosexual Experiences Survey (HSES). The SES contains 12 items describing different forms of sexual aggression, and respondents indicate in a yes-no format whether or not they have ever experienced (female version) or committed (male version) that form of sexual aggression. The SES is widely used in the heterosexual domain as a measure of sexual aggression with demonstrated reliability and validity (Karabatsos, 1997; Koss & Gidycz, 1985; Koss & Oros, 1982; Ullman, Karabatsos, & Koss, 1999). Since it is geared toward the measurement of heterosexual aggression, the majority of items refer to vaginal intercourse. These items had to be modified to reflect the different sexual practices prevalent among homosexual men.

The HSES was derived from a German version of the SES used in a series of studies to investigate heterosexual aggression among German adolescents (e.g., Krahe, Scheinberger-Olwig, & Waizenhofer, 1999; Krahe, Scheinberger-Olwig, Waizenhofer, & Kolpin, 1999). The German adaptation of the heterosexual SES was developed in a careful process of translation and back translation and was shown to have retest reliabilities of 95.6% (male respondents) and 94.3% (female respondents) over a period of 3-4 weeks (cf. Krahe, Reimer, Scheinberger-Olwig & Fritsche, 1999).(2) Thus, the instrument which formed the basis for the HSES has been established to be psychometrically sound for the use with a German sample.

The revised version breaks down reports of sexual aggression and victimization into different types of relationships between aggressor and victim: (a) an (ex-)partner, (b) a friend or acquaintance, and (c) an unknown man. The relevance of these categories is supported by evidence on the sexual victimization of men (e.g., Stermac et al., 1996). Three different forms of coercing another person into sexual acts against his will are explored: (a) the use (or threat) of physical force, (b) the exploitation of another man's inability to resist unwanted sexual advances because he is in an incapacitated state; and (c) the use of verbal pressure. Each of these three forms of aggression is combined with four types of sexual acts: (a) touching; (b) masturbation; (c) oral sex; (d) anal sex. These four sexual acts were chosen on the basis of evidence on the predominant sexual practices of homosexual men (e.g. Bochow, 1997; Coxon, 1996). It was decided to include colloquial terms for the different practices to facilitate understanding in a sample with varying educational backgrounds (cf. Coxon).

The total HSES consisted of six items, three for each type of coercive strategy from the perpetrator perspective and three from the victim perspective. The full text of the items and the response format for each item are presented in the Appendix. The victimization and perpetration items were presented in descending order of severity (i.e., physical force followed by exploitation of incapacitated state and verbal pressure). The items were preceded by an introductory part which clarified that the HSES items did not refer to incidents of sexual abuse in childhood or in relationships with a clear power differential (e.g., teacher-pupil). To control for potential order effects in the presentation of the victimization and perpetration items, respectively, two versions of the HSES were created. Half of the respondents received the victimization items first, and the other half received the perpetration items first. Reliability analyses revealed the HSES to have high internal consistency, with [Alpha] = .90 for the victimization items and [Alpha] = .88 for the perpetration items.

The final part of the questionnaire contained four questions addressing the respondents' sexual lifestyle: (a) "How would you rate your sexual orientation?" Ratings were made on a seven-point scale ranging from (1) = homosexual to (7) = heterosexual;(3) (b) "Have you ever had sex in tea rooms, parks, etc.?" (Response options yes/no), (c) "Have you ever received money for sexual activities?", and (d) "Have you ever paid money for sexual activities?". Response options for items (c) and (d) were no, once, and repeatedly. In addition, respondents were asked to indicate the age of their first homosexual contact and the age at which they had their "coming out" that made their homosexual orientation known to others.

Procedure

The survey was conducted in the city of Berlin, which has a large homosexual community. Respondents were approached in a variety of places frequented by homosexual men, including social clubs, different gay events, and youth centers. According to Bochow's (1997) survey, the majority of homosexual men regularly visit such places. Locations frequented only for the sake of obtaining sexual contacts (public sex environments: Coxon, 1996) were excluded from the range of data collection sites. The questionnaires were handed out by male and female research assistants with prior training and experience in administering sexual aggression surveys. Care was taken to ensure that respondents completed the questionnaire individually in a quiet place arranged in the respective settings. The average time for completing the questionnaire was 15-20 minutes. Respondents were paid 10 DM (approximately US $5) for their participation. After returning the questionnaires, all respondents received a list of local counseling agencies offering advice to victims and perpetrators of sexual violence.

