Recalled Sexual Experiences in Childhood with Older Partners: A Study of Brazilian Men Who Have Sex with Men and M-T-F Transgender Persons

Abstract

This study assessed the prevalence of recalled childhood sexual experiences with an older partner among men who have sex with men (MSM) and/or male-to-female transgender persons recruited in Campinas, Brazil. It also analyzed associations between such recalled experiences and sexual risk behavior in adulthood. Participants recruited using respondent driven sampling completed a self-administered, computer-based questionnaire, and underwent HIV testing. For data analysis, raw scores were weighted based on participants’ reported network size. Of 575 participants (85% men and 15% transgender), 32% reported childhood sexual experiences with an older partner. Mean age at first experience was 9 years, partners being on average 19 years old, and mostly men. Most frequent behaviors were partners exposing their genitals, mutual fondling, child masturbating partner, child performing oral sex on partner, and child being anally penetrated. Only 29% of the participants who had had such childhood sexual experiences considered it abuse; 57% reported liking, 29% being indifferent, and only 14% not liking the sexual experience at the time it happened. Transgender participants were significantly more likely to report such experiences and, compared with men, had less negative feelings about the experience at the time of the interview. No significant associations were found between sexual experiences in childhood and unprotected receptive or insertive anal intercourse in adulthood. Results highlight the importance of assessing participants’ perception of abuse, regardless of researchers’ pre-determined criteria to identify abuse. MSM and transgender people may experience childhood sexual experiences with older partners differently from other populations (e.g., heterosexuals), particularly in countries with different cultural norms concerning sexuality than those prevalent in Europe and the U.S.

INTRODUCTION

Defining childhood sexual abuse is an arduous task, and researchers have operationalized the construct in many different ways (for reviews, see Holmes & Slap, 1998; Putnam, 2003; Rind, Tromovitch, & Bauserman, 1998). Although the extreme cases in which children are deceived, degraded or hurt for the sexual enjoyment of an adult constitute clear cases of abuse, as one zooms out from that focus and considers the vast variety of childhood sexual experiences with older partners the need for operationalization of the construct becomes more evident. There is a need to define who is considered a child, what sexual behaviors should be considered, who constitutes an older partner, what is attributed to the age difference between partners, and the role played by gender. There is lack of agreement in the field with respect to these issues. For example, Rind et al. (1998) suggest distinguishing children’s experiences from those of adolescents, but this is seldom done and frequently anybody under age 18 is considered a child; also, there is lack of agreement concerning the meaning of age difference between the child and the adult (see Card, 2002) and there are contrasting opinions concerning the role that gender plays in childhood sexual experiences with older partners (see Bauserman & Rind, 1997; Dube et al., 2005). Furthermore, the role played by social and cultural norms, tradition, religion, and the law of the land, among other factors, in attributing meaning to the sexual act(s) is seldom considered (Crossette, 1999). Last but not least, insufficient attention has been paid to the perception of those involved in the events (who may or may not consider the experience abuse; e.g., see Carballo-Diéguez & Dolezal, 1995; Sandfort, 1982), and the degree to which those perceptions can be said to be undistorted by guilt, denial, reinterpretation of the past, or other mechanisms.

If one wants to establish how prevalent childhood sexual experiences with older partners are in society, further challenges appear, such as how to sample the “hidden” population likely to have had the target sexual experiences, how to ask about such experiences without biasing responses (Finkelhor, Hotaling, Lewis, & Smith, 1990; Hulme, 2004), and how to overcome the natural resistance to disclose a stigmatized behavior.

It has been pointed out that advantages could be derived from reaching consensus on a definition of childhood sexual abuse (Dolezal & Carballo-Diéguez, 2002; Holmes & Slap, 1998; Rind et al., 1998) and some valiant attempts have been made to review the field in search of common ground (Basile & Saltzman, 2002; Holmes & Slap, 1998) or to evaluate instruments for the assessment of childhood sexual abuse across studies (Hulme, 2004). Yet, consensus remains elusive. Although it could be argued that the absence of a dogmatic definition may favor rich and diverse research around a central theme, comparisons across studies are extremely difficult to make in the current circumstances.

Despite the lack of a shared definition, there is a growing body of literature focusing on men who have sex with men (MSM) who experienced either childhood sexual experiences with older partners or outright sexual abuse. Studies have found associations between those early sexual experiences and higher likelihood of psychiatric disorders, substance use, suicidality, and high risk sexual behavior (Briere & Elliott, 2003; Carballo-Diéguez & Dolezal, 1995; O’Leary & Gould, 2008; Paul, Catania, Pollack, & Stall, 2001). These findings have led researchers to propose interventions to decrease negative sequelae of childhood sexual abuse in males (Sikkema et al., 2007; Williams et al., 2008).

