Childhood gender nonconformity
Childhood gender nonconformity—behaving like the other sex—is a strong correlate of adult sexual orientation that has been consistently and repeatedly replicated (Bailey & Zucker, 1995). More specifically, childhood gender nonconformity comprises the following phenomena among boys: cross-dressing, desiring to have long hair, playing with dolls, disliking competitive sports and rough play, preferring girls as playmates, exhibiting elevated separation anxiety, and desiring to be—or believing that one is—a girl. In girls, gender nonconformity comprises dressing like and playing with boys, showing interest in competitive sports and rough play, lacking interest in conventionally female toys such as dolls and makeup, and desiring to be a boy. Childhood gender nonconformity usually emerges by preschool age. It is important to add that children are usually considered to be gender-nonconforming only if they persistently engage in a variety of these behaviors, as opposed to engaging in a single behavior once or twice. Furthermore, childhood gender nonconformity is not an either/or trait but, rather, a dimensional one, so the differences we discuss are a matter of degree and not of kind.
Children who grow up to be nonheterosexual are substantially more gender nonconforming, on average, than children who grow up to be heterosexual. There are two types of studies supporting this. In retrospective studies, homosexual and heterosexual (and sometimes bisexual) adults are asked about their childhood behavior. In prospective studies, extremely gender nonconforming children are followed into adulthood to assess their sexual orientations. Both kinds of studies have limitations that in isolation might lead to skepticism about their validity. Retrospective studies rely on childhood memories, which might be distorted in a way that supports stereotypes about homosexuality and heterosexuality. Prospective studies often focus on highly unusual children, and one might worry about generalizing findings from these studies to less atypical children. Nonetheless, the two kinds of studies have produced highly convergent findings that support a strong association between childhood gender nonconformity and adult nonheterosexuality.
A 1995 review of retrospective studies found large differences in retrospectively reported childhood gender nonconformity between homosexual and heterosexual men (32 studies; d = 1.3) and women (16 studies; d = 1.0; Bailey & Zucker, 1995). The review estimated that with respect to male sexual orientation, 89% of homosexual men exceeded the heterosexual median score, whereas 2% of heterosexual men exceeded the homosexual median. For female sexual orientation, the respective figures were 81% and 12%. Although this review was published 20 years ago, there is little reason to suspect that its results would change much if updated. Furthermore, despite the skepticism of some writers (e.g., Fausto-Sterling, 2014), no study has persuasively demonstrated that retrospective studies exaggerate orientation differences associated with childhood gender nonconformity. Retrospective self-report measures have been supplemented by childhood home videos provided by heterosexual and homosexual adults, with similar results: Viewers of the videos could tell, at far better than chance levels, which children would grow up to be homosexual and which would grow up to be heterosexual (Rieger, Linsenmeier, Gygax, & Bailey, 2008). The retrospective studies suggest that some degree of childhood gender nonconformity is a common precursor of adult homosexuality in both sexes. This conclusion holds for both Western and a wide array of non-Western cultures, including those in Asia, Latin America, Polynesia, and the Middle East (Bartlett & Vasey, 2006; Cardoso, 2005, 2009; Vasey, VanderLaan, Gothreau, & Bartlett, 2011; Whitam & Mathy, 1986).
Several prospective studies of highly gender nonconforming children have been conducted, most focusing on boys (for reviews, see Bailey & Zucker, 1995; Zucker, 2014). These studies have identified children through clinical interventions to help the children accept their birth sex identity, as opposed to obtaining sex reassignment surgery in adulthood. Thus, in general, these children exhibited extreme gender nonconformity as well as discomfort with or confusion about their gender identity. For males, all studies have found most men to be homosexually or bisexually oriented at follow-up,8 with rates ranging from 64% (N = 129; Singh, 2012) to 80% (N = 66; Green, 1987). The two studies including females were smaller and had more variable outcomes: 32% of 25 subjects (Drummond, Bradley, Peterson-Badali, & Zucker, 2008) and 100% of 10 subjects (Wallien & Cohen-Kettenis, 2008) endorsed bisexual or homosexual feelings. Not all of the studies had control groups, but all the rates of nonheterosexual outcomes far exceed plausible epidemiological estimates for the general population, which might be generously estimated at 5% (see Fig. 1). Thus, the prospective studies also suggest a large association between (extreme) childhood gender nonconformity and adult nonheterosexuality. In nearly all the studies, there was a significant tendency for subjects self-reporting as heterosexual to be younger than those self-reporting as nonheterosexual. This suggests that the prospective studies may underestimate the likelihood of a nonheterosexual outcome because younger nonheterosexual individuals are more likely to provide false claims that they are heterosexual (Bailey & Zucker, 1995).
Both retrospective and prospective studies support a strong association between childhood gender nonconformity and adult nonheterosexuality. This does not mean, of course, that all nonheterosexual individuals were more gender nonconforming than average, much less that they were all extremely gender nonconforming. There are at least three potentially important implications, however. First, long before sexual attraction emerges, some children who will become nonheterosexual are markedly different than other children. Indeed, childhood gender nonconformity is often evident by age 2 (Cohen-Kettenis & Pfäfflin, 2003). Second, there is little evidence that gender nonconforming children have been encouraged or taught to behave that way; rather, childhood gender nonconformity typically emerges despite conventional socialization. Third, the differences between children who will become nonheterosexual and those who will become heterosexual are related to sex-typed behavior and gender identity (Bailey & Zucker, 1995; Zucker & Bradley, 1995). The content of these differences provides some potential clues about the causes of sexual orientation, which we discuss further below.