RESULTS

Sexual Background Variables

The sexual orientation of the present sample was predominantly homosexual, with a mean score of M = 1.71 (SD = .91) on a seven-point scale ranging from (1) = homosexual to (7) = heterosexual. Fifty-one percent of respondents indicated that they had had sexual contacts in anonymous places. Money was accepted in return for sex on one occasion by 12.3% of respondents or repeatedly by 11%. Money was paid in return for sex by 4.8% on one occasion and by 2.6% repeatedly. The average age of first homosexual contact was M = 16.55 years (SD = 2.80), the average age at which respondents had their "coming out" was M = 17.84 years (SD = 2.45).

Prevalence of Sexual Aggression and Victimization

Since the aim of the study was to establish the prevalence of sexual aggression among homosexual men, frequency analyses were performed on the endorsement rates for the different forms of sexual aggression and victimization tapped by the HSES. Since the combination of four sexual acts, three coercive strategies, and three relationship types yields a large number of cells with relatively small frequencies, the analyses were conducted in two steps to provide a more clear-cut picture of the results. In the first step, the focus was on the different types of sexual acts, and data were collapsed across relationship types. In the second step, the focus was on relationship types, and the data were collapsed across sexual acts. The findings concerning sexual victimization will be presented first (Table 1), followed by the findings for the perpetrator perspective (Table 2).

Table 1. Prevalence of Sexual Victimization: Coercive Strategy, Type of Sexual Act, and Victim-Perpetrator Relationship

                          Coercive strategy

                                   Exploitation
                                        of
                        Physical   incapacitated
                         force         state

Type of sexual act
  Sexual touch            12.6         15.1
  Masturbation             6.8         11.3
  Oral sex                 7.1         11.3
  Anal sex                 5.5          9.0
  Total                   15.2         20.0

Victim-perpetrator
relationship
  (Ex-)partner             3.5          6.1
  Friend/acquaintance      6.8          9.0
  Unknown man             10.6          8.7

                        Coercive strategy

                         Verbal
                        pressure   Total

Type of sexual act
  Sexual touch             7.1      24.2
  Masturbation             5.8      17.1
  Oral sex                 6.8      17.4
  Anal sex                 3.9      13.5
  Total                   10.3

Victim-perpetrator
relationship
  (Ex-)partner             4.2      11.0
  Friend/acquaintance      2.6      14.2
  Unknown man              3.9      16.5

Note. N = 310. Values indicated in %. Multiple responses were possible across coercive strategies, types of sexual acts, and victim-perpetrator relationships. Therefore, "Total" scores are lower than the sum of the single categories.

Table 2. Prevalence of Sexual Aggression: Coercive Strategy, Type of Sexual Act, and Victim-Perpetrator Relationship

                           Coercive strategy

                                   Exploitation
                                        of
                        Physical   incapacitated
                         force         state

Type of sexual act
  Sexual touch            4.5          14.2
  Masturbation            2.9          10.3
  Oral sex                2.6           9.4
  Anal sex                1.6           3.9
  Total                   5.5          16.8

Victim-perpetrator
relationship
  (Ex-)partner            3.2           8.1
  Friend/acquaintance     2.6          11.3
  Unknown man             1.3           2.3

                        Coercive strategy

                         Verbal
                        pressure   Total

Type of sexual act
  Sexual touch            5.2       18.4
  Masturbation            3.9       13.9
  Oral sex                3.2       13.2
  Anal sex                1.9        6.5
  Total                   6.5

Victim-perpetrator
relationship
  (Ex-)partner            4.5       12.6
  Friend/acquaintance     2.6       12.9
  Unknown man             0.3        1.3

Note. N = 310. Values indicated in %. Multiple responses were possible across coercive strategies, types of sexual acts, and victim-perpetrator relationships. Therefore, "Total" scores are lower than the sum of the single categories.