Keeping the above-mentioned constraints in mind, we studied the childhood sexual experiences with older partners and childhood sexual abuse reported by two groups: In one group, MSM; and in the other group, participants born male who identified as women, transvestites, or male-to-female transsexuals, for brevity here referred to as transgender persons. Our study sought to answer some of the questions still open despite the growth observed in the scientific literature in this field in recent decades. For example, numerous studies have reported differences in childhood sexual abuse across racial and ethnic groups, with higher rates among African American and Latino MSM in the U.S. (Arreola, Neilands, Pollack, Paul, & Catania, 2005; Holmes & Slap, 1998; Jinich et al., 1998; Mimiaga et al., 2009; Parsons, Bimbi, Koken, & Halkitis, 2005), yet little has been published that explores in a detailed manner childhood sexual experiences with older partners within these groups (Dolezal & Carballo-Dieguez, 2002). Second, few published studies on childhood sexual experiences with older partners and childhood sexual abuse among MSM have been conducted outside Western European countries or North America (Bauserman & Rind, 1997). Third, while the literature on childhood sexual abuse has broadened from an original focus on girls’ experiences to encompass boys’ experiences and those of boys who later identify as gay or bisexual men, there are few studies (e.g., Finlinson, Colón, Robles, & Soto, 2008) on childhood sexual abuse and childhood sexual experiences with older partners of boys who later identify not as men but rather as women, transvestite, transsexual or transgender. Finally, insufficient attention has been paid to individuals’ perceptions of their own early sexual experiences with older partners, especially whether they label them as abuse or not (Carballo-Diéguez & Dolezal, 1995; Dolezal & Carballo-Diéguez, 2002; Holmes, 2008; Stanley, Bartholomew, & Oram, 2004).

Our research questions were: (1) What is the prevalence of childhood sexual experiences with older partners among MSM and transgender persons in an urban location in Brazil? (2) What proportion of those childhood sexual experiences with older partners were perceived as physically or emotionally hurtful by the participants? (3) What proportion of individuals reporting childhood sexual experiences with older partners considered that the events constituted childhood sexual abuse?, and (4) What significant correlates, if any, exist between hurtful childhood sexual experiences with older partners as well as self-defined childhood sexual abuse and adult sexual risk behavior that may lead to HIV infection? For our study, we operationalized childhood sexual experience with an older partner as any manual, oral, genital, or anal sexual contact between a male younger than 13 years of age and a partner at least four years older.

Several reasons led us to choose these parameters. First, although undoubtedly some sexual experiences that do not include sexual contact (e.g., witnessing a rape) can have traumatizing consequences for a child, in general it is considered that experiences that involve physical contact are more likely to carry sequelae than those with no physical contact (Collings, 1995; Rind et al., 1998); therefore, we chose to inquire only about the former. Second, our focus was on children, not adolescents. There is a tendency in Western societies to encourage males to initiate their sexual exploration in adolescence, and for males interested in other males, older partners may appear “safer” than peers who may expose them to ridicule or even aggression if same sex interest is expressed (Coleman, 1989). Kilpatrick (1987) stated that “when generalizations are made across [prepubertal and postpubertal children], important data may be lost and findings may be misleading” (p. 177). Therefore, we decided to focus exclusively on children, i.e., those under 13 years of age, a threshold frequently used in the literature. The four-year age difference between partners had been used by our team in prior studies of Latino MSM (Carballo-Diéguez & Dolezal, 1995; Dolezal & Carballo-Diéguez, 2002); we maintained the same parameter to allow comparisons. We operationalized childhood sexual abuse as a childhood sexual experience with an older partner in which the child felt emotionally or physically hurt by the experience.

Given that convenience samples (e.g., those recruited from clinics or from the legal system) are likely to have unknown sample bias, and considering that random sampling methods are not a good strategy to sample populations that are rare and/or with stigmatized behaviors (Magnani, Sabin, Saidel & Heckathorn, 2005), we embedded our questions on childhood sexual experiences in a larger survey (Mello et al., 2007) for which participants were recruited using respondent-driven sampling (RDS, Heckathorn, 1997, 2002). RDS is a variant of chain-referral sampling that employs a dual system of structured incentives to reduce biases associated with chain-referral methods. Participants’ network size is assessed and used to weight the data based on likelihood of inclusion in the sample. RDS claims to make asymptotically unbiased estimates about hidden populations (Salganik & Heckathorn, 2004).

METHOD

The study was conducted in Campinas, a city with over 1 million inhabitants, located in Southeastern Brazil. Campinas is considered a reference city in health care since it lies at the geographical center of a metropolitan area composed of 19 municipalities, encompassing approximately 3,600 square km, and it is home to 2.33 million inhabitants, according to the 2000 Census (Instituto Brasileiro de Geografia e Estatística [IBGE], 2000). The study was approved by the Institutional Review Boards of all participating institutions. For more detail on study methods, please see Mello et al. (2007).

Participants

MSM were recruited into the study between October 2005 and October 2006 through RDS, a method in which participants recruit peers into the study (Heckathorn, 1997). The initial 10 individuals enrolled (“seeds”) were selected based on their high level of motivation to participate in the study, self-reported large social networks, and their popularity in the gay community as identified by peers. In addition, seeds were selected to represent diversity in socioeconomic characteristics and self-reported HIV serological status. These seeds were given three coupons each to invite peers to join the study. The coupons identified the study as “Semear Saúde” (Sowing Health) and encouraged participation to “contribute to improve the health of the community.”

