Childhood Sexual Experiences with an Older Partner among Men Who Have Sex with Men in Buenos Aires, Argentina

Abstract

This study sought to describe childhood sexual experiences with older partners (CSEOP) among men who have sex with men (MSM) in Buenos Aires, Argentina. MSM were recruited through Respondent Driven Sampling. They responded to a computer administered self-interview with questions on CSEOP, operationalized as manual, oral, genital, or anal contact prior to age 13 with a partner at least 4 years older. Of the 500 respondents, only 25% identified as gay. Eighteen percent of the respondents reported CSEOP, the majority of whom did not feel they were hurt by the experience and did not consider it to be childhood sexual abuse (CSA). Over two-thirds of MSM who reported CSEOP said that their older partner was a female. Only 4% of those with a female partner felt their experience was CSA compared to 44% of those who had a male partner. Among all men reporting CSEOP, those who felt sexually abused were more likely to have been physically forced or threatened, physically hurt, and emotionally hurt than those who did not feel sexually abused. Having CSEOP, being hurt by the experiences, and perceiving the experiences as sexual abuse were not associated with current HIV sexual risk or substance use behavior. In this sample of MSM in Argentina, a substantial minority reported CSEOP. Those who felt they had been sexually abused were much more likely to have had an older male partner than an older female partner, and were more likely to report having been physically forced and threatened by their older partner.

Introduction

Childhood sexual experiences with older partners (CSEOP) are frequently studied to identify possible correlates with behaviors, emotions, and attitudes in adulthood. Some of CSEOP can be considered childhood sexual abuse (CSA). Yet, the lack of a shared definition of the latter often hinders comparisons across studies and the drawing of definite conclusions about its potential effects. For this study, we operationalized CSEOP as any sexual contact (manual, oral, genital, or anal) prior to age 13 with a partner who was at least four years older. This operationalization includes a subset of experiences that we labeled CSA, defined as a CSEOP in which the child felt emotionally or physically hurt (for a detailed discussion on the topic and rationale for our operationalization, see Carballo-Diéguez, Balan, Dolezal, & Mello, 2012).

Gay men have been found to be much more likely than heterosexual men to have experienced unwanted sexual encounters in childhood (27% vs. 1%; Finlinson et al., 2003) and to have had childhood sexual abuse involving penetration (22% vs. 6%; Balsam, Rothblum, & Beauchaine, 2005. Balsam et al. (2005) hypothesized that young men who appear to be gender non-conformant may be specifically targeted for sexual abuse because of their perceived sexual orientation.

Among MSM, a history of CSA has been associated with a wide range of negative outcomes in adulthood. Many of these outcomes are behaviors that may increase the risk of HIV transmission: unprotected receptive anal sex, unprotected sex with a serodiscordant partner, commercial sex trade, and injection drug use (Bartholow et al., 1994; Brennan, Hellerstedt, Ross, & Welles, 2007; Carballo-Diéguez & Dolezal, 1995; Kalichman, Gore-Felton, Benotsch, Cage, & Rompa, 2004; Mimiaga et al., 2009). In fact, HIV prevalence has been found to be higher among men reporting CSA than those who do not (Brennan et al., 2007; Kalichman et al., 2004; Lloyd & Operario, 2012). CSA among MSM has also been associated with mental health problems, such as depression, psychological distress, suicidality, eating disorders, drug use, and requiring hospitalization for mental health conditions (Balsam, Lehavot, & Beadnell, 2011; Bartholow et al., 1994; Feldman & Meyer, 2007; Mimiaga et al., 2009; Ratner et al., 2003).

The long list of negative outcomes emphasizes the serious consequences such experiences may have for children, even into adulthood. Yet, many children seem to survive CSA without developing such problems. In a meta-analysis of 59 studies, Rind, Tromovitch, and Bauserman (1998) concluded that negative consequences were not pervasive in most studies and that men, compared to women, were much less likely to report negative reactions (see also Dallam et al., 1998 and Ondersma, Chaffin, Berliner, Cordon, & Goodman, 1998, for a critique of Rind's article.) Stanley, Bartholomew, and Oram (2004) examined psychological adjustment in a sample of MSM and found that men who reported experiences that were non-coercive and not perceived as abusive did not report poorer adjustment, and those who had coercive or abusive experiences did report poorer adjustment. Arreola, Neilands, Pollack, Paul, and Catania (2008) found childhood sexual experiences to be associated with depression and suicidal ideation but only among those who reported being forced. Those who felt the experiences were consensual did not differ from those who did not have childhood sexual experiences on depression and suicidality, although they did report more HIV transmission risk. Jinich et al. (1998) also found that the perception of having been coerced was associated with higher rates of sexual risk behavior.