Sexual victimization. The prevalence rates for sexual victimization are presented in Table 1. The Total row in the top half of Table 1 shows that just over 15% of the respondents experienced some form of physical victimization (or threat thereof) from another man to make them comply with his sexual demands. Twenty percent reported sexual victimization in an incapacitated state where the perpetrator exploited their inability to offer resistance. Ten percent gave in to unwanted sexual acts following verbal coercion. In interpreting these findings, it should be borne in mind that the HSES allows for multiple responses, so that the figures in Table 1 should not be added up across the different categories.

Beyond providing total prevalence rates, Table 1 breaks down victimization reports into different sexual acts. Unwanted touch, such as kissing and stroking, was the most frequently reported form of sexual victimization, followed by oral sex, masturbation and anal sex. Table 1 also speaks to the prevalence of sexually aggressive acts which meet the legal definition of rape: oral and anal sex involving the use of force or the exploitation of incapacitation after alcohol or drugs. Here, figures range from 5.5% (anal sex through the use of force) to 11.3% (oral sex through the use of alcohol/drugs). Aggregating across the four forms of sexual acts, 16.1% of respondents reported victimization meeting the legal definition of rape.

As far as the relationship between victim and perpetrator is concerned, the bottom half of Table 1 reveals few substantial differences in the prevalence of the three different forms of sexual coercion in the relationship categories considered here. A noteworthy exception is victimization through the use of physical force, which is about twice as prevalent in relationships with

friends/acquaintances compared to (ex-) partners and even three times more frequent with unknown men.

Perpetration of sexual aggression. The prevalence rates for the perpetration of sexually aggressive acts are presented in Table 2. Prevalence rates for the perpetration of sexual aggression are lower than the corresponding victimization figures. Nevertheless, over 5% of respondents indicated that they had used, or threatened to use, physical force to make an unwilling man comply with their sexual demands; over 16% reported exploiting a person's inability to resist their advances; and over 6% admitted to the use of verbal coercion. Parallel to the victimization reports, unwanted touch was found to be the most frequent form of sexual act forced upon another man, followed by masturbation, oral sex, and anal sex. Prevalence rates for sexual acts qualifying as rape ranged from 1.6% (anal sex through the use of physical force) to 9.4% (oral sex through exploitation of the victim's incapacitated state), with an overall prevalence of 11.3%.

With respect to victim-perpetrator relationship, the data reveal that exploiting the victim's inability to resist is the most prominent coercive strategy in each of the three relationship categories. It is particularly frequent in relationships with (ex-) partners and friends or acquaintances. In fact, perpetration rates are higher than the corresponding victimization rates for these two forms of sexual coercion. Overall, however, it may be concluded that sexual aggression occurs in different types of relationships and that there does not seem to be a strong link between the nature of the relationship and the use of a particular coercive strategy.

Order of Presentation of Victimization and Perpetration Questions

For each of the entries in Tables 1 and 2, nonparametric tests (Mann-Whitney U-Tests) were conducted to examine whether the order of presentation of the victimization and perpetration items had an effect on prevalence rates. None of the tests produced a significant difference between the two versions (all U-values with p [is greater than] .15). Thus, it may be concluded that the order of presentation did not systematically affect reporting rates.

Accepting Money for Sex and Having Sex in Anonymous Places

A series of loglinear analyses was conducted to examine the impact of having sex in anonymous places and accepting money for sex on the likelihood of sexual victimization and the likelihood of committing sexual aggression (Norusis, 1993). The number of respondents who reported paying money for sexual contacts was too small to include this variable in the analyses. Since multiple responses were allowed across coercive strategies as well as relationship types, these variables were not independent and thus could not be entered simultaneously into a single loglinear analysis. Separate analyses were therefore conducted for each of the three types of coercion (physical force, alcohol/drugs, and verbal pressure) and the three relationship types with respect to (a) victimization by and (b) perpetration of sexual aggression.

With respect to the risk of sexual victimization, having sex in anonymous places was not related to any of the coercive strategies or relationship types. For acceptance of money for sex, three significant effects were found, as shown in Table 3.

Table 3. Percentage of Sexual Victimization as a Function of Having Accepted Money for Sex

                                Acceptance of
                                money for sex

                                No         Yes
                             (n = 233)   (n = 72)

Coercive strategy involved
  Physical force               12.4        23.6
  Alcohol/drugs(*)             15.5        33.3
  Verbal pressure               9.4        12.5

Relationship type
  (Ex-)partner(*)               8.2        18.1
  Friend/acquaintance(*)       10.3        25.0
  Unknown man                  13.7        26.4

(*) p < .05.