Due to slow recruitment, 20 additional seeds were enrolled. Of a total of 30 seeds, eight did not recruit any participants. A total of 658 participants entered the study. After excluding those who chose “Other” in the gender identity questions (see Measures, below) as well as those lacking data regarding childhood sexual experiences with older partners, we were left with 575 individuals born male of whom 82 identified as women, transvestites, or transsexuals. When weighting was used to adjust these numbers, the sample resulted in 490 men and 84 transgender persons (the discrepancies in N are an artifact of the weights; when applied, the size of the sample and subgroups will vary slightly). All of them had sex with men. In terms of sexual attraction, 81.7% of transgender participants reported being attracted only to men, while 14.6% said they were attracted both to women and men. Among those identifying as men, 71.1% reported being attracted exclusively to men while 21.9% said they were attracted to both women and men.

In order to participate in the study, men had to be recruited by a previous participant, bring a valid recruitment coupon to the appointment, be 14 years of age or older, reside in the metropolitan area of Campinas, have had oral or anal sex with a man within the last six months, and be willing to provide a blood sample for syphilis rapid testing. HIV rapid testing was offered to all participants. Prior to enrolling in the study, written informed consent was obtained from all participants.

Procedures

Completion of all study procedures lasted approximately two hours. Data were collected through face-to-face interviews, audio computer-assisted self-interviews (ACASI), and blood samples.

Face-to-face interviews

Face-to-face interviews were used to collect information on the participants’ social networks, information that was essential for the estimation of weights in RDS. These questions included: First, “Approximately how many men or transvestites do you know, who have had at least one sexual relation (anal or oral) with a man or transvestite, and who (1) live in the Campinas region, (2) you could contact in person or by phone, and (3) you have spoken with in the past month.” Second, “Of these people, how many could you invite to this study?” Reponses to the second question were used to estimate network size, as discussed in the data analysis section.

ACASI

The ACASI included a question on gender identity (“Do you consider yourself a man, woman, transvestite, male-to-female transsexual, or other?”), sexual behavior during the prior two months, condom use, drug and alcohol use, homophobic abuse, perceptions related to HIV, HIV/STI testing and diagnosis history, and exposure to HIV prevention activities. The ACASI also included questions on childhood sexual experiences with older partners.

Assessment of childhood sexual experiences with older partners

The assessment of childhood sexual experiences with older partners appears in the Appendix. It closely follows our operationalization of such experiences and of childhood sexual abuse. It started by identifying age of the child when the first sexual contact took place and age difference with adult. If our definition criteria were met (child less than 13 years, adult at least four years older), then questions on what happened, with whom, and for how long were posed. Words such as “victim” and “perpetrator” were not used to avoid injecting value judgments in the inquiry. At the end of the inquiry, we asked if the participant felt physically or emotionally hurt, and if he considered it sexual abuse. There was no need to define childhood sexual abuse for the participant at this point given that the intention of the question was to understand the participant’s perception of the event(s) previously described.

Compensation

For their time and effort, participants could choose from a few similar value compensations (e.g., two cinema tickets or manicure and pedicure). For each person the participant recruited who enrolled in the study, the participant could select a slightly smaller compensation (a cinema ticket or a manicure service). Reimbursement was provided for public transportation to the study site.

Data Analysis

Weighing data according to personal network size is considered crucial in RDS recruited samples (Heckathorn, 2007). Consequently, we calculated a weight for each participant using the inverse of the participant’s personal network size plus 1 (to account for reciprocity). To achieve results that had a comparable N to the original sample, weights were multiplied by the sample size divided by the sum of the original weights. The weight calculation is therefore:

• (1/ (PNS + 1)) * (N / Σweights)

Data analyses were weighted based on these values. Yet, a literature review of studies using samples recruited with RDS show that while some researchers follow Heckathorn’s advice to weight the data (see Des Jarlais et al., 2007; Johnston, Malekinejad, Kendall, Iuppa, & Rutherford, 2008), others prefer to use unweighted data for their analyses (Blankenship, West, Kershaw, & Biradavolu, 2008; Kuhns, Vazquez, & Ramirez-Valles, 2008; Ramirez-Valles, Garcia, Campbell, Diaz, & Heckathorn, 2008; Xia et al., 2006). While deriving our conclusions from the weighted results, we also present unweighted results in our tables to allow readers to see the differences brought in by the weighting procedure.

Participants who self-identified as “man” were compared to participants who self-identified as “woman,” “transvestite,” or “transsexual (male-to-female),” grouped under the “transgender” label at the time of analysis. Comparisons were made on age, education, HIV status, and whether they had had childhood sexual experiences. For those who reported having had sexual experiences prior to age 13 with a partner at least four years older, the same two categories (man, transgender) were compared on various characteristics of their experiences. T-tests were used for continuous variables and chi-square tests were used for dichotomous/categorical variables. Finally, ANOVAs were used to compare three groups of men (those who did not have childhood sexual experiences, those who did but did not feel hurt by the experiences, and those who felt physically or emotionally hurt by the experiences) on sexual risk behavior and substance use variables. Frequency of unprotected receptive anal intercourse, unprotected insertive anal intercourse, number of male partners, and number of commercial sex partners were log-transformed prior to analyses due to skewed distributions.

RESULTS

Demographic Analysis

As shown in Table 1, transgender participants did not significantly differ from men in terms of age, but they did so in terms of education (lower level of education among transgender persons). Transgender participants were also significantly more likely to report having had childhood sexual experiences with an older partner. Overall, considering both groups, about a third of the participants had had such experience with an older partner.