In two studies of Latino MSM living in the United States (Carballo-Diéguez & Dolezal, 1995; Dolezal & Carballo-Diéguez, 2002), sexual risk behavior in adulthood was associated with CSEOP and was more frequent among those who were coerced, felt physically or emotionally hurt, or who felt their experiences constituted sexual abuse. In the later study, among MSM reporting CSEOP, those who perceived it as sexual abuse were younger at the time it happened and were more likely to report that they were forced, physically hurt, emotionally hurt, and/or threatened. Therefore, the study of early sexual experiences should take into consideration the characteristics of those experiences given that not all experiences are equally traumatic or coercive.

Several studies of MSM in the United States have found higher rates of CSA among Latinos, compared to non-Latino white participants (Arreola, Neilands, Pollack, Paul, & Catania, 2005; Balsam, Lehavot, Beadnell, & Circo, 2010; Doll et al., 1992; Newcomb, Munoz, & Carmona, 2009; Parsons, Bimbi, Koken, & Halkitis, 2005). However, little is known about the prevalence and severity of CSEOP or CSA among MSM in Latin American countries. A recent study of CSEOP among MSM in Brazil (Carballo-Diéguez, Balan, Dolezal, & Mello, 2012) found that 29% of non-transgender participants had sexual contact prior to age 13 with a partner that was at least four years older, and 39% of those men reported that they were physically or emotionally hurt by the experience. Those with CSEOP reported more male sex partners than those who did not have childhood sexual experiences, but their rates of unprotected sex were not significantly different. Among those who had CSEOP, those who considered it to be sexual abuse had more unprotected insertive anal sex than those who did not, but the groups did not differ on frequency of unprotected receptive sex or number of male sex partners. Of note, Carballo-Diéguez et al. (2012) and Parker (1991) discussed the importance of taking into account cultural factors that may value differently insertive versus receptive anal intercourse, as well as popular childhood games, such as the meia or troca-troca played by boys in which partners are said to take turns masturbating, fellating or, most commonly, penetrating one another. Parker (1991) stated that:

“Just as older males instruct younger ones in the intricacies and techniques of lovemaking with women, and often even arrange for heterosexual initiation, in meia or troca-troca, the older males offer an initiation into homosexual practices by symbolically feminizing their partners. As one frequently cited expression puts it, Homem, para ser homem, tem que dar primeiro - in other words, `A man, in order to be a man, has to give (to take the passive role in anal intercourse) first.'” (p. 128)

Unfortunately, there is a dearth of similar cultural analysis of CSEOP conducted in other Latin American countries.

Given the insufficient information on CSEOP among MSM in Latin American countries, we explored the issue as part of a larger study designed primarily to estimate the prevalence of HIV, STIs, and sexual risk behavior among MSM in Buenos Aires, Argentina (Carballo-Diéguez et al., 2011; Pando et al., 2012). Our research questions were: (a) What is the prevalence estimate of CSEOP and CSA among MSM in Argentina? (b) What are the characteristics of those experiences? (c) Are CSEOP and CSA associated with sexual risk behavior and substance use in adulthood? and (d) To what extent do participants equate their CSEOP with CSA?

Methods

We used Respondent Driven Sampling (RDS; Heckathorn, 1997, 2002) to recruit MSM for this study. Recruitment procedures are described in detail elsewhere (Carballo-Diéguez et al., 2011). In brief, research staff identified 16 MSM (seeds) who purportedly had social networks that could allow them to refer other “men like them” to the study. After being interviewed, respondents were given three coupons each (with an ID number and study contact information) and were asked to distribute the coupons to peers. An incentive was given to each respondent whose referrals contacted us, until the target sample of 500 was reached. Eligibility criteria were (a) 18 years old or older, (b) self-identify as a man, (c) sex with a man or male-to-female transvestite at least 10 times in their life and at least once in the past 6 months, and (d) willing to provide a blood sample for HIV/STI testing. The research study received approval from the Institutional Review Boards of the New York State Psychiatric Institute in the United States and the Comité Independiente de Ética en Investigación de la Facultad de Medicina at the Universidad de Buenos Aires, in Argentina. Written consent was obtained from all participants prior to enrollment in the study.

