The “mostly heterosexuals” make up a variety of people. As the authors point out, some may actually be bisexuals or homosexuals who refrain from identifying as a sexual minority; some could be completely heterosexual “but claim a nonheterosexual label, attraction, or behavior for reasons other than their sexual orientation, such as liberal social views, political correctness, or a desire for the gaze of the same sex.” Some may have answered out of confusion or failure to understand the question.28 There is no way to estimate the size of these groups from the data at hand.
Those who genuinely fit the authors’ definition of “mostly heterosexual” can include those who have never acted on their same-sex attractions. Here are the two examples of “mostly heterosexual” respondents that the authors described in the article:
For example, an 18-year old New England girl identified as mostly heterosexual because “I sort of like that it doesn’t just have a completely or just a bisexual, but it has in between… there isn’t always that black and white picture.” A boy in the same study explained, “I’m basically attracted to girls, but I’ve felt like kind of attracted to guys before, but not to like some great extent.… I’ve never felt I was attracted enough to a guy to like go out with them or something like that or like having a relationship with a guy.”29
The authors presented data on 60 studies published from 1994 to 2012 with samples from the United States, Canada, the United Kingdom, Norway, Australia, and New Zealand and including a wide range of age groups. The authors of the Bailey survey article used a subset of these studies that comprised 71,190 adult males and 117,717 adult females. Here are the results by sex:30
Completely heterosexual
Women: 86.8%
Men: 93.2%
Mostly heterosexual
Women: 10.1%
Men: 3.9%
Bisexual
Women: 1.4%
Men: 0.6%
Mostly homosexual
Women: 0.5%
Men: 0.7%
Completely homosexual
Women: 0.6%
Men: 1.4%
The results reinforce the point made earlier: The estimates of homosexuality for the United States from the Gallup polls and the General Social Survey are much higher than those from other studies. For current purposes, the main point of the results is the small proportion of people who self-identify as “mostly heterosexual” and the extremely small proportion who self-identify as bisexual.31 Are bisexuals undercounted by these studies? Ritch Savin-Williams and Kenneth Cohen argue that the answer is yes, giving a “best estimate” (the authors’ phrase) of male bisexual prevalence of about 10–20 percent.32
It’s partly a technical issue: The two “mostly” categories actually identify bisexuals. If the table above had done that, the male bisexual figure would have been 5.2 percent instead of 0.6 percent. The figure would be augmented by the small percentages of men who do not answer “bisexual” but identify as pansexual, transgender, fluid, questioning, or kink-oriented. In most cases these individuals are deleted from the calculations of prevalence rates. Savin-Williams also argues that some men who don’t answer the question are actually bisexual, and points to nonresponse rates in national studies that sometimes reach 10 percent.
The problem ranked first on Savin-Williams’s list is definitional. He argues that a major contributor to the undercount is the failure of many assessments to include “nonsexual domains such as romantic attractions, infatuations, and relationships.”33 But should romantic but nonsexual attraction be classified as evidence of bisexuality? For example, consider a man who had a one-time crush on another man as a teenager but never was sexually aroused by him and has had enthusiastic sex exclusively with women all his subsequent life. In trying to characterize the sexual orientation of the population, it doesn’t seem realistic to categorize him as bisexual. On the other hand, one may argue that a man who has had many such crushes on men throughout his life but never acted upon them is psychologically bisexual.
No matter how detailed the data collection becomes, people with different perspectives will make different judgments about such cases. A common assumption in the technical literature is that to move someone off the “completely heterosexual” end of the continuum, “attraction” needs to include actual sexual arousal by someone of the same sex. If there’s no arousal, it doesn’t count. I share that position. A same-sex attraction without any desire for sex is too easily confused with the genuine but nonerotic love that heterosexual women can have for other women and heterosexual men can have for other men.
The Transgender Phenomenon
Transgendered identity has become a major issue in academia and the media. In the popular understanding, transgender people see themselves as trapped in the body of the other sex. It’s more complicated than that.34
Childhood-onset gender dysphoria. This comes closest to the popular understanding, and it occurs in both girls and boys. It is characterized by behavior typical of the opposite sex from an early age. In the published literature, findings are that the majority (60–90 percent) of children with this kind of gender dysphoria have become comfortable with their sex of birth by adulthood, and thus have avoided the need for a sex change.35 Adults whose childhood-onset gender dysphoria persists typically do report that they have felt more like the other sex from their earliest memories. Childhood-onset gender dysphoria is highly correlated with adult sexual attraction to one’s birth sex, especially in males.