Table 1.

Sample Description

	Unweighted scores						Weighted scores
	Men (N = 493)	Trans (N = 82)				Full sample (N = 575)	Men (N = 490)	Trans (N = 84)a				Full sample (N = 575)
	M (SD) range	M (SD) range	t/χ2	df	p	M (SD) range	M (SD) range	M (SD) range	t/χ2	df	p	M (SD) range
Age (years)	25.0 (7.2) 14–56	27.3 (10.1) 14–63	1.94	95	.055	25.3 (7.7) 14–63	25.3 (7.8) 14–56	26.9 (10.4) 14–63	1.28	100	.203	25.6 (8.2) 14–63
Education (years)	11.7 (2.6) 1–21	8.9 (3.0) 2–18	−8.55	571	<.001	11.3 (2.8) 1–21	11.5 (2.7) 1–21	8.5 (3.0) 2–18	−8.80	109	<.001	11.1 (3.0) 1–21
	%	%				%	%	%				%
Tested HIV+	6	14	5.90	1	.017	7	6	13	4.36	1	.051	7
Childhood sexual experience
No sex pre-13 with 4-year older partner	70	45	20.36	2	<.001	67	71	49	16.49	2	<.001	68
Had sex, but not physically or emotionally hurt	18	32				20	18	30				20
Had sex, physically or emotionally hurt	12	23				13	11	21				12

^aMen + Transgender Ns do not add up to 575 as function of weights.

Childhood Sexual Experiences with Older Partners

Table 2 focuses on the 185 cases (after weighting) in our sample that reported childhood sexual experiences with older partners. It shows that the mean age at the first experience was nine years, with the partner’s age being, on average, about 19 years. There were no statistically significant differences between men and transgender persons in terms of number of sexual occasions (note, however, that prior to weighting, transgender participants appeared to have significantly more sexual occasions than men; yet, this difference was reversed and no longer significant after weighting). For the great majority of the participants, the partner was a non-relative male. Among relatives, male cousins were the most frequent sexual partners. Six percent of participants said the sexual partner was an older female.

Table 2.

Characteristics of Childhood Sexual Experience Prior to Age 13 with a Partner at Least 4 Years Older

	Unweighted scores						Weighted scores
	Men (N = 146)	Trans (N = 45)				Full sample (N = 191)	Men (N = 142)	Trans (N = 43)				Full sample (N = 185)
	M (SD)	M (SD)	t/χ2	df	p	M (SD)	M (SD)	M (SD)	t/χ2	df	p	M (SD)
Age at first experience	9.19 (2.54)	8.87 (2.56)	<1	189	ns	9.12 (2.54)	9.11 (2.71)	9.03 (2.65)	<1	183	ns	9.09 (2.69)
Partner’s age	18.94 (7.23)	18.51 (6.06)	<1	189	ns	18.84 (6.96)	18.85 (6.88)	18.79 (6.03)	<1	183	ns	18.84 (6.68)
Number of sexual occasionsa	17.67 (89.55)	22.60 (50.30)	2.02	170	.045	18.91 (81.42)	32.13 (154.58)	14.42 (28.42)	<1	164	ns	27.84 (135.38)
Partner’s relationship to child:b	%	%				%	%	%				%
Male non-relative	58	66	2.23	1	ns	60	59	69	2.18	1	ns	62
Male cousin	19	23				20	20	24				21
Uncle	10	7				10	8	5				7
Female non-relative	3	0				3	5	0				4
Brother	3	2				3	4	2				3
Other male relative	2	0				2	3	0				2
Step-father	1	2				2	1	0				1
Female cousin	1	0				1	1	0				1
Sister	1	0				1	1	0				1
Behavior at first sexual experience:
Fondled by partner	89	89	0.04	1	ns	89	92	91	0.21	1	ns	92
Partner exposed genitals	91	89	0.37	1	ns	91	92	83	2.60	1	ns	90
Child fondled partner	78	78	0.02	1	ns	78	82	84	0.04	1	ns	83
Child masturbated partner	75	84	1.83	1	ns	77	77	84	0.83	1	ns	79
Performed oral sex	71	78	0.77	1	ns	73	70	77	0.69	1	ns	72
Receptive anal intercourse	52	82	13.55	1	<.001	59	51	86	17.53	1	<.001	60
Child exposed genitals	58	36	6.62	1	.011	52	58	33	8.37	1	.005	52
Masturbated by partner	50	38	1.92	1	ns	47	54	40	2.70	1	ns	51
Deep kissing	29	49	6.08	1	.018	34	31	44	2.55	1	ns	34
Received oral sex	31	36	0.36	1	ns	32	33	30	0.11	1	ns	32
Rimmed by partner	26	42	4.39	1	.041	30	23	40	4.53	1	.048	27
Insertive anal intercourse	19	20	0.05	1	ns	20	20	23	0.33	1	ns	21
Child rimmed partner	7	11	0.98	1	ns	8	8	14	1.57	1	ns	9
Vaginal sex	3	0	0.68	1	ns	2	5	0	1.75	1	ns	4
Cunnilingus	1	0	0.42	1	ns	1	1	0	0.42	1	ns	1
Other characteristics of first sexual experience:
Was forced	33	16	5.27	1	.025	29	27	12	4.50	1	.040	24
Was threatened	23	16	0.93	1	ns	21	20	14	0.60	1	ns	18
Was physically hurt	13	27	4.51	1	.038	16	9	23	6.32	1	.014	12
Felt bad about experience	38	22	3.65	1	.072	34	37	28	1.06	1	ns	35

Note. Percents may not sum to 100 due to rounding.