Measures

Participants completed a web-based, computer-assisted self-interview (CASI) on computers at the study office. The survey covered many topics, including demographics, current sexual behavior, substance use, and childhood sexual experiences. Current (past two months) sexual behavior included the frequency of unprotected receptive (with a male or transsexual partner) and insertive (with a male, transsexual, or female partner) anal intercourse and the number of male sex (oral and/or anal) partners. Substance use was assessed separately for 15 substances, and frequency of use in the previous 2 months was measured with a 7-point scale ranging from never to more than once a day.

CSEOP was assessed by a detailed interview that began by asking “¿Qué edad tenía cuando tuvo su primer contacto sexual? (manual, oral, genital, o anal)” (“How old were you when you had your first sexual contact of any kind [manual, oral, genital, or anal]?”). Note that a detailed sexual interview occurred earlier in the survey, where definitions of these sexual activities were provided. If the participant reported engaging in sexual contact prior to age 13, they were asked “Aproximadamente, ¿qué edad tenía la otra persona?” (“Approximately, how old was the other person?”). A follow-up question was also used to determine whether the participant had had any sexual contact prior to age 13 with a partner that was four or more years older than he was. If the participant had more than one older partner prior to the age of 13, the instructions told the participant to answer the questions that followed making reference to the first older partner they had. The age of 13 and four-year age difference were selected to identify those who had sexual contact at a young age and to eliminate sexual experiences with peers. Although sexual experiences with partners close in age can also be abusive, for the purposes of this study, we wanted to examine specifically those experiences which involved men or women who were substantially older. If men reported having such a partner, they were asked several questions about the characteristics of that partner (age, gender, relationship) and what behaviors took place. Participants were also asked whether physical threats or force were used, whether they were emotionally or physically hurt, and whether they considered the experience sexual abuse. The questions were carefully drafted to avoid judgment values (e.g., we did not refer to the child as a victim or the partner as a perpetrator), and they were computer administered to minimize drifting that could have been introduced in face-to-face interviews. Participants were asked only at the end of the questionnaire if they considered the experience as abuse. This survey is virtually identical to the one used in Brazil that is published in the appendix to Carballo-Diéguez et al., (2012) and was also used in previous studies of Latino MSM in New York (Carballo-Diéguez & Dolezal, 1995; Dolezal & Carballo-Diéguez, 2002). For the purpose of these analyses, participants who reported CSEOP were further divided into those who felt they had been hurt by the experience and those who did not and also into those who felt they were the victims of sexual abuse and those who did not.

Data Analysis

Because this sample was recruited using RDS, all data were weighted prior to analyses using an RDS II estimator (Salganik & Heckathorn, 2004; Volz & Heckathorn, 2008). This approach gives greater weight to men with a small personal network size (PNS), as presumably those men are less likely to be recruited into the study. Weights were calculated as the inverse of the participant's PNS. This value was then multiplied by the sample size (N) divided by the sum of weights (Σw). The following weighting formula reflects the original sample size of 500:

$${(1∕\text{PNS})}^{\ast }(\mathrm{N}∕\Sigma \mathrm{w})$$

Note that when data are weighted, the Ns produced in different analyses are often slightly discrepant. Therefore, the tables report only percentages to avoid confusion.

Statistical tests compared groups on the characteristics of their experiences, substance use, and sexual risk behavior. The number of childhood sex occasions and current sexual risk behavior variables were log-transformed prior to analyses due to skewed distributions. T-tests for continuous variables and Fisher's exact tests for dichotomous variables were used for 2-group comparisons: participants with a female versus male partner and participants who felt they were abused versus those who did not. ANOVAs were used for three-group comparisons: participants who did not have any sexual contact prior to age 13 with a partner who was at least four years older (no CSEOP), participants who reported CSEOP but did not feel either physically or emotionally hurt by the experience, and participants who reported CSEOP and did feel physically or emotionally hurt. In order to identify findings that were affected by data weighting, all statistical comparisons were repeated using unweighted data. Any differences are reported in the results.