The significant effects indicate that the percentage of sexually victimized respondents in the "no acceptance of money" group is significantly lower and the percentage of respondents in the "acceptance of money" group significantly higher than would be expected if the two variables were independent. Having accepted money for sex significantly increased the likelihood of being sexually victimized through the use of alcohol or drags (Partial [chi square] [df = 1, N = 303] = 8.788, p [is less than] .01), of being victimized by an (ex-)partner (Partial [chi square] [df= 1, N = 303] = 3.788, p [is less than] .05) and of being victimized by a friend/acquaintance (Partial [chi square] [df = 1, N = 303] = 7.809, p [is less than] .01). The parameter estimates for these effects all had Z-values higher than +/- 1.96 (i.e., were significant at least at p [is less than] .05).

However, each of the three effects was qualified by a significant interaction with having sex in anonymous places. These interactions indicate that the lower risk of victimization for respondents who do not accept money for sex is only true for men who did not have sex in anonymous places. For these respondents, the prevalence of sexual victimization went up from 11.0% to 44.4% as a function of accepting money for sex. For those who report sex in anonymous places, a much smaller increase was found from 20.0% to 30.2% (Pearson [chi square] [df = 1, N = 303] = 3.924, p [is less than] .05). In a similar way, the prevalence of victimization by an (ex-)partner increased from 4.7% to 27.8% as a function of accepting money for respondents who did not report sex in anonymous places. For those who did have sex in anonymous places, the increase in risk as a function of taking money is only from 12.4% to 15.1% (Pearson [chi square] [df= 1, N = 303] = 5.162, p [is less than] .05). Finally, the prevalence of victimization by a friend or acquaintance went up from 7.9% to 38.9% as a function of accepting money for respondents who do not report sex in anonymous places, while the difference was only 12.4% vs. 18.9% for those who did have sex in anonymous places (Pearson [chi square] [df = 1, N = 303] = 4,255, p [is less than] .05). Again, all parameter estimates were significant at p [is less than] .05. Thus, adding sex in anonymous places as a risk factor attenuated the difference in risk between respondents who do not and those who do accept money for sex.

Parallel loglinear analyses were conducted on the prevalence rates for the perpetration of sexually aggressive acts. Again, having sex in anonymous places was not linked to higher prevalence rates. Acceptance of money for sex affected the likelihood of sexual aggression in five of the six analyses, as shown in Table 4.

Table 4. Percentage of Perpetration of Sexual Aggression as a Function of Having Accepted Money for Sex

                              Acceptance of
                              money for sex

                              No         Yes
                           (n = 233)   (n = 72)
Coercive strategy used
  Physical force(*)           3.4        12.5
  Alcohol/drugs(*)           12.0        31.9
  Verbal pressure             4.3        11.1

Relationship type
  (Ex-)partner(*)             9.4        20.8
  Friend/Acquaintance(*)      9.0        25.0
  Unknown man(*)              1.3         8.3

(*) p < .05.

Note that significant effects indicate that the percentage of sexually aggressive respondents in the "no acceptance of money" group is significantly lower and the percentage of respondents in the "acceptance of money" group significantly higher than would be expected if the two variables were independent. Having accepted money for sex significantly increased the likelihood of obtaining sexual contacts through the use of force (Partial [chi square] [df = 1, N = 303] = 6.353, p = .01) and through the use of alcohol or drugs (Partial [chi square] [df = 1, N = 303] = 14.976, p [is less than] .001). Acceptance of money for sex was found to interact with having sex in anonymous places for the use of verbal pressure (Pearson [chi square] [df= 1, N = 303] = 4.809, p [is less than] .05): the difference in perpetration rates associated with accepting money for sex was significant only for respondents who did not report sex in anonymous places (1.6% vs. 16.7%), while the difference was insignificant for respondents who did have sex in anonymous places (7.6% vs. 9.4%).

