Arch Sex Behav DOI 10.1007/s10508-016-0781-8

SPECIAL SECTION: THE PUZZLE OF SEXUAL ORIENTATION

Recalled Gendered Behavior in Childhood: A Comparison of Androphilic Men, Gynephilic Men, and Androphilic Women in Japan Lanna J. Petterson1 · Chelsea R. Wrightson1 · Paul L. Vasey1

Received: 13 April 2013 / Revised: 11 May 2016 / Accepted: 25 May 2016 © Springer Science+Business Media New York 2016

Abstract The current study tested the hypothesis that men who are androphilic (sexually attracted to adult men) in a nonWestern, developed country—Japan—would recall engaging in more female-typical behavior, and less male-typical behavior, in childhood, compared to men who are gynephilic (sexually attracted to adult women). Androphilic men, androphilic women, and gynephilic men (N = 302) responded to the Female-Typical Behavior Subscale and the Male-Typical Behavior Subscale of the Childhood Gender Identity Scale, which asked participants to recall their childhood behavior. Results indicated that gynephilic men scored highest on the Male-Typical Behavior Subscale and lowest on the Female-Typical Behavior Subscale. Androphilic women scored the highest on the Female-Typical Behavior Subscale and lowest on the Male-Typical Behavior Subscale. Androphilic men scored intermediately for both the Male- and FemaleTypical Behavior Subscales. The results supported the hypothesis that Japanese androphilic men would recall greater gendernonconforming childhood behavior compared to gynephilic men. These results further reinforce the conclusion that childhoodgender-nonconformingbehavior isacross-culturallyuniversal aspect of psychosexual life course development in androphilic men. We discuss why this may be the case, as well as why crosscultural variation occurs in the magnitude with which recalled childhood gender nonconformity is reported by androphilic males. Keywords Sexual orientation · Gender nonconformity · Androphilia · Japan

Introduction Androphilia refers to attraction and arousal to adult males. Gynephilia refers to attraction and arousal to adult females.1 Retrospective studies conducted in Western cultures, such as Canada, the U.S., and Australia, indicate that men who are androphilic in adulthood recall significantly higher levels of childhood gender-nonconforming behavior compared to gynephilic men (e.g., Bailey & Zucker, 1995; Bogaert, 2003; Dunne, Bailey, Kirk, & Martin, 2000; VanderLaan, Gothreau, Bartlett, & Vasey, 2011; VanderLaan, Petterson & Vasey, 2015; Zucker et al., 2006). A meta-analytic review of this literature by Bailey and Zucker confirmed an association between recalled childhood gender nonconformity and adulthood androphilia among Western men. Furthermore, the effect sizes associated with the male sexual orientation group difference in recalled childhood gender-nonconforming behavior are large by the standards of most behavioral sciences research; a meta-analytic review of 32 studies yielded a mean effect size of d = 1.37 (SD = .38). The review estimated that 89 % of androphilic men exceeded the median gender nonconformity score of gynephilic men, whereas only 2 % of gynephilic men exceeded the median gender nonconformity score of androphilic men. Further, prospective research indicates that boys who are gender nonconforming in childhood are overwhelmingly androphilic in adulthood (Green, 1987; Singh, 2012; Steensma, van der Ende,Verhulst,&Cohen-Kettenis,2013;Zucker,2014).Assuch, the prospective research is consistent with the findings from 1

& Lanna J. Petterson [email protected] 1

Department of Psychology, University of Lethbridge, Lethbridge, AB T1K 3M4, Canada

In Western culture, men who are gynephilic are labeled heterosexual, whereas men who are androphilic are labeled gay or homosexual. These identity categories are not necessarily recognized cross-culturally (including in Japan; McLelland, 2000) and, consequently, represent poor constructs for the type of research conducted here. As such, we employ the terms androphilic and gynephilic, which lend themselves much more easily to cross-cultural comparisons.

