The Future of Sex and Gender in Psychology: Five Challenges to the Gender Binary PDF

American Psychologist © 2018 American Psychological Association 2019, Vol. 74, No. 2, 171–193...

American Psychologist © 2018 American Psychological Association 2019, Vol. 74, No. 2, 171–193 0003-066X/19/$12.00 http://dx.doi.org/10.1037/amp0000307 The Future of Sex and Gender in Psychology: Five Challenges to the Gender Binary Janet Shibley Hyde Rebecca S. Bigler University of Wisconsin—Madison University of Texas at Austin Daphna Joel Charlotte Chucky Tate Tel-Aviv University San Francisco State University This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. This document is copyrighted by the American Psychological Association or one of its allied publishers. Sari M. van Anders University of Michigan The view that humans comprise only two types of beings, women and men, a framework that is sometimes referred to as the “gender binary,” played a profound role in shaping the history of psychological science. In recent years, serious challenges to the gender binary have arisen from both academic research and social activism. This review describes 5 sets of empirical findings, spanning multiple disciplines, that fundamentally undermine the gender binary. These sources of evidence include neuroscience findings that refute sexual dimorphism of the human brain; behavioral neuroendocrinology findings that challenge the notion of genetically fixed, nonoverlapping, sexually dimorphic hormonal systems; psychological findings that highlight the similarities between men and women; psychological research on transgender and nonbinary individuals’ identities and experiences; and developmental research suggesting that the tendency to view gender/sex as a meaningful, binary category is culturally determined and malleable. Costs associated with reliance on the gender binary and recommendations for future research, as well as clinical practice, are outlined. Keywords: gender, sex differences, transgender, neuroscience, social neuroendocrinology From its beginnings in the 1800s, psychological research the early part of the 20th century, Woolley (1910) had and practice firmly espoused the assumption that there are written a review of psychological research on the differ- two and only two categories of people: women and men. By ences between women and men. In the 1930s, psychologists developed the concept of psychological masculinity- femininity (Terman & Miles, 1936) and argued that mascu- linity was necessary for good adjustment for men, as was This article was published Online First July 19, 2018. femininity for women (Pleck, 1981). These approaches are Janet Shibley Hyde, Department of Psychology and Department of Gender & Women’s Studies, University of Wisconsin—Madison; Rebecca based on what is referred to as the gender binary. In S. Bigler, Department of Psychology, University of Texas at Austin; addition to the core belief that there are two discrete cate- Daphna Joel, School of Psychological Sciences and Sagol School of gories into which all individuals can be sorted, the gender Neuroscience, Tel-Aviv University; Charlotte Chucky Tate, Department of binary system also typically assumes that one’s category Psychology, San Francisco State University; Sari M. van Anders, Depart- ments of Psychology, Women’s Studies, and Neuroscience, University of membership is biologically determined, apparent at birth, Michigan; as of July 2018, Departments of Psychology, Gender Studies, stable over time, salient and meaningful to the self, and a and Neuroscience, Queen’s University. powerful predictor of a host of psychological variables. Parts of this research were supported by the Israel Science Foundation Over the past two decades, however, a confluence of (Grant 217/16) to Daphna Joel, and by the National Science Foundation (REC 0635444) to Janet Hyde. The authors thank Kelley Kidwell for forces has challenged psychology’s assumption of the gen- creating Figure 1. Coauthors after Janet Shibley Hyde are listed alphabet- der binary. These forces range from the transgender activist ically. movement (Martinez-San Miguel & Tobias, 2016; Stryker, Correspondence concerning this article should be addressed to Janet 2008) and the intersex activist movement (Dreger & Hern- Shibley Hyde, Department of Psychology, University of Wisconsin— Madison, 1202 West Johnson Street, Madison, WI 53706. E-mail: don, 2009; Reis, 2007) to research in neuroscience and [email protected] psychological science. This article synthesizes research that 171  172 HYDE, BIGLER, JOEL, TATE, AND VAN ANDERS pectations for males and females, which vary as a function of intersections with other factors (Cole, 2009; Else-Quest & Hyde, 2016), as well as psychological processes such as identity, femininity, masculinity, and gender-conformity and nonconformity. The term transgender is used in this article as an umbrella term for individuals who self-label differently than their birth-assigned category (for a list of terms concerning gender that are transgender inclusive, see American Psychological Association [APA], 2015), and cisgender is used to refer to individuals whose self-labeling is the same as their birth-assigned category. Finally, the term nonbinary is used here for individuals who self- This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. identify in ways outside the two categories of female and This document is copyrighted by the American Psychological Association or one of its allied