Results

The RDS sample

Of the 16 seeds, seven did not recruit any participants and four recruited only one. By the end of recruitment, 89% of the participants could be traced back to a single seed after 22 waves of referrals. Participant's personal network size ranged from 1 to 50, with a median of 3. Further details and discussion on this RDS recruited sample appear in Carballo-Diéguez et al., 2011.

A description of the sample is found in Table 1. On average, men were slightly over 30 years old and most had not graduated from high school. Only 25% of the sample self-identified as gay or homosexual, with over a third identifying as bisexual. Sexual identity was overall consistent with participants' behavioral reports, with over two-thirds of men reporting sex with a woman in the past year (not shown).

Table 1.

Sample description (N=500)

	M (SD) range
Age (years)	30.5 (11.5) 18–77
Educationa	4.3 (1.7) 1–9
	%b
Self-label
Gay / Homosexual	25
Bisexual	36
Heterosexual	22
Otherc	17
Tested HIV+	17
Childhood sexual experience with an older partner
No sex pre-13 with 4-year older partner	82
Had sex, but not physically or emotionally hurt	14
Had sex, and was physically or emotionally hurt	4
Older partner's relationship to childd
Male non-relative	24
Father	1
Step-father	1
Brother	1
Uncle	1
Male cousin	9
Female non-relative	50
Aunt	2
Female cousin	14

^aEducation measured on a 9-point scale from 0 = none to 9 = completed college (4=partial high school).

^bPercents are of those who have non-missing data.

^cFor example, “hombre” (man) or “activo” (active).

^dPercents are of the 90 men who report childhood sexual experiences with an older partner. Due to weighting, some categories have percentages less than 1 and have been rounded up. Therefore the percentages sum to greater than 100.

Prevalence and Characteristics of Childhood Sexual Experiences

Eighteen percent of the men reported having sexual contact prior to age 13 with a partner who was at least four years older (Table 1). Most of the older partners were non-relatives, and approximately two-thirds of the men reported that the older partner was a female.

Table 2 shows descriptive details of the early childhood experiences. On average, the experiences took place when the participants were 10 years old with a partner who was 20. One man reported 1,000 sex occasions with his partner, and although this seems high, the response was consistent with his other responses, which indicated a very sexually active childhood. In terms of the behavior involved, manual contact was quite common, and oral sex was reported by approximately half of the men. Most men (84%) who had a male partner reported engaging in receptive anal sex, and all of the participants with a female partner reported engaging in vaginal sex. Most men with a history of CSEOP did not report that they had been forced, threatened, or hurt, and only 18% felt that the experience was sexual abuse.

Table 2.

Characteristics of childhood sexual experiences prior to age 13 with a partner at least 4 years older

	Total Sample (n = 90)		Female Partnera (n = 59b)		Male Partnera (n = 32b)
	M (SD)	range	M (SD)	range	M (SD)	range	p c
Age at first experience	10.2 (2.7)	1–12	10.6 (2.7)	1–12	9.4 (2.6)	4–12	.049
Partner's age	20.1 (7.7)	10–48	19.2 (5.8)	14–40	21.9 (10.2)	10–48	ns
Number of sexual occasionsd	19.5 (84.7)	1–1000	20.3 (102.5)	1–1000	17.9 (20.9)	1–150	ns
	%e		%e		%e		p c
Sexual behavior
Partner exposed genitals	84		86		79		ns
Child exposed genitals	54		57		47		ns
Fondled by partner	94		97		89		ns
Child fondled partner	82		95		59		<.001
Deep kissing	52		67		26		.001
Child masturbated partner	57		50		70		ns
Masturbated by partner	70		81		51		.005
Performed fellatio	23				67		na
Received fellatio	51		66		23		<.001
Receptive anal intercourse	29				84		na
Insertive anal intercourse	42		56		19		.001
Cunnilingus	30		46				na
Vaginal sex	65		100				na
Other characteristics of experiences
Was physically forced	13		6		25		.014
Was threatened	13		5		27		.007
Was emotionally hurt	18		14		26		ns
Was physically hurt	13		11		18		ns
Consider experience sexual abuse	18		4		44		<.001

^aThese columns compare those whose first older partner was a female vs. a male.