123

Arch Sex Behav

retrospective studies that link childhood gender nonconformity in boys to adult androphilia in men. Taken together, these studies haveledmanyresearcherstoconcludethatgendernonconformity in boyhood is a developmental precursor of androphilia in adulthood. Indeed, Whitam (1980) has argued that gender-nonconforming behavior in boyhood should be considered a cross-culturally universal and normative aspect of the psychosexual life course development of the majority of androphilic men. Proponents ofthisdevelopmental femininityhypothesisgenerally adhere to the conclusion that innate factors influence the expression of childhood gender nonconformity in (pre)androphilic boys more so than social factors (for further discussion, see Bailey & Zucker, 1995). One potential criticism of this perspective is that the majority of the existing relevant studies have been confined to Western populations. Consequently, some scholars have argued that these findings are better explained by social influences unique to Western cultures, as opposed to shared biological factors that transcend cultural contexts and influence both childhood behavior and patterns of sexual preference (Gottschalk, 2003; Risman & Schwartz, 1988). A cross-cultural perspective would help elucidate whether gender-nonconforming behavior in (pre)androphilic boys is cross-culturally universal or the consequence of culturally unique influences. When adopting such a perspective, this criticism appears indefensible, because the relationship between childhood gender nonconformity and adulthood sexual orientation in males2 has been quantitatively investigated in a wide array of disparate non-Western cultures, including Brazil, Guatemala, the Philippines, Samoa, Thailand, and Turkey. Cardoso (2005) examined the recalled childhood behavior of exclusively androphilicmales in a Brazilian fishing village, known locally as paneleiros. Paneleiros were more likely than gynephilic men to recall engaging in female-typical play activities, such as playing with dolls and ring-around-the-rosy. Paneleiros also reported preferring girls as playmates more frequently than gynephilic men. Subsequently, Cardoso (2009) undertook a comparative examination of recalled childhood behavior and adult sexual orientation among males in three culturally disparate nations: Brazil, Turkey, and Thailand. In line with his previous work, androphilic males in all three of these cultures were more likely than gynephilic men to recall engaging in female-typical behavior and a preference for female play partners in childhood. Whitam (1980) collected data on recalled childhood behavior and adult sexual orientation in males from populations in Guatemala, Brazil, and the Philippines. In all three of these societies, androphilic males recalled greater interest in female-typical toys, activities, and dressing in girls clothing during childhood

compared to gynephilic men. Additionally, in all three of these cultures, androphilic males recalled a greater preference for girls as play partners than did gynephilic men. Bartlett and Vasey (2006) examined recalled childhood behavior among Samoan fa’afafine. In Samoa, fa’afafine are recognized as belonging to a third-gender category that is distinct from that of “man” and of “woman.” Fa’afafine tend to be notably feminine with respect to their gender presentation and are, almost without exception, exclusively androphilic. Bartlett and Vasey found that fa’afafine recalled engaging in significantly more female-typical, and significantly less maletypical, behavior in childhood than gynephilic men. As one of their fa’afafine participants stated: “I really didn’t like to be with the boys…I wanted to be with the girls, play with the girls, play with the girls’ toys” (p. 663). Another fa’afafine participant stated: “Yeah, They forced me to be on the rugby team. They wanted me to act “straight” and be on the team. They were too rough and I didn’t want to have anything to do with it” (p. 663; emphasis in original). To date, all of the non-Western research linking gendernonconforming behavior in childhood with male androphilia in adulthood has been conducted in developing countries (Bartlett & Vasey, 2006; Cardoso, 2005, 2009; Whitam, 1980, 1997).3 Given this, a gap exists in the literature with respect to whether androphilicmales from non-Western,developed countriesrecall significantly more gender-nonconforming behavior in childhood than gynephilic men. To fill in this gap in the literature, the current study examined the relationship between childhood behavior and adult sexual orientation among men in Japan, one of the few developed, non-Western societies (Sullivan & Sheffrin, 2003). In light of previous research on this topic, we hypothesized that Japanese androphilic men would recall engaging in more female-typical behavior, and less male-typical behavior, in childhood, compared to gynephilic males.

Method Participants The present study included 79 gynephilic men, 80 androphilic women, and 143 androphilic men (N = 302). Of the gynephilic men, in the year prior to being interviewed, 93.7 % reported experiencing “sexual feelings only towards females” (Kinsey 0) and 6.3 % reported experiencing “most feelings towards females, but occasional feelings towards males” (Kinsey 1). Of the androphilic women, in the year prior to being interviewed, 3

2

We have used the term male, as opposed to men, here because many androphilic biological males in non-Western cultures identify as belonging to a “third” gender category that exists beyond that of “man” or “woman.”

123

Developing countries are those with a low standard of living, undeveloped industrial base, and low Human Development Index (HDI) relative to other countries (Sullivan & Sheffrin, 2003). The HDI is a composite statistic of life expectancy, education, and income indices used to rank countries into four tiers of human development (United Nations Development Project, 2013).