^aData log-transformed prior to analyses due to skewed distribution.

^bChi-square test comparing the two groups on whether partner was male.

For most participants, their first childhood sexual experience with an older partner included the partner exposing his/her genitals, mutual fondling, the child masturbating the older partner, the child performing oral sex on the partner, and the child being anally penetrated. Compared to men, transgender participants were significantly more likely to have been anally penetrated (86% vs. 51%), less likely to have exposed their genitals (33% vs. 58%), and more likely to have experienced anilingus (40% vs. 23%).

A quarter of the participants reported being physically forced during the sexual contact, with a statistically significant difference between men and transgender persons; the latter were less likely to have felt forced. Almost one-fifth of participants felt threatened. Only 1 in 10 participants felt physically hurt during the experience, and transgender participants were statistically significantly more likely to feel that they had been hurt. Overall, 66% of participants experiencing childhood sexual experience with an older partner did not feel forced, threatened or hurt by the experience (not shown in table).

Childhood Sexual Abuse

Almost one-third of the participants who had an early sexual experience with an older partner considered it childhood sexual abuse when reporting it during the interview, with no statistically significant difference between transgender participants and men. Fifty five percent of participants reported liking the experience at the time it occurred, 29% were indifferent, and only 14% stated they did not like it. At the time of the interview, these numbers had varied slightly in the direction of more transgender persons reporting liking the experience. At the time of the interview, the difference between transgender persons and men was statistically significant (the former were more likely to report liking the experience). Further analyses, not presented in the table, showed that 22 (19%) of 114 participants who had a childhood sexual experience with an older partner but did not feel hurt by it considered it abuse compared to 41 (54%) of 76 participants with such experience who felt hurt (Fisher’s statistic = 24.46; df = 1; p < .001). In other words, those who felt hurt by their experience were more likely to consider it sexual abuse than those who did not feel hurt. There were no statistically significant differences between the two groups in terms of education or the proportion of participants who had been anally penetrated.

Childhood Sexual Experiences with Older Partners, Childhood Sexual Abuse, and Sexual Risk Behavior in Adulthood

As shown in Table 3, childhood sexual experiences with older partners for both the hurt and not hurt groups was not associated with frequency of unprotected receptive or insertive anal intercourse. There were significant group differences in the number of male sex partners and commercial sex partners, with highest numbers among those who had childhood sexual experience with older partners but had not felt physically or emotionally hurt by the experience (interestingly, the unweighted results showed a higher number of sexual partners and commercial sex partners among those who had the experience and felt hurt by it, which is the association one would expect based on prior findings, e.g., Dolezal & Carballo-Diéguez, 2002). Individuals with childhood sexual experience with older partners who felt hurt by it were statistically more likely to use marijuana and those who had the experience but did not feel hurt by it were more likely to use cocaine.

Table 3.

Association between Childhood Sexual Experiences and Adult Sexual Risk Behavior and Substance Use

Mean frequency (SD) of:	Unweighted scores						Weighted scores
	No CSEOPa	CSEOPa but not hurtb	CSEOPa and hurtb				No CSEOPa	CSEOPa but not hurtb	CSEOPa and hurtb
	M (SD)	M (SD)	M (SD)	F	df	p	M (SD)	M (SD)	M (SD)	F	df	p
Unprotected receptive anal intercourse	2.02 (5.98)	2.43 (6.78)	1.84 (4.40)	<1	2	ns	1.87 (5.21)	2.20 (6.26)	1.24 (3.30)	1.32	2	ns
Unprotected insertive anal intercourse	1.89 (4.43)	2.60 (7.17)	1.96 (4.30)	<1	2	ns	1.92 (4.42)	2.59 (7.32)	1.62 (3.57)	<1	2	ns
Male partners	3.32 (5.17)	16.72 (94.41)	20.30 (88.22)	18.85	2	<.001	2.92 (4.27)	20.26 (116.14)	5.49 (21.59)	19.38	2	<.001
Commercial sex partners	2.78 (31.21)	14.68 (95.60)	25.61 (134.84)	11.21	2	<.001	2.45 (27.42)	18.72 (117.75)	4.08 (34.46)	12.04	2	<.001
Alcohol usec	2.67 (0.88)	2.75 (1.02)	2.77 (0.99)	<1	2	ns	2.63 (0.89)	2.67 (1.03)	2.67 (0.98)	<1	2	ns
Marijuana usec	1.42 (0.84)	1.58 (1.04)	1.68 (1.24)	3.01	2	.050	1.37 (0.79)	1.57 (1.02)	1.65 (1.23)	4.13	2	.017
Cocaine usec	1.24 (0.66)	1.31 (0.80)	1.38 (0.90)	1.37	2	ns	1.18 (0.56)	1.51 (1.06)	1.32 (0.81)	9.54	2	<.001

Note. Sexual risk variables log-transformed prior to ANOVA analyses due to skewed distributions.

^aCSEOP = Childhood sexual experience (before age 13) with an older (at least 4 years) partner.