In terms of victim-perpetrator relationship, acceptance of money for sex was associated with a significant increase in the percentage of respondents who had sexually victimized an (ex-)partner (Partial [chi square] [df= 1, N = 303] = 7.966, p [is less than] .01), a friend/acquaintance (Partial [chi square] [df= 1, N = 303]) = 10.949, p [is less than] .001) and an unknown man (Partial [chi square] [df= 1, N = 303] = 7.062, p [is less than] .01). All parameter estimates were significant at least at p [is less than] .05.

Overlap of Victim and Perpetrator Roles

A final question of interest refers to the independence versus interdependence of victim and perpetrator roles. It is a commonly held assumption that victimization is a risk factor of perpetration, in other words, that individuals who were sexually victimized have a higher risk of subsequently committing sexually aggressive acts. Evidence for this assumption comes from the child sexual abuse literature, where men who sexually abused children were frequently found to have been abused themselves in childhood (Browne, 1994).

To explore this link, we examined the distribution of respondents across three categories: (a) victim only (reporting victimization experiences but no perpetration); (b) perpetrator only (reporting perpetration, but no victimization); and (c) victim and perpetrator (indicating both victimization and perpetration experiences). Since this analysis excludes respondents who reported neither victimization nor perpetration, the N was reduced to 161 respondents. Of these, 57 (35.4%) were both victims and perpetrators of sexual aggression, 77 respondents (47.8%) were in the victims-only group, and 27 (16.8%) fell into the perpetrator-only group. The fact that over a third of the respondents who had been involved in sexual aggression reported both victimization and perpetration lends some support to the assumption that victimization and sexual aggression are related, but the causal direction of the link remains undetermined in the present design.

DISCUSSION

The present study investigated the prevalence of sexual aggression among homosexual men, aiming to contribute evidence from a larger data base to the predominantly small-scale surveys and case studies reported in previous research. The findings show that over 15% of the present sample reported experiences where another man had used, or threatened to use, physical force to make them comply with his sexual advances. One in five respondents reported having been subjected to unwanted sexual acts when the assailant exploited their inability to offer resistance. Verbal pressure was experienced by over 10% as a means to make them engage in sexual contacts against their will. Sexual acts which meet the legal definition of rape (penetration of the body through use of force or exploitation of the victim's incapacitated state) were experienced by 16.1% of the sample. Even though these rates are lower than those found for female victims of heterosexual aggression using similar instruments (Krahe 1998; Tyler et al., 1998), they identify sexual victimization as a serious problem among homosexual men which requires attention.

In contrast to previous studies, the present survey asked not only for self-reports of sexual victimization, but also for incidents in which the respondents had inflicted sexual aggression on another man. Self-reports of sexual aggression were found to be substantially lower than the corresponding victimization rates. Similar discrepancies are routinely found in studies where victimization and perpetration rates are collected from different individuals. For example, in the Tyler et al. (1998) study, 11.2% of the female respondents reported that they had been held down physically by a man to force them into sexual intercourse, while 2.6% of the male respondents reported having held down a woman to force her to have sexual intercourse with them (cf. also Koss et al., 11987; Koss & Oros, 1982; Krahe, 1998). The discrepancy may be explained in large part by the fact that perpetrators are asked to report behavior which is highly undesirable and potentially subject to legal prosecution.

Given the general tendency to underreport sexually aggressive behavior, it is noteworthy that over 5% of the respondents indicated that they had used, or threatened to use, force to make another man engage in sexual acts against his will, 16.8% acknowledged exploitation of another man's incapacitated state, and 6.5% indicated the use of verbal pressure. More than 1 in 10 respondents (11.3%) reported sexual coercion which met the legal definition of rape.

Because reports of both perpetration and victimization were collected, it was possible to examine the relationship between the two roles. Of the respondents who had been involved in sexually aggressive encounters, over a third had experiences as both victims and perpetrators. This finding suggests that it is not uncommon for the victim and assailant roles to coincide in one individual. However, because the present survey asked for life-time prevalence, it is impossible to determine the temporal sequence of the victim and assailant experiences, which would be a prerequisite for any causal assumption about the relationship between victim and perpetrator role. As an alternative interpretation, the possibility must be considered that some men who committed sexually aggressive acts also reported victimization as a blame reduction strategy. However, the fact that the order of presentation of the victimization and perpetration items did not affect prevalence rates speaks against this interpretation.