Arch Sex Behav

93.8 % reported experiencing “sexual feelings towards males only” (Kinsey 0) and 6.3 % reported experiencing “most sexual feelings towards males, but occasional sexual feelings towards females” (Kinsey 1). Of the androphilic men, in the year prior to being interviewed, 84.9 % reported experiencing “sexual feelings towards males only” (Kinsey 6) and 15.1 % reported experiencing “most sexual feelings towards males, but occasional sexual feelings towards females” (Kinsey 5). Data were collected by PLV during seven separate trips to Japan, from November 2009 to November 2015. During the first six trips to Japan, four Japanese research assistants (two gynephilic men and two androphilic men) recruited all participants from neighborhoods and bars in the Kansai region of Japan (population: 22,757,897; Ministry of Internal Affairs and Communications Statistics Bureau, 2011). Only Japanese men and women who were 18 years of age and older were recruited. Androphilic men were largely recruited from a Kansai gaisen bar located in a neighborhood that contains numerous of bars popular with androphilic men. Gaisen is a Japanese term that refers to Japanese androphilic men who prefer foreign men (usually Caucasian) as sexual partners. Because Japanese bars tend to be extremely small by Western standards (i.e., a typical one might seat\10 customers), there can be many bars clustered within the same building and even more clustered within the same neighborhood. With respect to the Kansai region overall, Osaka (metro population: 19,341,976; Statistics Bureau Ministry of Affairs and Communications Japan, 2015) has a far greater number of bars that are popular with androphilic men than do the other major cities in the region, such as Kyoto or Nara. If androphilic men living in cities such as Kyoto or Nara cities want to socialize with other androphilic men, they almost invariably go to Osaka, which is a relatively short train ride away. Osaka offers a greater selection of venues than other cities in the region as well as the anonymity of a big city. Gaisen bars were initially targeted for participant recruitment given that Caucasian men are welcome. Caucasian men are often unwelcome in non-gaisen bars. This is because each bar has very limited space and caters to specific clientele who are seekinga very specific sexual “type” (e.g., Japanese “bears”). If a customer does not conform to the “type” to which the bar caters, then they are perceived as taking up valuable space, detracting from the bars’ atmosphere, and potentially making other customers feelthat they are wasting their time and money. To ensure that the sample of androphilic males was not collected entirely from a gaisen bar, data were also collect at the Osaka Gay Pride Festival during a seventh trip to Japan in 2015. The festival is a day-long event in central Osaka that attracts a diverse array of Japanese androphilic men. Three native Japanese speaking research assistants (all gynephilic women) helped with participant recruitment and data collection at the festival. PLV assisted with participant recruitment at both the

gaisen bar and the Osaka Gay Pride Festival. Of the sample of androphilic men, 50 were collected from the gaisen bar and 93 were collected from Kansai neighborhoods, or non-gaisen bars, or from the Osaka Gay Pride Festival. Potential participants were told that we were undertaking a study about the childhood behavior of men and women. Participants were given questionnaires, which they completed on their own. Research assistants were available to answer questions. Questionnaires were returned in blank manila envelopes to ensure anonymity. Of the individuals recruited, 42 were excluded because they did not meet the inclusion criteria (e.g., participants were required to report androphilic or gynephilic attraction, report that they were a manor a woman,complete the majority of the questionnaire). We are unable to assess how many individuals were approached about participating in the present study but either declined or did not return their questionnaires. Measures Questionnaires were provided to participants in Japanese after being translated and translated back by two fluent JapaneseEnglish speakers. The questionnaires included items pertaining to participants’ biographic information, sexual orientation, and childhood behavior. Biographic information collected included participants’ biological sex, gender, age, and childhood socioeconomic bracket. Sexual orientation was assessed using a Kinsey Scale for sexual attraction (Kinsey, Pomeroy, & Martin, 1948). Participants were asked to describe their feelings over the past year. Participants were given the response range from 0 = “Sexual feelings toward females only” to 6 = “Sexual feelings toward males only.” Thus, this scale assessed the extent to which participants experienced androphilic attraction, gynephilic attraction, or non-exclusive attraction to either sex. Recalled childhood behavior was assed using the Childhood Gender Identity Scale (CGIS; see Appendix). The CGIS is an adaptation of the psychometrically validated Gender Identity Questionnaire for Children (Johnson et al., 2004). The CGIS consists of a Female-Typical Behavior Subscale and a MaleTypical Behavior Subscale, which were used to measure the extent to which participants recalled behavior that conformed to, or deviated from, behavior typical of their gender, between the ages of 6–12 years. The mean score was calculated for each of the two subscales, thus providing a Female-Typical Behavior and Male-Typical Behavior score. Standardized inter-item reliabilities (alphas) for the childhood Female-Typical Behavior Subscale ranged from .66 to .86 for all groups and the sample total. Standardized inter-item reliabilities (alphas) for the childhood Male-Typical Behavior Subscale ranged from .68 to .81 for all groups and the sample total.

123

Arch Sex Behav

Statistical Analysis Biographic variables were first examined for possible group differences. Descriptive statistics for each biographical variable are shown in Table 1. First, we sought to determine whether the gaisen bar recruited androphilic men differed from those who were recruited from other locations with respect to their recalled childhood behavior. The rationale for this comparison was that Japanese men who frequent gaisen bars may have been more Westernized and this may have influenced their responses, thus undermining the raison d’être of the study. The responses of androphilic men recruited in the gaisen bar and those recruited elsewhere were compared using a two-way mixed model repeated measures ANOVA, for which Group (i. e., recruitment location category) was a between-subject factor and CGIS Subscale (i.e., Female-Typical Behavior and MaleTypical Behavior Subscales) was the within-subject factor. Mean and SD values as well as inferential statistics and significance values for the two-way mixed model repeated measures ANOVA are shown in Table 2 by group. Independent sample t-tests were conducted to compare the recruitment group’s responses to each of the CGIS Subscales. The central research question was then addressed. To do so, androphilic men’s, gynephilic men’s, androphilic women’s responses to the CGIS Subscales were compared using a twoway mixed model repeated measures ANOVA, for which Group (i.e., sexual orientation category) was the between-subject factor and CGIS Subscale (i.e., Female-Typical Behavior and MaleTypical Behavior Subscales) was the within-subject factor. If a significant Group x Subscale interaction was found, direct group comparisons of each subscale were then conducted. Betweengroup comparisons were performed using one-way ANOVAs, with the alpha levels adjusted to a = .025 to maintain a Type I Error rate of a = .05 across multiple comparisons. Post hoc direct