^b“Hurt” = physically or emotionally.

^cReported on a 5-point scale from none to daily.

Finally, we compared those individuals who labeled their childhood sexual experience with older partners as childhood sexual abuse with those who did not. Table 4 shows that only one comparison showed statistical significance: Individuals who considered themselves to have been sexually abused were more likely to engage in unprotected insertive anal intercourse than those who did not consider themselves abused.

Table 4.

Association between Self-reported Childhood Sexual Abuse and Sexual Risk Behavior/Substance Use

Mean frequency (SD) of:	Unweighted Scores					Weighted Scores
	Not CSA	CSA				Not CSA	CSA
	M (SD)	M (SD)	t	df	p	M (SD)	M (SD)	t	df	p
Unprotected receptive anal intercourse	2.48 (6.64)	1.67 (4.19)	−0.75	187	ns	2.17 (5.85)	1.23 (3.35)	−1.40	188	ns
Unprotected insertive anal intercourse	2.03 (6.36)	2.90 (5.79)	1.95	187	.053	1.81 (5.85)	3.11 (6.44)	2.19	188	.030
Male partners	19.23 (99.28)	15.83 (75.85)	−0.21	188	ns	17.20 (105.91)	5.57 (18.16)	−0.28	136	ns
Commercial sex partners	21.46 (126.74)	14.47 (78.97)	0.48	185	ns	15.73 (109.17)	4.66 (19.50)	0.94	185	ns
Alcohol usea	2.69 (1.01)	2.88 (0.99)	1.22	181	ns	2.58 (0.99)	2.88 (1.00)	1.81	180	.071
Marijuana usea	1.69 (1.22)	1.49 (0.92)	−1.13	182	ns	1.61 (1.14)	1.53 (0.95)	−0.48	182	ns
Cocaine usea	1.37 (0.86)	1.28 (0.80)	−0.68	181	ns	1.48 (1.03)	1.28 (0.75)	−1.46	120	ns

Note. Sexual risk variables log-transformed prior to ANOVA analyses due to skewed distributions.

CSA = childhood sexual abuse.

^aReported on a 5-point scale from None to Daily.

DISCUSSION

Our study, the largest conducted to date with MSM in Brazil which included an extensive assessment of childhood sexual experiences with older partners, found that 29% of boys later in life identified as men, and 51% of those later identified as transgender persons reported having sexual experiences before age 13 with a partner at least four years their senior. Of those participants reporting childhood sexual experiences with older partners, 73% of men and 88% of transgender persons denied being forced into the sexual activity, 80% and 86% (respectively) denied feeling threatened, and 91% and 77% (respectively) denied feeling physically hurt. Furthermore, a high percentage of participants reported liking or feeling indifferent about the childhood sexual experience with older partners, both at the time it happened and also at the time of our interview. Only one-third of participants with such early experiences reported they felt bad about it, and slightly less than one-third considered it childhood sexual abuse. The latter were more likely to have been forced than those not reporting childhood sexual abuse.

The existent scientific literature offers divergent interpretations of this type of finding. On one side, Jinich et al. (1998) argued that one cannot assume that because an individual reports that a sexual experience was non-coercive that it was non-exploitative or free of any subsequent detrimental effects. Along this same line, King (2000) stated that a low level of rejection or condemnation of the childhood sexual experiences with older partners is a defensive mechanism by which the child tries to “tough it out” when abused, rather than admit his hurt, fear, confusion, and need.

On the other side, Kilpatrick (1987) argued that there is a need to distinguish “abuse as some type of harm (a consequence of sexual activity that can be quantitatively measured) [from] abuse [as] violation of social norms” (p. 175). Rind et al. (1998) stated that “Classifying behavior as abuse simply because it is generally viewed as immoral or defined as illegal is problematic, because such a classification may obscure the true nature of the behavior and its actual causes and effects” (p. 45). Previous findings of perceptions of childhood sexual abuse among MSM have led researchers to propose that the developmental role of childhood sexual experiences with older partners among MSM must be considered differently than among the general population, due to social constraints on homosexual behavior (Stanley et al., 2004; Steever, Follette, & Naugle, 2001). Furthermore, discounting the report of the participants who found the childhood sexual experiences with older partners not harmful cannot be easily justified without also opening the possibility of discounting the reports of those who found it harmful.

Our data did not allow us to draw conclusions on the overall associations of childhood sexual experiences with older partners with psychological functioning in adulthood, since that is not what we set out to measure. We focused exclusively on the hypothesized association between such early sexual experiences and HIV sexual risk behavior in adulthood. Such an association has been found in many studies conducted in the U.S. and Europe, including studies conducted by members of our team with two different cohorts of Latino men recruited in the New York City metropolitan region for which the same operational definition of childhood sexual experiences with older partners and childhood sexual abuse and the same assessment instruments were utilized (Carballo-Diéguez & Dolezal, 1995; Dolezal & Carballo-Diéguez, 2002). Yet, we found no such differences between participants with childhood sexual experiences with older partners and those without such experiences in the Brazilian sample. Although those reporting early sexual experience who did not feel hurt by it had a larger number of sexual partners than those in the other two groups, this cannot be interpreted as a risk correlate per se, given that we found no differences in terms of unprotected sex.