Autogynephilic gender dysphoria. This form occurs only in males. Autogynephilia taken alone refers to sexual arousal at the idea of having a female body or behaving like a female. It is usually a gradual process that begins in adolescence or later. Many men with autogynephilia live their lives as heterosexual men with wives and children while often engaging in practices such as cross-dressing in private. Males with autogynephilic gender dysphoria (as distinct from autogynephilia alone) are more likely to live as homosexuals and often possess a strong desire to obtain sex reassignment surgery, after which they often identify as lesbian.36
Rapid-onset gender dysphoria (ROGD). This is a newly documented phenomenon that might not even have existed a decade ago. It is characterized by adolescents, mostly female, who showed no signs of gender dysphoria as children and apparently abruptly decide they are transgender as teenagers.
As I write, only one systematic study of the ROGD phenomenon is available, based on 256 surveys from parents who had reported that their teenage and young adult children had exhibited rapid onsets of gender dysphoria.[37] The author, Lisa Littman, a gynecologist specializing in gender dysphoria, reported that none of the parents’ children in the study (83 percent were girls) would have been likely to have met diagnostic criteria for gender dysphoria in childhood. They did have a variety of psychiatric problems, however. Many (48 percent) had experienced a traumatic event prior to their declaration of gender dysphoria, and 63 percent had one or more diagnoses of a psychiatric or neurodevelopmental disability.
The results implicated an important role for social influences. Sixty-seven percent of the adolescents had one or more friends who declared they were transgender at about the same time. Increased time on social media was common, and 69 percent of the parents believed (for reasons they explained) that their children were using language that they had found online. “Within friendship groups,” Littman writes, “the average number of individuals who became transgender-identified was 3.5 per group. In 36.8% of the friend groups described, the majority of individuals in the group became transgender-identified.… Parents described intense group dynamics where friend groups praised and supported people who were transgender-identified and ridiculed and maligned non-transgender people. Where popularity status and activities were known, 60.7% of the [children] experienced an increased popularity within their friend group when they announced a transgender-identification and 60.0% of the friend groups were known to mock people who were not transgender or LGBTIA (lesbian, gay, bisexual, transgender, intersex,
or asexual).”38
Littman’s is the only systematic study, but two leading scholars of LGBTIA issues, Michael Bailey and Ray Blanchard, have expressed some preliminary thoughts based on interviews with mothers of ROGD children and clinicians who work with ROGD children and parents, case studies, and the limited quantitative data. They surmise that “ROGD is a socially contagious phenomenon in which a young person—typically a natal female—comes to believe that she has a condition that she does not have. ROGD is not about discovering gender dysphoria that was there all along; rather, it is about falsely coming to believe that one’s problems have been due to gender dysphoria previously hidden (from the self and others).”39
One thing seems clear: The rise in ROGD is concentrated among adolescent females. Historically and internationally, males have had higher rates of gender dysphoria than females. For childhood-onset gender dysphoria, this continues to be true in data from Canada, the Netherlands, and the UK.40 But these same sources also found that the ratio reversed itself among adolescents during the period 2006–10.41 In the UK, which operates the largest child and adolescent gender services in the world, the Gender Identity Development Service (GIDS), the sex ratio has become rapidly more disproportionate. In 2010, 52 percent of adolescent referrals were female; by 2016, that figure had risen to 72 percent.42 The UK data also show a steep rise in referrals of adolescents, from 92 in 2010 to 1,497 in 2016, but it is difficult to disentangle how this growth is divided among increases in incidence, awareness of the existence of the service, and GIDS’s area of coverage.
Taken together, how many people qualify as transgender? The fifth (2013) edition of the Diagnostic and Statistical Manual of Mental Disorders of the American Psychiatric Association put the incidence rate at 0.5–1.4 percent for males and 0.2–0.3 percent for women. A 2015 meta-analysis of prevalence studies reached much smaller estimates of 0.0068 percent for women and 0.0026 percent for men.43 Kenneth Zucker pulled together the available epidemiological data in a 2017 article and reached an overall estimate of 0.5–1.3 percent.44 Whichever of these numbers is closest to the truth, only a small proportion of such people have undergone the major sex reassignment surgery that is popularly associated with transgender identity.45
I have left many aspects of sexual identity unaddressed not because they are intrinsically unimportant but because of my limited objective: The first five chapters of Human Diversity discuss sex differences, making it important to establish that those chapters are not ignoring large proportions of people who do not fit the ordinary definition of male or female. The evidence is incomplete in many respects, but not the basics. Homo sapiens is a typically dimorphic species with regard to sexuality. It consists of two sexes with an extremely small proportion of biological anomalies. Self-identified homosexuality or bisexualism in the United States is somewhere around 5 percent and usually less than that elsewhere. There is room for disagreement about the precise percentages, but not about the truth that humans are a normally dimorphic species.
Appendix 3
Sex Differences in Brain Volumes and Variance
This appendix deals with two important ways in which males and females differ biologically: larger male brain volumes and generally greater male variance in a wide variety of physiological and cognitive traits.
Brain Volumes
The biological reality of larger male brain volumes has been established, but no consensus has been reached on what they mean.