group comparisons were performed using Dunnett T3 tests. Paired sample t-tests were used toperform post hoc within-group comparisons of CGIS Subscale (i.e., Female-Typical Behavior and Male-Typical Behavior) scores. The alpha levels were adjusted to a = .017 to maintain a Type I Error rate of a = .05 across multiple comparisons. Effect sizes for paired sample (d) were calculated by dividing the mean by the SD. Mean and standard deviation values, as well as inferential statistics and significance values, for the one-way ANOVA and paired sample t-test analyses are shown in Table 3 by group. This analysis was repeated with men recruited from gaisen bars excluded to examine whether they influenced the original analysis (only the post hoc comparisons that include androphilic males are presented to limit redundancy).

Results Comparison of Androphilic Men Recruited From a Gaisen Bar and Those Recruited Elsewhere Biodemographic Analysis Groups did not differ significantly in age, t(141)\1, Hedge’s g = .11 or social economic status during childhood (which was bifurcated [lower/lower-middle class versus middle class and higher] to permit group comparisons), likelihood-ratio χ2(1) = 1.43, p = .231. Group Comparison of Gender Subscales A two-way mixed model repeated measure ANOVA revealed a main effect of Subscale (corrected with Greenhouse-Geisser), a main effect of Group, and a Subscale x Group interaction

Table 1 Descriptive statistics of biographic information by group

Socioeconomic status Upper/upper-middle class (%)

Gynephilic men (n = 79)

Androphilic women (n = 80)

Androphilic men (n = 143)

41.3

49.4

33.8

Middle class (%)

41.8

63.7

48.3

Lower/lower-middle class (%)

8.9

2.5

10.5

30.44 (11.31)

36.19 (13.97)

30.33 (8.07)

18–65

19–82

18–52

Age in years M (SD) Age range

Androphilic men recruited from a gaisen bar (n = 50) Androphilic men not recruited from a gaisen bar (n = 93) Socioeconomic status Upper/upper-middle class (%) 48 Middle class (%) 42 Lower/lower-middle class (%) 10 Age in years M (SD) Age range

123

37.6 51.6 10.8

29.74 (7.89)

30.66 (8.19)

19–48

18–52

Arch Sex Behav Table 2 Mean and SD values, inferential statistics, and significance values for recruitment location group comparison on the measure of recalled childhood behavior Androphilic men not recruited from a gaisen bar (n = 93) Androphilic men recruited from a gaisen bar (n = 50) M

SD

M

SD

Female-Typical Behavior Subscale 1.96

.66

1.95

.61

Male-Typical Behavior Subscale

2.67

.82

3.11

.85

Main effect of Subscale Main effect of Group

F(1, 141) = 83.80, p\.001, η2p = .37 F(1, 141) = 7.18, p = .008, η2p = .05

Subscale 9 Group interaction

F(1, 141) = 4.71, p = .032, η2p = .03

(corrected with Greenhouse-Geisser). The groups did not differ significantly in their response to the Female-Typical Behavior Subscale, t(141)\1, Hedges’ g = -.01. However, androphilic men recruited from the gaisen bar scored significantly higher on the Male-Typical Behavior Subscale than androphilic men who were recruited at the other locations, t(141) = 3.02, p = .003, Hedges’ g = .53. In light of these results, we conducted two sets of analyses. The first compared gynephilic men, androphilic women, and our combined sample of androphilic men (gaisen and nongaisen). The second compared gynephilic men, androphilic and only the sample of non-gaisen androphilic men. Comparison of Gynephilic Men, Androphilic Women, and All Androphilic Men

(Hedges’ g = .88). Androphilic men scored significantly higher than androphilic women (Hedges’ g = .99). Within-group analysis indicated that androphilic men scored significantly higher on the Male-Typical Behavior Subscale than on the Female-Typical Behavior Subscale. Gynephilic men scored significantly higher on the Male-Typical Behavior Subscale than the Female-Typical Behavior Subscale. Androphilic women scored significantly higher on the Female-Typical Behavior Subscale than the Male-Typical Behavior Subscale.

Comparison of Gynephilic Men, Androphilic Women, and Non-Gaisen Androphilic Men Biodemographic Analysis

Biodemographic Analysis Participant differed significantly in age, Brown-Forsythe F(2, 193.33) = 7.31, p = .001, ηp2 = .05. However, age did not correlate with our variables of interest (p = .062 to p = .808). Groups did not differ significantly regarding socioeconomic status during childhood (which was bifurcated to permit group comparisons), χ2(2) = 4.00, p = .135.

Participant groups were shown to differ significantly in age, (ANOVA) F(2, 202.18) = 6.58, p = .002, ηp2 = .05. However, age did not correlate with our variables of interest (p = .062 to p = .180). Groups did not differ significantly regarding socioeconomic status during childhood(which was bifurcated to permit group comparisons), χ2(2) = 4.40, p = .111.