Although we acknowledge that failure to prove a relationship does not prove that a relationship does not exist (Ondersma et al., 2001), the findings were still puzzling. We can only speculate about possible explanations. For example, the Brazilian sample had very low proportions of men or transgender persons who had felt forced, hurt or threatened in the sexual act. It may also be that the sexual abuse construct has less salience in the Brazilian social discourse, especially when related to men, than in that of U.S. or European countries. Furthermore, there may be elements in the Brazilian culture that facilitate sexual expression more than in other cultures and that mitigate negative correlates in adulthood. Parker (1991), in discussing the “celebration of the flesh” during Brazil’s famous carnival, stated that “sexual transactions that cross the lines of class, age, and race, lesbian and homosexual interactions, exhibitionism, and any number of other marginal pleasures become possible in a world where repression and oppression cease to exist” (p.148). Such transactions did not appear to be exclusively restricted to carnival time.

In our study, rates of anal intercourse were much higher than what has been previously reported in studies with the same operationalization of childhood sexual experiences with older partners (Dolezal & Carballo-Dieguez, 2002; Paul et al., 2001; Ratner et al., 2003). This finding may be related to the cultural importance of penetrative sex in Brazil (Taquette, Vilhena, Paulo dos Santos, & Vianna de Barros, 2005), and to what Parker (1991) called the structure of erotic practice in contemporary Brazil. Parker described the meia or troca-troca game played by boys in which partners are said to take turns masturbating, fellating, or, most commonly, penetrating one another. Parker states 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).

Findings from this study also offer insights into how the childhood sexual experiences with older partners of boys who later in life did not identify as men might differ from those who do. Compared to participants who identified as men, those who did not were more likely to have had a childhood sexual experience before the age of 13 with a partner at least four years older, were more likely to have engaged in receptive anal intercourse, less likely to have felt physically forced into the experience (although more likely to have felt hurt), and more likely to view it favorably at the time of the interview. This raises the question of whether transgender people, when children, demonstrated more feminine qualities that made them targets for penetrative sex by older partners (in this study, we did not collect information on gender behavior in childhood). Among adults, there is evidence that MSM are inclined to take a more active, penetrative role according to the perceived effeminacy of a partner with respect to themselves (Carballo-Diéguez et al., 2004), and the behavior of the adult partners of our participants may have been influenced by the degree of masculine demeanor (or lack of it) of the boys. It could also be that the types of sexual behavior that took place during these childhood sexual experiences contributed to the boys’ identification, not as a man, but as a woman, transvestite, or transsexual. Maybe being the object of sexual interest of an older partner was interpreted by the boys in a flattering manner and, for this reason, they gave a favorable opinion of the event. The observed lower level of education among transgender persons in the sample than among those who identified as men is not surprising, considering high levels of stigma and discrimination in the country, including within the school environment (Castro, Abramovay, & Silva, 2004; Brazil, 2004). In addition, it is not uncommon for transgender persons to be rejected by family members and have to find their own means of subsistence in early adolescence while abandoning school (Peres, 2004). Concerned about high levels of violence and discrimination against transgender people in Brazil, the National AIDS Program has launched several campaigns and educational materials, including one specific for transvestites, aimed at schools to combat this violation of human rights. (Brazil, 2004)

Strengths and Limitations

Our study had several strengths: (1) we focused on individuals born male who later identified as men (independently of sexual orientation, i.e., homosexual, heterosexual, or bisexual) but also on those who in adulthood identified as women, transvestites or transsexuals, a seldom studied population; (2) we used RDS, a method designed to generate unbiased estimates of hidden populations (Heckathorn, 1997), and we weighted the results; (3) we embedded the study of childhood sexual experiences with older partners in a larger health study (Mello et al., 2007), publicized as such, thus avoiding the bias that could have resulted if the study has been advertised as one of problems in childhood; (4) we used ACASI, a technology that results in higher reports of socially censored behaviors (Turner et al., 1998); (5) our approach to assessing childhood sexual experiences with older partners was neutrally worded, gathered details of the experience, and only then asked the participant to report his reaction to it and whether he considered it abuse; (6) we did not restrict our inquiry to just one or two questions on childhood sexual abuse (for a critique of this approach, see Hulme, 2004).

Our study also had several limitations. First, although we used RDS for its reported strength to control sample bias, RDS is still a relatively new sampling methodology that requires further testing and statistical development. In some cases, weighting the data resulted in differences in the opposite direction than that shown by the raw data, which was in agreement with prior literature findings. This leaves as an open question whether the weighting process based on participants’ reported network size corrects or distorts results.

Second, as we have pointed out previously, while a majority of participants did not feel hurt by the experience, this study did not assess current functioning across a variety of psychological domains, relying solely on the participants’ report of sexual behavior. As such, our study did not explore the presence of mood or behavioral problems.

Third, recall of events that took place many years before the interview may be subject to reinterpretation, distortion, and partial repression. This would bring unknown biases to the reported data, and may also result in underestimates of the true prevalence of childhood sexual experiences with older partners (Putnam et al., 2003). Assessment of reaction to childhood sexual experiences with older partners shortly after the time they happened may elicit different evaluations of the experiences; yet, a longitudinal study of such experiences in a cohort would be required to establish the comparative risk of participants’ sexual risk behavior in adulthood.