In addition to prevalence rates for the total sample, the impact of two potential risk factors was examined: having sexual contacts in anonymous places and accepting money in return for sexual contacts. Having sex in anonymous places did not significantly affect prevalence rates of sexual victimization. Acceptance of money for sex was associated with a higher incidence of sexual victimization in an incapacitated state and in relationships with (ex-)partners and friends or acquaintances. These effects were qualified, however, by an interaction between accepting money and having sex in anonymous places: The protective effect of not accepting money for sex was true only for those who refrained from sex in anonymous places. If men did report sex in anonymous places, their risk of sexual victimization did not vary as a function of accepting money for sex.

In terms of the perpetration of sexual aggression, prevalence rates were significantly increased for each coercive strategy and each relationship type for men who accepted money for sex. Having sex in anonymous places failed to exert an independent influence. These findings suggest that accepting money for sex enhances the vulnerability for sexual aggression, both from the victim and the perpetrator perspective. The fact that taking money for sex was found in the present study to be related to increased prevalence rates, not just for casual sexual relationships (acquaintances, unknown men) but also in relationships with (ex-)partners, corroborates the idea that this behavior is generally associated with a higher risk of sexual aggression (see also Bochow, 1997).

Taken together, the findings from this study show that a substantial proportion of homosexual men experienced unwanted sexual contacts which they were subjected to through force, exploitation of their inability to resist, or verbal coercion. A smaller, but still considerable number of men reported having used aggressive strategies to pursue their sexual interests.

It is important, however, to note the limitations of the present database. First, even though the sample is large compared to most other studies on sexual assault among homosexual men, it is still a convenience sample from which generalizations to the homosexual population at large have to remain tentative. Second, the HSES, while identifying specific forms of sexual coercion, is not exhaustive. For example, no distinction was made between the active (performer) and passive (recipient) role in the sexual acts involved. In order to take this difference into account, the instrument would have become too long and complex. Even as it stands, the HSES requires considerable attention from the participants, and the likelihood of fatigue affecting responses would undoubtedly be increased by breaking down the response format even further. Furthermore, it must be borne in mind that this study, just like almost all other studies on sexual aggression, elicited self-reports of victimization and perpetration which are potentially subject to socially desirable response biases. However, social desirability concerns are more likely to depress rather than to inflate prevalence rates. Therefore, the prevalence data reported in this study may be seen as a conservative estimate of the scale of sexual aggression among homosexual men.

In addition, the lack of related research investigating homosexual aggression limits the interpretation of the present prevalence rates. Comparative data in particular are needed to see if the simultaneous collection of victimization and perpetration had an effect on prevalence rates. Research on heterosexual aggression, which found similar discrepancies between victimization reports from women and perpetration reports from men, speaks against such a possibility. Nonetheless, further evidence is needed from the homosexual domain to rule out this potential source of reporting bias. Furthermore, data on sexual victimization of heterosexual men by male perpetrators is needed to assess the particular vulnerability of homosexual men. Given that sexual assault has a lasting, often traumatic impact on the victims and that empirical evidence is still scarce, it is time for the study of male-on-male sexual aggression to move up the agenda of sexual aggression research. Intensified research efforts are needed to gain a clearer understanding of the scale, the risk factors, and the consequences of sexual aggression among homosexual men.

(1) Of the 325 respondents who agreed to participate, 12 were excluded because they indicated that they had not had any homosexual experience, and three were excluded because they had missing data on all critical items of sexual aggression and victimization.

(2) Percentages rather than chance-corrected measures of retest reliability were used with these data because of the highly skewed response distribution in the direction of "no" responses to the sexual aggression and victimization items (cf. Krahe, Reimer, et al., 1999 for a detailed discussion).

(3) A graded rating scale rather than a forced-choice format was used to measure sexual orientation to allow for the possibility that respondents were unable to or did not want to classify themselves as exclusively homosexual or heterosexual. As Radkowsky and Siegel (1997, p. 193) note, regarding one's sexual attraction to other men as "just a phase" or "not really gay" is a common response to the discovery of same-sex attraction in young men. Therefore, presenting heterosexual vs. homosexual orientation as a continuum was regarded as more appropriate for this relatively young sample.