Group Comparison of Gender Subscales

Group Comparison of Gender Subscales

A two-way mixed model repeated measures ANOVA, corrected with Greenhouse-Geisser, Geisser, revealed a Group x Subscale interaction, F(2, 299.00) = 140.95, p\.001, η2p = .49. Because an interaction was found, we proceeded with direct group comparisons of each CGIS Subscale. For the Female-Typical Behavior Subscale, a main effect of Group was observed. Androphilic women scored significantly higher than gynephilic men (Hedges’ g = 2.14) and androphilic men (Hedges’ g = 1.50). Androphilic men scored significantly higher than gynephilic men (Hedges’ g = .75). For the Male-Typical Behavior Subscale, a main effect of Group was observed. Gynephilic men scored significantly higher than androphilic women (Hedges’ g = 2.06) and androphilic men

A two-way mixed model repeated measures ANOVA, corrected with Greenhouse-Geisser, revealed a Group x Subscale interaction, F(2, 249.00) = 134.96, p\.001, η2p = .52. Because a significant interaction was found, we proceeded with direct group comparisons of each CGIS Subscale. For the Female-Typical Behavior Subscale, a main effect of Group was observed, Brown-Forsythe F(2, 172.28) = 106.00, p\.001, η2p = .46. Androphilic men scored significantly higher thangynephilicmen (p\.001,Hedges’ g= .77)andsignificantly lower than androphilic women (p\.001, Hedges’ g = −1.41). For the Male-Typical Behavior Subscale, a main effect of Group was observed, Brown-Forsythe F(2, 246.56) = 79.68, p\.001, η2p = .39. Androphilic men scored significantly higher than

123

Arch Sex Behav Table 3 Mean and SD, inferential statistics, and significance values for sexual orientation group comparison on the measure of recalled childhood behavior Gynephilic mend

Androphilic womend Androphilic mend

n = 79

n = 80

M Female-Typical Behavior Subscalea,b,c 1.53 Male-Typical Behavior Subscale Paired sample t tests

a,b,c

3.55

One-way ANOVA

n = 143

SD

M

SD

M

SD

F(2, 299)=

.40

3.11

.96

1.96

.64

114.52, p\.001, η2p = .43

.78

2.03

.69

2.82

.85

72.98, p\.001, η2p = .33

t(78) = −18.53, d = −2.08 t(79) = 7.04, d = .79

t(142) = −8.80, d = −.74

a

Gynephilic men differ significantly from androphilic women (p\.001)

b

Gynephilic men differ significantly from androphilic men (p\.001)

c

Androphilic women differ significantly from androphilic men (p\.001)

d

Group’s responses to the Female-Typical Behavior Subscale and Male-Typical Behavior Subscale differ significantly (p\.001)

androphilic women (p\.001, Hedges’ g = .83) and significantly lower than gynephilic men (p\.001, Hedges’ g = −1.10). Within-group analysis indicated that androphilic men scored significantly higher on the Male-Typical Behavior Subscale than on the Female-Typical Behavior Subscale, t(92) = −5.68, p\.001, d = −.59.

Discussion In line with our prediction, Japanese androphilic men recalled engaging in more female-typical behavior and less male-typical behavior in childhood than gynephilic men. Consequently, the results of this study, which was conducted in a non-Western developed country, are consistent with data collected in variety non-Western developing countries (Bartlett & Vasey, 2006; Cardoso, 2005, 2009; Whitam, 1980, 1997) and Western developed countries (Bogaert, 2003; Dunne et al., 2000; VanderLaan et al., 2011, 2015; Zucker et al., 2006; for a review, see Bailey & Zucker, 1995). As such, the results of this study lend further support to Whitam’s (1980) assertion that childhood gender nonconformity is a cross-culturally consistent pattern of psychosexual development for most androphilic males. It is important to note, however, that the degree to which androphilic males recall engaging in gender-nonconforming behavior in childhood appears to vary cross-culturally. Japanese androphilic men recalled engaging in more male-typical behavior and less female-typical behavior in childhood than gynephilic women. Thus, Japanese androphilic men scored intermediately between androphilic women and gynephilic men on measures of recalled childhood female- and male-typical behavior. In contrast, in Samoa, androphilic males (fa’afafine) recalled greater gender-nonconforming behavior than gynephilic men, yet their recalled female- and male-typical behavior did not differ significantly from that of Samoan androphilic women (Bartlett & Vasey, 2006). The question then arises as to why variation occurs