Onset of sexual attraction
Children cannot know their sexual orientations until they experience sexual attraction toward the same or the other sex. When does this happen? At present in North America, it is impossible to study actual children as they become sexually aware. Such a study would be a political minefield if not an ethical one. Therefore, the onset of sexual attraction has been studied using retrospective reports of adolescents and adults.
Studies have shown that subjects recall first having feelings of sexual attraction at age 10, on average (McClintock & Herdt, 1996). Male and female subjects report similar ages. Importantly, so do homosexual and heterosexual subjects. Age 10 is several years before the typical age of onset of sexual activity (Cavazos-Rehg et al., 2009). This is consistent with an earlier retrospective study showing that homosexual men and women recalled their first homosexual feelings as preceding their first homosexual experiences by 3 years (Bell, Weinberg, & Hammersmith, 1981). Although this gap may have changed since that publication, the important point—to which we return—is that the large majority of nonheterosexual people recall that homosexual desires preceded homosexual experiences.
How do nonheterosexual people experience the recognition of homosexual feelings? In a retrospective study, homosexual men reported that their early sexual attractions were not necessarily a source of any distress (Savin-Williams, 1996). Same-sex attractions were often experienced as an obsession with being near masculine, often older, same-sex teenagers and adults, such as male teachers and coaches. Eventually, these men recognized that their same-sex desires were rarely shared by others, were not condoned, and should be hidden. Although first homosexual male experiences can occur in the context of romantic relationships, it is not uncommon for these experiences to occur with strangers (Savin-Williams & Diamond, 2000).
On average, homosexual women describe their early same-sex attractions as more emotionally than sexually charged. Lesbian or female bisexual self-identification often precedes the onset of homosexual activity, and first same-sex sexual attractions and behaviors are often experienced within the context of a romantic relationship (Savin-Williams & Diamond, 2000).
Bisexual-identified individuals typically experience later recognition of same-sex attractions compared with homosexual men and women. This may reflect the fact that bisexual identity usually follows the establishment of a heterosexual identity (Fox, 1995; Weinberg, Williams, & Pryor, 1994). Homosexual people are atypical in two ways: their lack of attraction to the other sex and their increased attraction to their own. In contrast, bisexual people are atypical in only the latter sense, and this may delay their recognition of their own difference.
Adult gender nonconformity
It would be surprising if the large orientation differences in childhood gender nonconformity disappeared without a trace by adulthood. Several differences between heterosexual and nonheterosexual adults persist. They include patterns of occupational and recreational interests (Lippa, 2005a, 2005b). Research indicates that heterosexual men have greater interest in occupations and hobbies focusing on things and less interest in those focusing on people, compared with heterosexual women (Lippa, 2005a; Su, Rounds, & Armstrong, 2009). In contrast, homosexual men show a somewhat feminine pattern of interests, and homosexual women a somewhat masculine one. The orientation differences are large, although smaller than the sex differences. They are consistent with stereotypes about occupational differences between homosexual and heterosexual people. This makes them more difficult to interpret than childhood differences because both homosexual and heterosexual adults may—either consciously or unconsciously—mold their behavior in accordance with societal expectations.
Other differences between homosexual and heterosexual adults include patterns of movement (i.e., gestures and walking) and speech (i.e., articulation), physical presentation (i.e., clothing choices and hairstyles; Bailey, 2003; Rieger et al., 2008; Rieger, Linsenmeier, Gygax, Garcia, & Bailey, 2010), and even facial appearance (Rule, Ambady, Adams, & Macrae, 2008). These differences also tend to be large and help explain the phenomenon by which individuals can sometimes judge a target’s sexual orientation by observing superficial aspects of the target’s nonsexual behavior (“gaydar”). As with gender nonconformity in interests, gender nonconformity in movement, speech, and physical presentation might be influenced by societal expectations. Homosexual and heterosexual individuals also tend to differ in their performance on cognitive and personality tests that show sex differences. For example, studies of visuospatial abilities and verbal fluency show that homosexual individuals are shifted in the direction of the other sex (M. Peters, Manning, & Reimers, 2007; Rahman, Abrahams, & Wilson, 2003). Similarly, studies of neuroticism9 have indicated that homosexual individuals are shifted in a sex-atypical direction (Lippa, 2008). It is important to note, however, that the effect sizes for cognitive ability and personality are much smaller—both for sex and sexual orientation—than the effect sizes for patterns of interest.
Before leaving the topic of gender nonconformity, we address a commonly raised question: Might the gender-atypicality of adult homosexual men and women simply reflect a culturally influenced self-fulfilling prophecy? In other words, given that society expects homosexual individuals to be gender atypical, and given that LGB communities often support and facetiously celebrate such gender atypicality, perhaps some homosexual people adopt gender-atypical characteristics to conform to their own stereotypes. Because of the evidence we have reviewed—indicating that gender nonconformity often begins before a prehomosexual child even has a sexual orientation or is aware of cultural stereotypes, and that the link between gender nonconformity and nonheterosexual orientation has been found in a wide variety of cultures—we think it is highly unlikely that gender nonconformity in LGB populations represents a self-fulfilling prophecy due to cultural beliefs. It is possible, however, that cultural stereotypes sometimes amplify gender nonconformity among LGB people. Many LGB individuals report that they have always been fairly gender-typical in dress, appearance, and interests. It is possible that as these individuals come to identify as LGB and participate in the LGB community, they adopt aspects of gender-atypicality.