REFERENCES

Abbey, A. A., McAuslan, P., & Ross, L. T. (1998). Sexual assault perpetration by college men: The role of alcohol, misperception of sexual intent, and sexual beliefs and experiences. Journal of Social and Clinical Psychology, 17, 167-195.

Abbey, A. A., Ross, L. T., McDuffie, D., & McAuslan, P. (1996). Alcohol and dating risk factors for sexual assault among college women. Psychology of Women Quarterly, 20, 147-169.

Allison, J. A., & Wrightsman, L. S. (1993). Rape: The misunderstood crime. Newbury Park, CA: Sage.

Anderson, C. L. (1982). Males as sexual assault victims: Multiple levels of trauma. Journal of Homosexuality, 7, 145-162.

Bochow, M. (1997). Schwule Manner und AIDS [Gay men and AIDS]. Bonn: Deutsche AIDS-Hilfe e.V.

Browne, K. (1994). Child sexual abuse. In J. Archer (Ed.), Male violence (pp. 210-230). London: Routledge.

Coxell, A. W., & King, M. B. (1996). Male victims of rape and sexual abuse. Sexual and Marital Therapy, 11, 297-308.

Coxon, A. (1996). Between the sheets. Sexual diaries and gay men's sex in the era of AIDS. London: Cassell.

Craig, M. E. (1990). Coercive sexuality in dating relationships: A situational model. Clinical Psychology Review, 10, 395-423.

Duncan, D. F. (1990). Prevalence of sexual assault victimization among heterosexual and gay/lesbian university students. Psychological Reports, 66, 65-66.

Hickson, F. C. I., Davies, P. M., Hunt, A. J., Weatherburn, P., McManus, T J., & Coxon, A. P. M. (1994). Gay men as victims of nonconsensual sex. Archives of Sexual Behavior, 23, 281-294.

Himelein, M. J. (1995). Risk factors for sexual victimization in dating. Psychology of Women Quarterly, 19, 31-48.

Hodge, S., & Canter, D. (1998). Victims and perpetrators of male sexual assault. Journal of Interpersonal Violence, 13, 222-239.

Karabatsos, G. (1997). The Sexual Experiences Survey. Interpretation and validity. Journal of Outcome Measurement, 1, 305-328.

Kaufman. A., Divasto, P., Jackson, R., Voorhees, D., & Christy, J. (1980). Male victims: Noninstitutionalized assault. American Journal of Psychiatry, 137, 221-223.

King, M. B. (1992). Male sexual assault in the community. In G. C. Mezey & M. B. King (Eds.), Male victims of sexual assault (pp. 3-12). Oxford, England: Oxford University Press.

Koss, M. E, & Gidycz, C. A. (1985). Sexual Experiences Survey: Reliability and validity. Journal of Consulting and Clinical Psychology, 53, 422-423.

Koss, M. P., Gidycz, C. A., & Wisniewski, N. (1987). The scope of rape: Incidence and prevalence of sexual aggression and victimization in a national sample of higher education students. Journal of Consulting and Clinical Psychology, 55, 162-170.

Koss, M. P., Goodman, L. A., Browne, A., Fitzgerald, L. F., Keita, G. P., & Russo, N. F. (1994). No safe haven: Male violence against women at home, at work, and in the community. Washington, DC: American Psychological Association.

Koss, M. P., & Oros, C. J. (1982). Sexual Experiences Survey: A research instrument investigating sexual aggression and victimization. Journal of Consulting and Clinical Psychology, 50, 455-457.

Krahe, B. (1998). Sexual aggression among adolescents: Prevalence and predictors in a German sample. Psychology of Women Quarterly, 22, 537-554.

Krahe, B., Reimer, T., Scheinberger-Olwig, R., & Fritsche, I. (1999). Measuring sexual aggression: The reliability of the Sexual Experiences Survey in a German sample. Journal of Interpersonal Violence, 14, 91-100.

Krahe, B., Scheinberger-Olwig, R., & Waizenhofer, E. (1999). Sexuelle Aggression zwischen Jugendlichen: Eine Pravalenzerhebung mit Ost-West-Vergleich. [Sexual aggression among adolescents: A prevalence survey including an East-West comparison]. Zeitschrift fur Sozialpsychologie, 30, 165-178.

Krahe, B., Scheinberger-Olwig, R., Waizenhofer, E. & Kolpin, S. (1999). Childhood sexual abuse and revictimization in adolescence. Child Abuse and Neglect, 23, 383-394.