123

in the magnitude with which recalled childhood gender nonconformity is reported by androphilic males cross-culturally. Variation in the expression of male androphilia across cultures may, in part, explain variation in recalled childhood gendernonconforming behavior. The expression of male androphilia typically takes one of two forms, which are related to gender role enactment (VanderLaan, Ren, & Lance, 2013). Cisgender androphilic males occupy the gender role typical of their sex and identify as “men.” In contrast, in many non-Western cultures, androphilic males are recognized as members of a “third” or alternative gender category and are markedly feminine in terms of their gender presentation throughout their lifetime. While Samoan fa’afafine belong to a third-gender category and overwhelmingly exhibit the feminine form, androphilic men in Japan overwhelmingly exhibit the cisgender form of male androphilia. Although it appears that androphilic males recall elevated childhood gender nonconformity, regardless of cultural context, cisgender androphilic males do so to a lesser degree. Whitam (1997) conducted qualitative ethnographic interviews both third gender/feminine and cisgender androphilic males from within the same non-Western cultures (e.g., Brazil, Guatemala, Indonesia, the Philippines, Thailand) and reported that the former recalled more childhood gender-nonconforming behavior compared to the latter. Some researchers (e.g., Gottschalk, 2003; Ross, 1980) might argue that the study’s findings can be accounted for by the selective recall/memory distortion hypothesis, which suggests that males who adopt a “gay” identity are socialized to think that they are more feminine than gynephilic men. Consequently, to have a cohesive self-narrative that fits with social messages about what it means to be “gay,” such males bias their recall so as to develop a linear personal narrative in which there is a logical progression from a feminine childhood to a feminine adulthood (e.g., Fausto-Sterling, 2014; Gottschalk, 2003; Risman & Schwartz, 1988; Ross, 1980). It is important to note that none of the proponents of the selective recall/memory distortion hypothesis have offered any

Arch Sex Behav

supporting empirical evidence. Moreover, there are at least three lines of converging evidence that challenge this hypothesis. First, prospective studies demonstrate a strong association between childhood gender nonconformity and adulthood sexual orientation (Green, 1987; Singh, 2012; Steensma et al., 2013; Wallien & Cohen-Kettenis, 2008; Zucker, 2014). Second,Rieger, Linsenmeier, Gygax,and Bailey (2008) found that, when rating the behavior of androphilic and gynephilic men’s behavior via their childhood videos, judges who were unaware of sexual orientation status of the targets rated androphilic men as more gender nonconforming; these ratings were consistent with androphilic men’s self-report. Third, the consistency with which childhood gender nonconformity is recalled by androphilic males across highly disparate cultures is inconsistent with the conclusion that such behavior is socialized. Additionally, Gottschalk (2003) has suggested that the association between childhood gender nonconformity and adulthood sexual orientation is tenuous because the connection is not invariant among all sexual orientation minority individuals and, thus, may not be biologically based. However, Zucker (2005) rightly noted that childhood gender-nonconforming behavior and adulthood same-sex sexual orientation can be correlated, without the two sharing the same biological causation (although they might). Adopting some of the methodological approaches used in the present study may benefit future quantitative cross-cultural studies. First, in their meta-analytic review of research examining the association between childhood gender-nonconforming behavior and adult sexual orientation, Bailey and Zucker (1995) recommend the use of multi-item scales to increase reliability. Other studies that have assessed gender (non)conformity among non-Western samples have assessed recalled childhood behavior using two or three items and have used either a dichotomous scale or nominations of favorite activities (Cardoso, 2005; Whitam, 1980). We employed 5-point ordinal rating scales of both male-typical and female-typical childhood behaviors, which included five and six items, respectively. Second, we recommend measuring sexual orientation based on sexual feelings as opposed to self-ascribed sexual orientation identity, which is not necessarily concordant with an individual’s sexual feelings (Beckstead, 2001; Dunne et al., 2000). For example, Cardoso (2005) described a group of men in a Brazilian fishing village who self-identified as heterosexual but preferred feminine males (paneleiros), over women, as sexual partners. Moreover, Western categories of sexual and gender identities are not necessarily relevant or even recognized in nonWestern cultures, rendering cross-cultural comparisons of such identities highly problematic. In contrast, androphilic and gynephilic feelings are cross-cultural universals, facilitating their comparisons between cultures. Lastly, we recommend comparing androphilic males to both gynephilic men and androphilic women. To establish a sex difference, both androphilic men and gynephilic women must be included. It can then be determined whether a male sexual orientation difference exists