^bDue to weighing, the subsamples sum to 91, rather than 90.

^cP-values from t-tests (continuous variables) or Fisher's exact tests (dichotomous variables).

^dValues were log-transformed prior to t-tests due to a skewed distribution

^ePercents are of those who have non-missing data

ns = p > .05; na = not applicable

Characteristics Associated With the Gender of the Older Partner

Because a large proportion of the men in the sample reported that their partner had been a female, we compared the characteristics of the experiences of participants with a female partner with men with a male partner (Table 2). Those participants with a female partner were slightly over a year older when the experience happened, and this difference is statistically significant. Male partners were slightly older than female partners, but this difference was not statistically significant. The range for the number of occasions is large, with one participant with a female partner reporting 1,000 sex occasions. The median number of occasions was five for the total sample and for those participants with a male partner and three for those with a female partner. In the total sample, 22% report that they only had one sex occasion with that partner, and 80% report 10 or fewer occasions. Apart from the one participant reporting 1,000 occasions, only four others report high numbers, ranging from 100–200.

Several behaviors were more common when the older partner was a female: fondling the older partner, deep kissing, being masturbated, receiving fellatio, and insertive anal intercourse. Those with a female partner were significantly less likely to report that they were physically forced or threatened, although there were no differences between the two groups on whether they felt physically or emotionally hurt. Only 4% of those with a female partner felt that the experience was sexual abuse, compared to 44% of those who had a male partner. Only one man with a female partner identified as gay at the time of the interview, compared to 75% of those with a male partner (not shown).

Association of Childhood Sexual Experiences and the Perception of Sexual Abuse With Adult Sexual Risk Behavior and Substance Use

Participants who reported no CSEOP, reported CSEOP but did not feel hurt by the experience, and reported feeling they were hurt by the experience were compared on their current sexual risk behavior and substance use (Table 3). Alcohol, marijuana, and cocaine are included because they were the only substances used by more than 20% of the sample. No statistical differences were found. Likewise those who felt the experience was sexual abuse did not differ from those who did not think it was abuse on current sexual risk behavior and substance use.

Table 3.

The association of childhood sexual experiences with an older partner (CSEOP) and perceived childhood sexual abuse with adult sexual risk behavior and substance use^a

	No CSEOP (n = 398)	CSEOP but not hurt (n = 70)	CSEOP and hurt (n = 20)	Not Abuse (n = 73)	Abuse (n = 17)
	M (SD)	M (SD)	M (SD)	M (SD)	M (SD)
Unprotected receptive anal intercourseb	1.1 (5.2)	2.5 (9.6)	1.3 (3.6)	2.0 (8.7)	3.2 (8.9)
Unprotected insertive anal intercourseb	2.8 (7.2)	4.9 (10.9)	3.2 (9.5)	4.3 (10.0)	5.5 (12.9)
Male partnersb	8.3 (49.1)	6.3 (7.9)	7.8 (8.8)	6.8 (8.2)	6.0 (7.5)
Alcohol usec	3.7 (1.8)	3.7 (1.6)	4.0 (2.0)	3.7 (1.7)	4.1 (1.6)
Marijuana usec	2.7 (2.3)	2.7 (2.5)	2.9 (2.1)	2.9 (2.5)	1.9 (1.7)
Cocaine usec	1.8 (1.4)	1.8 (1.4)	2.1 (1.8)	1.9 (1.7)	1.9 (1.8)

^aANOVA tests for differences among the 3 CSEOP groups, and t-tests for differences between the 2 abuse groups were all non-significant.

^bValues were log-transformed prior to ANOVAs due to skewed distributions.

^cSubstance use, past 2 months, measured on a 7-point scale from 0 = `never' to 7 `more than once a day'.

Characteristics Associated With a Perception of Sexual Abuse

The characteristics of childhood sexual experiences were compared for participants who felt they were sexually abused and participants who did not feel they were abused (Table 4). Participants who felt they were abused were much more likely to report that they were physically forced, threatened, and emotionally and physically hurt. They were also significantly more likely to report that their older partner was male. Those who felt they were abused were less likely to report that they had engaged in vaginal or anal sex, although rates of these behaviors were high for both groups. The two groups did not differ on their age when it happened, the age of their older partner, or the number of times they had sex with that person.