Human Males Have Much Larger Mean Brain Volume Than Human Females
In the 1990s, the first studies of sex differences in brain volume using MRI scans began to be published.1 By 2014, the literature had grown so large that an international team of neuroscientists drawn primarily from Cambridge University (first author was Amber Ruigrok) could publish “A Meta-analysis of Sex Differences in Human Brain Structure” that combined 77 different studies involving 14,597 individuals.
The table below shows the results for the basic volume measures.
META-ANALYSIS OF SEX DIFFERENCES IN THE MAJOR VOLUMES
Volume: Intracranial volume
Studies: 77
Sample size: 14,957
Mean difference (ml): 135.3
Percentage difference: 12.0
Effect size (d): –3.03
Volume: Total brain volume
Studies: 31
Sample size: 2,532
Mean difference (ml): 131.0
Percentage difference: 10.8
Effect size (d): –2.10
Volume: Cerebrum
Studies: 22
Sample size: 1,851
Mean difference (ml): 51.1
Percentage difference: 9.8
Effect size (d): –3.35
Volume: Gray matter
Studies: 60
Sample size: 7,934
Mean difference (ml): 56.5
Percentage difference: 9.4
Effect size (d): –2.13
Volume: White matter
Studies: 57
Sample size: 7,515
Mean difference (ml): 44.4
Percentage difference: 12.9
Effect size (d): –2.06
Volume: Cerebrospinal fluid
Studies: 35
Sample size: 4,484
Mean difference (ml): 18.7
Percentage difference: 11.5
Effect size (d): –1.21
Volume: Cerebellum
Studies: 19
Sample size: 1,842
Mean difference (ml): 7.8
Percentage difference: 8.6
Effect size (d): –1.68
Source: Ruigrok, Salimi-Khorshidi, Lai et al. (2014): Table 3.
The effect sizes for two of the measured volumes qualify as “very large” by Cohen’s guidelines, while the other five of them are “huge.”
Sex Differences in Brain Volume Extend to the Subcortical Regions
Good information on volumes of the regions of the brain in vivo had to wait for MRI technology. The last half of the 1990s saw more than a dozen early studies showing that a variety of subcortical volumes were larger in males, but small sample sizes left uncertainty about the consistency and magnitude of the differences.2 That problem continued into the 2000s. In the Ruigrok meta-analysis of 2014, the median sample size of the 25 studies that examined sex differences in specific regions was just 86 and only 10 of them had samples of more than 100.3
In 2018, many of the uncertainties of earlier work were put to rest by a team of British scholars (first author was Stuart Ritchie) who analyzed subcortical regional means and variances based on a sample of 5,216 persons from the UK Biobank data.4 The average male total brain volume was 1,234 cm3 compared to 1,116 cm3 for females, which amounts to a difference of 118 cm3, representing an effect size of –1.41, smaller than the –2.10 effect size in the Ruigrok meta-analysis.
The Ritchie study reported volume estimates for the 68 subcortical regions in the Desikan-Killiany neuroanatomical atlas.5 Males had larger volumes in all 68. The smallest effect size was –0.24 and the largest was –1.03, with a mean of –0.67. All were highly statistically significant. The table below shows the Ritchie results for some of the most important subcortical regions. Highly statistically significant effect sizes are shown in boldface.
EFFECT SIZES (D) FOR RAW AND ADJUSTED DIFFERENCES IN BRAIN VOLUMES
Total brain
Raw: –1.41
Adjusted for…
Height: –0.42
Total volume:
Gray matter
Raw: –1.28
Adjusted for…
Height: –0.31
Total volume: *
White matter
Raw: –1.49
Adjusted for…
Height: –0.47
Total volume: *
Left hippocampus
Raw: –0.55
Adjusted for…
Height: –0.17
Total volume: –0.02
Right hippocampus
R
aw: –0.54
Adjusted for…
Height: –0.18
Total volume: –0.01
Left accumbens
Raw: –0.39
Adjusted for…
Height: –0.06
Total volume: +0.08
Right accumbens
Raw: –0.31
Adjusted for…
Height: –0.05
Total volume: +0.10
Left amygdala
Raw: –0.59
Adjusted for…
Height: –0.21
Total volume: –0.18
Right amygdala
Raw: –0.51
Adjusted for…
Height: –0.18
Total volume: –0.18
Left caudate
Raw: –0.66
Adjusted for…
Height: –0.20
Total volume: –0.01
Right caudate
Raw: –0.65
Adjusted for…
Height: –0.19
Total volume: 0.00
Left pallidum
Raw: –0.77
Adjusted for…
Height: –0.24
Total volume: –0.16
Right pallidum
Raw: –0.78
Adjusted for…
Height: –0.24
Total volume: –0.12
Left putamen
Raw: –1.01
Adjusted for…
Height: –0.30
Total volume: –0.22
Right putamen
Raw: –1.08
Adjusted for…
Height: –0.35
Total volume: –0.25
Left thalamus
Human Diversity Page 42