Larimer, M. E., Lydum, A. R., Anderson, B. K., & Turner, A. P. (1999). Male and female recipients of unwanted sexual contact in a college student sample: Prevalence rates, alcohol use, and depression symptoms. Sex Roles, 40. 295-308.

Mezey, G. C., & King, M. B. (1989). The effects of sexual assault on men: A survey of 22 victims. Psychological Medicine, 19, 205-209.

Mezey, G. C., & King, M. B. (Eds.) (1992). Male victims of sexual assault. Oxford, England: Oxford University Press.

Norusis, M. J. (1993). SPSS for Windows. Advanced statistics (Release 6.0). Chicago: SPSS Inc.

Postmes, T., & Spears, R. (1998). Deindividuation and antinormative behavior: A meta-analysis. Psychological Bulletin, 123, 238-259.

Radowsky, M., & Siegel, L. M. (1997). The gay adolescent: Stressors, adaptations, and psychosocial interventions. Clinical Psychology Review, 17, 191-216.

Scarce, M. (1997). Male on male rape. The hidden toll of stigma and shame. New York: Plenum.

Schwartz, M. D., & DeKeseredy, W. S. (1997). Sexual assault on the college campus. The role of male peer support. Thousand Oaks, CA: Sage.

Stermac, L., Sheridan, P. M., Davidson, A., & Dunn, S. (1996). Sexual assault of adult males. Journal of Interpersonal Violence, 11, 52-64.

Turner, S. (1992). Surviving sexual assault and sexual torture. In G. C. Mezey & M. B. King (Eds.), Male victims of sexual assault (pp. 75-86). Oxford, England: Oxford University Press.

Tyler, K. A., Hoyt, D. R., & Whitbeck, L. B. (1998). Coercive sexual strategies. Violence and Victims, 13, 47-61.

Ullman, S. E., Karabatsos, G., & Koss, M. P. (1999). Alcohol and sexual assault in a national sample of college women. Journal of Interpersonal Violence, 4, 603-625.

Waterman, C. K., Dawson, L., & Bologna, M. J. (1989). Sexual coercion in gay male and lesbian relationships: Predictors and implications for support services. The Journal of Sex Research, 26, 118-124.

Manuscript accepted December 8, 1999

APPENDIX

HOMOSEXUAL EXPERIENCES SURVEY (HSES)

Victimization Items

"Has a man ever made you have sex with him against your will by using physical force or threatening to do so?"

"Has a man ever made you have sex with him against your will by exploiting the fact that you could not resist (e.g. after too much alcohol or drugs)?"

"Has a man ever made you have sex with him against your will by using verbal pressure?"

Perpetration Items

"Have you ever made a man have sex with you against his will by using physical force or threatening to do so?"

"Have you ever made a man have sex with you against his will by exploiting the fact that he could not resist (e.g. after too much alcohol or drugs)?"

"Have you ever made a man have sex with you against his will by using verbal pressure?"

EXAMPLE OF RESPONSE FORMAT

Has a man ever made you have sex with him against your will by using physical force or threatening to do so?

[] No

[] Yes ....

... my (ex-)partner:

[] Touching (Kissing, stroking)

[] Masturbating ("Wanking")

[] Oral sex ("Sucking")

[] Anal sex ("Fucking")

... a friend or acquaintance:

[] Touching

[] Masturbating

[] Oral sex

[] Anal sex

... an unknown man:

[] Touching

[] Masturbating

[] Oral sex

[] Anal sex

Barbara Krahe, Stephan Schutze, Immo Fritsche, and Eva Waizenhofer University of Potsdam, Germany

The study was supported by a grant from the Deutsche Forschungsgemeinschaft to the first author. The authors would like to thank Anja Berger, Andreas Haubner, Susanne Kolpin, Ingrid Moller, and Renate Scheinberger-Olwig for their support in conducting this study. The help and advice of Bastian Finke and his team at Mann-O-Meter e.V. Berlin is also gratefully acknowledged.

Address correspondence to Barbara Krahe, University of Potsdam, Department of Psychology, Postfach 60 15 53, D-14415 Potsdam, Germany; e-mail: [e-mail].