Table 4.

Characteristics associated with a perception of sexual abuse

	Not Abuse (n = 74a)	Abuse (n = 17a)
	M (SD)	M (SD)	p b
Age at first experience	10.3 (2.7)	9.4 (2.2)	.187
Partner's age	17.7 (6.5)	20.5 (9.6)	.271
Number of sexual occasionsc	20.1 (92.8)	16.5 (28.1)	.625
	%	%	p b
Partner was female	75	18	<.001
Was physically forced	5	50	<.001
Was threatened	1	59	<.001
Was emotionally hurt	12	44	.007
Was physically hurt	8	35	.009
Anal or vaginal sex	97	82	.043

^aDue to weighing, the subsamples sum to 91, rather than 90.

^bP-values from t-tests (continuous variables) or Fisher's exact tests (dichotomous variables).

^cValues were log-transformed prior to t-tests due to a skewed distribution.

Weighted Versus Unweighted Results

All of the statistical tests in Tables 2–4 were also conducted without weighting the data. Only four differences were noted. In Table 2, the t-test for the participant's age becomes nonsignificant, and the t-test for the partner's age becomes significant (p = .047). The Fisher's exact test for higher rates of being emotionally hurt among those with an older male partner was significant (p = .047). In Table 3, the 3-group ANOVA comparing CSEOP groups on frequency of unprotected receptive anal sex was significant (p = .049). All of these tests involve p-values that are extremely close to .05, so weighted and unweighted results are very similar.

Discussion

It is typically very difficult to compare data on childhood sexual experiences across studies because assessments and definitions vary dramatically from study to study (Purcell, Malow, Dolezal, & Carballo-Diéguez, 2004). One advantage of this study is that it used the same survey questions and the same definition of CSEOP as two other studies of Latino MSM: a convenience sample of Latino MSM living in the New York City [NYC] area (Dolezal & Carballo-Diéguez, 2002) and an RDS sample of MSM in Brazil (Carballo-Diéguez et al., 2012). When compared to those two studies, the results from these MSM in Buenos Aires, Argentina, are noticeably different. These differences indicate that the concepts of CSA and even MSM may vary widely from country to country.

First, 18% of the men in Argentina reported some sexual contact before age 13 with a partner who was at least four years older. Although this percentage represents a substantial minority, this prevalence rate is relatively low when compared to the other two studies (33% in NYC and 29% among the male-identified participants in Brazil; rates were even higher among the transsexual-identified participants in that study). Furthermore, among those who reported CSEOP, a smaller proportion of the Argentine men considered their experiences to be childhood sexual abuse (18%, compared to a third of the Brazil sample and 59% of the NYC sample). We have demonstrated that this low perception of abuse seems largely due to the gender of the older partner. Among the MSM in Argentina, 66% reported that the partner was a female, which was rare in the other two studies (7% among male-identified MSM in Brazil and 9% in NYC).

These differences demonstrate that one cannot assume homogeneity of populations from country to country when recruiting MSM. For example, the NYC sample of Latino MSM was predominately gay-identified and reported limited sexual experience with women. Although the case was the same in Brazil, among our RDS sample of Argentinian MSM, only 25% considered themselves gay and over two thirds reported also having sex with women in the past year. These studies were not based on probability sampling, so we cannot claim that they are representative of all MSM in each country, but sampling based on the criteria of having sex with men can result in dramatically different samples of men with very difference experiences and behaviors.

Therefore, keeping the uniqueness of this sample in mind, we turn to the relatively low prevalence rate of CSEOP, the importance of the gender of the older partner, and the lack of any associations between CSEOP and current risk behaviors. We have noted that 18% of these men reported CSEOP. Although this rate is lower than found in other studies, it is still a substantial minority of this population. It would be wrong to conclude that early childhood experiences with older partners are rare among Argentine MSM because nearly a fifth of this sample report having had such experiences. Therefore, any outreach or intervention to MSM in Argentina should assume that a substantial number of these men will have had such early sexual experiences. Although most men did not report traumatic experiences, several men report being physically forced, physically hurt, and sexually abused. Consequently, the relatively low prevalence rate should not detract from the fact that among MSM in Argentina, there are several who have experienced painful and abusive experiences as children, with presumably some negative consequences for their subsequent development and mental health.