and, in turn, whether androphilic males are male-typical, female-typical, or intermediate with respect to the variables under consideration. A potential limitation of the present study is that, in the first six field seasons, androphilic males were collected in a gaisen bar. Gaisen is a term that refers to androphilic Japanese men who prefer foreign men (usually Caucasian) as sexual partners. McLelland (2000) found that, as of 1994, the vast majority of Japanese androphilic men definitively rejected foreigners.4 As such, it is possible that our gaisen sample might not be entirely representative of Japanese androphilic men. Indeed, our analysis indicated that androphilic men whowere recruited from the gaisen bar recalled greater male-typical behavior, although similar female-typical behavior, than men recruited from neighborhoods, non-gaisen bars, and the Osaka Gay Pride Festival. As such, although the gaisen androphilic men may not be entirely representative of Japanese androphilic men, the inclusion of these men would provide us with a more conservative estimate of the gender nonconformity of Japanese androphilic men, not an exaggerated one. That being said, all of the between-group and within-group differences remained significant when the gaisen androphilic men were excluded from the direct group comparisons. Nonetheless, it is possible that the androphilic men who were willing to participate in the present study, includingthosewhowererecruitedoutsideofthegaisenbar,would have varied from those who displayed no interest in participating as wellasthosewhoareuninvolvedintheJapanese“gay”community and, thus, would not have been readily accessed using our recruitment methods. Apart from one study of Samoan fa’afafine (Bartlett & Vasey, 2006) and one study of Brazilian paneleiro (Cardoso, 2005), previous quantitative studies examining childhood gender variance and adult sexual orientation have all taken place in cultures in which androphilic males are largely cisgender with respect to their gender presentation. To further determine if cisgender and third gender/feminine androphilic males recall differing levels of male-typical and female-typical behavior, it will be necessary to conduct quantitative research on additional populations of third gender/feminine androphilic males (e.g., Zapotec muxes) (Chin˜as, 1992). Future studies should consider conducting withinculture comparisons between cisgender and third gender/feminine androphilic males if possible. Finally, although no easy task, undertaking a prospective study of childhood behavior and male androphilia in a non-Western culture would be highly valuable. Acknowledgments We thank Eiji Enomoto, Chiji Masafumi, Kiyoshige Murata, Ayumi Sawada, Yoshiko Sawada, Sachiko Shimada, Hideki Shiraume, Takashi Yanai, Ryoko Yoshikawa, and all of the individuals who agreed to participate in our study. Various stages of this research were supported by the University of Lethbridge, Social Sciences and 4

McLelland (2000) analyzed 607 personal ads by Japanese androphilic men in the June 1994 issue of the magazine Barazoku and the majority stated that they were not interested in foreign partners.

123

Arch Sex Behav Humanities Research Council of Canada (SSHRC) Masters Scholarship to LJP, Chinook summer research award to CRW, and SSHRC Insight Grant to PLV. All the authors declare that they have no conflict of interest. Ethical Approval TheUniversityofLethbridgeHumanSubjectResearch Committee approved this research. Participants were required to sign consent forms prior to taking part in the study. All procedures performed in this study were in accordance with the ethical standards of the Canadian Tri-Council PolicyStatement2(2014): EthicalConductforResearchInvolvingHumans.

Appendix

Table 4 Childhood Gender Identity Scale (CGIS) Female-Typical Behavior Subscale As a child (before the age of 12), how often did you do the following? (1) Play with girls? (2) Play with girls’ toys and girls’ games? (3) Take the girls’ role in pretend play such as when playing “house” or imitating a female character? (4) Put on girls’ makeup, or clothes and accessories? (5) Talk and act like a girl? (6) Do girls’ chores? Male-Typical Behavior Subscale As a child (before the age of 12), how often did you do the following? (1) Play with boys? (2) Play with boys’ toys and boys’ games? (3) Take the boys’ role in pretend play such as when playing “house” or imitating a male character? (4) Play rough games and sports? (5) Do boys’ chores? Scale: 1 = never; 2 = less than half the time; 3 = half the time; 4 = more than half the time; 5 = always/all the time; 6 = decline to answer

References Bailey, J. M., & Zucker, K. J. (1995). Childhood sex-typed behavior and sexual orientation: A conceptual analysis and quantitative review. Developmental Psychology, 31, 43–55. doi:10.1037/0012-1649. 31.1.43. Bartlett, N. H., & Vasey, P. L. (2006). A retrospective study of childhood gender-atypical behavior in Samoan fa’afafine. Archives of Sexual Behavior, 35, 659–666. doi:10.1007/s10508-006-9055-1. Beckstead, L.(2001). Curesversus choices: Agendasinsexual reorientation therapy. Journal of Gay and Lesbian Psychotheraphy, 5, 87–115. doi:10. 1300/J236v05n03_07. Bogaert, A. F. (2003). Interaction of older brothers and sex-typing in the prediction of sexual orientation in men. Archives of Sexual Behavior, 32, 129–134. doi:10.1023/A:1022827524721.