Fourth, we did not explore what identity labels, such as transvestite or transsexual, meant for participants. Finally, these results may be specific for the sample and not generalizable to other Brazilian populations.

Despite these limitations, our results underscore the complexity of issues involved in the study of childhood sexual abuse and childhood sexual experiences with older partners. Although childhood sexual experiences with older partners were not infrequent in this Brazilian sample, particularly among transgender individuals, the small proportion of participants who reported being harmed by such experiences and the absence of correlates in terms of sexual risk behavior in adulthood make childhood sexual experiences with older partners a poor marker to identify populations that could profitably be targeted for sexual risk prevention interventions. Our research highlights that different cultural environments may play important roles in the way individuals perceive these early sexual experiences and the way they may be affected by them.

APPENDIX

Assessment of Childhood Sexual Experiences with an Older Partner (CSEOP)

Carballo-Diéguez et al., 2005 (previously unpublished)

Now, I will ask you to remember your initial sexual experiences.

• 1.How old were you when you had your first sexual contact of any kind (manual, oral, genital, or anal) with another person?

  • __ __ IF 13 YEARS OR OLDER, GO TO NEXT SECTION. years

• 2.How old was the person with whom you had your first sexual contact?

  • __ __ IF AGE DIFFERENCE IS 4 YEARS OR MORE, GO TO 6. years

• 3.Before you turned 13, did you have sexual contact with anybody who was at least 4 years older than you?

  • _____ No. [IF NO, GO TO NEXT SECTION]

  • ____ Yes, I had sex with one person who was at least 4 years older than me.

  • _____ Yes, I had sex with more than one person who was at least 4 years older than me. (For the next question, think about the first one of these persons).

• 4.How old were you when you had the first sexual contact with that person? __ __ years
• 5.How old was that person? __ __ years
• 6.Approximately when did this sexual contact happen?

  • ___/____/_____ MM DD YYYY

• 7.How many times did it happen? __ __ __ IF ONLY ONE TIME, GO TO 9a
• 8.Approximately, until when did these sexual contacts take place?

  • ___/____/_____ MM DD YYYY

• 9a.Was the other person a man, woman or transvestite?

  • [ ] Man

    [ ] Woman

    [ ] Transvestite

• 9b.Who was s/he?_____________________________

  • ○ BROTHER	◦ GRANDFATHER
    ○ SISTER	◦ UNCLE
    ○ FATHER	◦ AUNT
    ○ STEPFATHER	◦ COUSIN
    ○ MOTHER	◦ FRIEND/SCHOOLMATE
    ○ STEPMOTHER	◦ OTHER: SPECIFY: ________________
    ○ GRANDFATHER

• 10.Please tell me if any of the following happened:

  • NO	YES
    ◦	◦	a. The other person showed his/her genitals to you.
    ◦	◦	b. You showed your penis to that person.
    ◦	◦	c. The other person fondled you
    ◦	◦	d. You fondled the other person
    ◦	◦	e. You two deep kissed
    ◦	◦	f. You masturbated the other
    ◦	◦	g. The other person masturbated you
    ◦	◦	h. You put your mouth on the other person’s genitals
    ◦	◦	i. The other person put your penis in her/his mouth
    ◦	◦	j. You licked the other person’s anus.
    ◦	◦	k. The other person licked your anus
    ◦	◦	j. The other person penetrated your anus
    ◦	◦	k. You penetrated the other person's anus
    ◦	◦	l. Other:_____________________________
    		[IF Q9a IS WOMAN, ALSO ASK THE FOLLWING;]
    ◦	◦	m. You put your tongue on her vagina
    ◦	◦	n. You penetrated her vagina
    ◦	◦	o. Other:_____________________________

• 11.Did the other person physically force you to do something you did not want to do?

  • ◦ NO

    ◦ YES

• 12.Did the other person threaten you in any way to get you to do something?

  • ◦ NO

    ◦ YES

• 13.Did you feel emotionally hurt by this sexual activity?

  • ◦ NO

    ◦ YES

• 14.Did you feel physically hurt by this sexual activity?

  • ◦ NO

    ◦ YES

• 15.Did you let anybody know what took place shortly after it happened (within a year)?

  • ◦ NO GO TO 18

    ◦ YES

• 16.Did you tell what happened to...

  • NO	YES
    ◦	◦	a. a family member
    ◦	◦	b. a peer
    ◦	◦	c. a teacher, clergy, police
    ◦	◦	d. other________________________

• 17.Those who came to know about the event...
  • ○were angry with you
  • ○were indifferent
  • ○were supportive
  • ○gave mixed responses
• 18.How did you react at the time of the event?
  • ◦I liked it
  • ◦I was indifferent
  • ◦I did not like it
• 19.How do you feel at present with respect to what happened?
  • ◦I like it
  • ◦I am indifferent
  • ◦I do not like it
• 20.Would you consider this event as sexual abuse?

  • ◦ NO

    ◦ YES

• 21.Have you discussed your experience with a health professional (e.g., psychologist or social worker) ?

  • ◦ NO

    ◦ YES

• 22.Have you read any books on sexual abuse?

  • ◦ NO

    ◦ YES

If these memories bother you and you would like a referral to a mental health practitioner, I will be happy to do it at the end of the interview.