Nevertheless, the majority of the men who had CSEOP did not seem to have a negative perception of their experiences, and this finding is partially explained by the gender of the older partner. Compared to those who did not feel they were abused, those who felt abused were more likely to report being physically forced and/or threatened. Both of these were more common among participants with a male older partner. As a result, 44% of participants with an older male partner felt they were sexually abused, compared to only 4% of those with an older female partner. This may also be useful information for those working with predominately gay-identified men who are less likely to have had sex with women. Traumatic experiences and feelings of abuse would clearly be more common among those men compared to bisexual or heterosexually identified men.

Cultural values may also be at play here. A man can interpret sexual contact with a female partner, even if it is an early experience, as falling within normative behavior and therefore not consider it abuse, but the situation is not parallel for those who had an older male partner. The latter may experience sexual contact with men as socially stigmatized behavior and therefore be more likely to interpret it as abusive.

In this study, we found no evidence that having early sexual experiences with an older partner or the severity of those experiences (being hurt or feeling they were abused) were associated with current sexual risk behavior or substance use. This is in contrast to other studies with Latino men conducted in the United States (Carballo-Diéguez & Dolezal, 1995; Dolezal & Carballo-Diéguez, 2002). However, among the sample of Brazilian MSM (Carballo-Diéguez et al., 2012), there was also no association found between CSEOP and adult risk behavior. Our results in Argentina may be the result of a lack of power for such analyses because relatively few men in this sample reported being hurt or abused. Yet, the concordance with the findings from Brazil may indicate further cultural issues that shape the construct of childhood sexual abuse. If early sexual experiences are more accepted by cultural norms, they may be less likely to cause the psychological stress that can lead to substance use or risky sexual behavior. An exploration of these cultural issues was beyond the scope of this study, but clearly further research is warranted.

Limitations

One limitation of these analyses may be the sample size. The original sample size of 500 was established based on power analyses that, at the time of study proposal, specified that a design effect of 2.0 was sufficient. In later publications, Salganik (2006) suggested that, in some simulations, RDS may require a design effect of 10 or higher. Therefore, caution should be used in interpreting these findings as representative of true population values. Furthermore, the use of RDS may have influenced the composition of the sample in various other ways. For example, incentives for recruiting others may have been more of a motivation for men with lower income, recruitment may oversample some subpopulations if several men begin to recruit only within a narrow social group, and eligibility criteria may be inadvertently disclosed to recruiters leading to the participation of ineligible men. Thus, it may be that the high proportion of bisexual- and heterosexual-identified men in this sample is the result of our sampling method.

Statistically, weighting based on a self-reported network size may underestimate the contribution of those with larger networks. Fortunately, in these analyses, weighted and unweighted results were very similar (see Carballo-Diéguez et al. [2011] for a more thorough discussion of the implications of RDS methods.)

Finally, because the characteristics of this sample and their childhood sexual experiences seem to be in contrast to other Latino MSM samples, there is a question as to whether any of these differences are related to cultural norms and values that may be unique to Argentina. Unfortunately, our survey did not explore this idea directly; further studies would be needed to see whether cultural differences exist and how they might moderate men's understanding of and reaction to childhood sexual experiences.

Conclusion

This study presents a unique sample of MSM in Buenos Aires, Argentina, among whom 25% identify as gay and sexual activity with women is common. This finding demonstrates that recruiting based on MSM in Latin American countries can result in very different kinds of men with varying experiences. CSEOP or CSA among MSM may also look quite different from country to country, both in terms of their prevalence and their characteristics.

The bisexual nature of this sample allowed for an interesting contrast between those who had a male versus a female older partner. Our findings show that the gender of the older partner is an important feature distinguishing their experiences. Surveys of childhood sexual experiences should include questions such as partner characteristics and coercion, as they are associated with the perception of sexual abuse among these MSM.

Finally, the fact that most of these men did not have negative reactions to their experiences provides another demonstration that there is a broad spectrum of childhood sexual experiences, including those that are perceived as positive and non-abusive. Furthermore, the lack of associations between CSEOP and adult risk behavior calls into question whether the construct of CSA as it is generally understood in North America and Europe is applicable in Latin American countries.