123

Cardoso, F. L. (2005). Cultural universals and differences in male homosexuality: The case of a Brazilian fishing village. Archives of Sexual Behavior, 34, 103–109. doi:10.1007/s10508-005-1004-x. Cardoso, F. L. (2009). Recalled sex-typed behavior in childhood and sports preferences in adulthood of heterosexual, bisexual, and homosexual men in Brazil, Turkey, and Thailand. Archives of Sexual Behavior, 38, 726–736. doi:10.1007/s10508-008-9312-6. Chin˜as, B. N. (1992).TheIsthmus Zapotecs: A matrifocal culture of Mexico. Fort Worth, TX: Harcourt Brace Jovanovich College Publishers. Dunne,M.P.,Bailey,J.M.,Kirk,K.M.,& Martin, N.G.(2000). Thesubtlety ofsex-atypicality.ArchivesofSexualBehavior,29,549–565.doi:10.1023/A: 1002002420159. Fausto-Sterling, A. (2014, October 13). The evidence of memory. Boston Review. Retrieved from http://www.bostonreview.net/wonders/faustosterling-evidence-memory. Gottschalk, L. (2003). Same-sex sexuality and childhood gender nonconformity: A spurious connection. Journal of Gender Studies, 12, 35–50. doi:10.1080/0958923032000067808. Green, R. (1987). The “sissy boy syndrome” and the development of homosexuality. New Haven, CT: Yale University Press. Johnson, L. L., Bradley, S. J., Birkenfeld-Adams, A. S., Radzins Kuksis, M. A., Maing, D. M., Mitchell, J. N., & Zucker, K. J. (2004). A parentreport Gender Identity Questionnaire for Children. Archives of Sexual Behavior, 33, 105–116. doi:10.1023/B:ASEB.0000014325.68094.f3. Kinsey, A. C., Pomeroy, W. B., & Martin, C. E. (1948). Sexual behavior in the human male. Philadelphia, PA: W. B. Saunders. McLelland, M. J. (2000). Male homosexuality in modern Japan: Cultural myths and social realities. Richmond, Surrey: Curzon Press. Ministry of Internal Affairs and Communications Statistics Bureau. (2011, October 26). 平成 22 年国勢調査の概要. Retrieved from http://www.stat.go.jp/data/kokusei/2010/kihon1/pdf/gaiyou2.pdf. Rieger, G., Linsenmeier, J. A. W., Gygax, L., & Bailey, J. M. (2008). Sexual orientation and childhood gender nonconformity: Evidence from home videos. Developmental Psychology, 44, 46–58. doi:10. 1037/0012-1649.44.1.46. Risman, B., & Schwartz, P. (1988). Sociological research on male and female homosexuality. Annual Review of Sociology, 14, 125–147. Retrieved from http://www.jstor.org/stable/2083313. Ross, M. W. (1980). Retrospective distortion in homosexual research. Archives of Sexual Behavior, 9, 523–531. doi:10.1007/BF01542156. Singh, D. (2012). A follow-up study of boys with gender identity disorder. Unpublished doctoral dissertation, University of Toronto, Toronto. Statistics Bureau Ministry of Affairs and Communications Japan. (2015). Statistical Handbook of Japan 2015. Retrieved from http:// www.stat.go.jp/english/data/handbook/pdf/2015all.pdf#page=17. Steensma, T. D., van der Ende, J., Verhulst, F. C., & Cohen-Kettenis, P. T. (2013).Gendervarianceinchildhoodandsexualorientationinadulthood: A prospective study. Journal of Sexual Medicine, 10, 2723–2733. doi:10.1111/j.1743-6109.2012.02701.x. Sullivan, A., & Sheffrin, S. M. (2003). Economics: Principles in action. Upper Saddle River, NJ: Pearson Prentice Hall. UnitedNations Development Project. (2013).Human DevelopmentReports: About Human Development. Retrieved from http://hdr.undp.org/en/ humandev/. VanderLaan, D. P., Gothreau, L. M., Bartlett, N. H., & Vasey, P. L. (2011). Recalled separation anxiety and gender atypicality in childhood: A study of Canadian heterosexual and homosexual men and women. Archives of Sexual Behavior, 40, 1233–1240. doi:10.1007/s10508010-9695-z. VanderLaan, D. P., Petterson, L. J., & Vasey, P. L. (2015). Elevated childhood separation anxiety: An early developmental expression of heightened concern for kin in homosexual men? Personality and Individual Differences, 81, 188–194. doi:10.1016/j.paid.2014.03. 018.

Arch Sex Behav VanderLaan, D. P., Ren, Z., & Vasey, P. L. (2013). Male androphilia in the ancestral environment. Human Nature, 24, 375–401. doi:10. 1007/s12110-013-9182-z. Wallien, M. S. C., & Cohen-Kettenis, P. T. (2008). Psychosexual outcome of gender-dysphoric children. Journal of the American Academy of Child and Adolescent Psychiatry, 47, 1413–1423. doi:10.1097/CHI. 0b013e31818956b9. Whitam, F. L. (1980). The prehomosexual male child in three societies: The United States, Guatemala, Brazil. Archives of Sexual Behavior, 9, 87–99. doi:10.1007/BF01542261. Whitam, F. L. (1997). Culturally universal aspects of male homosexual transvestites and transsexuals. In B. Bullough, V. Bullough, & J. Elias (Eds.), Gender blending (pp. 189–203). Amherst, NY: Prometheus.

Zucker, K.J. (2005). Commentary onGottschalk’s(2003) ‘Same-sex sexuality andchildhoodgendernon-conformity:Aspuriousconnection’.Journal of Gender Studies, 14, 55–60. doi:10.1080/0958923042000331498. Zucker, K. J. (2014). Gender dysphoria. In M. Lewis & K. Rudolph (Eds.), Handbook of developmental psychopathology (3rd ed., pp. 683–702). New York: Springer. Zucker, K. J., Mitchell, J. N., Bradley, S. J., Tkachuk, J., Cantor, J. M., & Allin, S. M. (2006). The Recalled Childhood Gender Identity/Gender Role Questionnaire: Psychometric properties. Sex Roles, 54, 469–483. doi:10.1007/s11199-006-9019-x.

123