publishers. male (e.g., agender, genderfluid, bigender; Bornstein, 1994; Brooks, 2017; Tate, Youssef, & Bettergarcia, 2014). The Challenge From Neuroscience The gender/sex binary is salient in neuroscience, and Janet Shibley especially in its popularized versions. For example, in her Hyde book The Female Brain, Brizendine (2006) claimed that scientists have documented an astonishing array of structural, challenges the gender binary from multiple perspectives, chemical, genetic, hormonal, and functional brain differences focusing especially on neuroscience, behavioral neuroendo- between women and men. We’ve learned that men and women crinology, research on gender similarities and differences, have different brain sensitivities to stress and conflict.... research on the experiences of transgender individuals, and Women may remember the smallest details of their first date, the developmental psychology underlying the psychological and their biggest fights, while their husbands barely remember process of categorizing by gender. Costs associated with the that these things happened. Brain structure and chemistry have gender binary framework are considered, as are implications everything to do with why this is so. (p. 4) for future research and clinical practice of moving beyond a There are indeed average differences between women and gender binary framework. men in brain structure and function (Lenroot & Giedd, Terminology in this area is complex and controversial.1 2010; Ruigrok et al., 2014). These differences are, however, Some authors have argued that sex should be used for often misinterpreted as innate or preprogrammed, context biologically based differences between males and females, independent, and stable over time (e.g., Joel & McCarthy, whereas gender should be used for differences between 2017). Furthermore, it is often implicitly assumed that these women and men that are produced socioculturally (e.g., differences add up to create two types of brains, one typical Muehlenhard & Peterson, 2011; Unger, 1979; West & Zim- of females and the other typical of males. For this assump- merman, 1987). Others have argued that biological and tion to be accurate, differences between females and males sociocultural factors are typically intertwined, and thus the in the structure of specific brain regions should be both distinction between the terms sex and gender should be highly dimorphic in the population and internally consistent abandoned (Yoder, 2003). In this article, the term gender/ in the individual. It turns out that they are neither. sex is frequently used, to recognize that the biological and the sociocultural are typically inseparable (van Anders, 2015; van Anders & Dunn, 2009). The term sex is used here Sex and Biological Dimorphism to refer to biological systems involving the X and Y chro- Two fundamental assumptions underlie current thinking mosomes, pre- and postnatal sexual differentiation, and about sex as a biological system and about its relations with hormones that influence sexual differentiation of the exter- other systems: (a) that sex is a dimorphic system (i.e., a nal genitals, which, in turn, serve as the basis for sex assignment at birth. Individuals with statistically atypical 1 The five authors have produced, independently, five lines of systematic genitals or internal reproductive structures can be termed empirical work that challenge the gender/sex binary. As is true of the field intersex or sex diverse, and constitute roughly 1% to 2% of of gender/sex research, we—as a group— do not agree completely on the the population (at least among White people; Blackless et terminology used to refer to men and women (i.e., sex versus gender) or some of the tenets of the development of men and women. Interested al., 2000; Lee et al., 2016). The term gender is used here to readers should consult the authors’ respective works for more detailed refer to sociocultural systems that include norms and ex- discussions.  FIVE CHALLENGES TO THE GENDER BINARY 173 the size of this nucleus falls in the female-typical range (Garcia-Falgueras, Ligtenberg, Kruijver, & Swaab, 2011). In terms of internal consistency, data on sex differences in the rodent brain suggest that internal consistency is rare. This is because sex differences in specific brain features can be different, and even opposite, under different environmen- tal conditions, and because these sex-by-environment inter- actions vary across brain features. Consider, for example, sex differences in the density of cannabinoid receptors in the rat hippocampus. Under typical laboratory conditions, the density of these receptors is higher in male than female rats. However, following 3 weeks of mild stress, the sex This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. difference in the dorsal part of the hippocampus is reversed; This document is copyrighted by the American Psychological Association or one of its allied publishers. the density of receptors in females is the same as that seen in nonstressed males, and the density of receptors in stressed males is the same as that seen in nonstressed females (Reich, Taylor, & McCarthy, 2009). In other words, the brain sex difference is context dependent. The experience of stress did not, however, reverse sex differences in the den- Rebecca S. sity of cannabinoid receptors in the entire brain. Different Bigler patterns were found, for example, in the ventral hippocam- pus. This example demonstrates that complex sex-by- environment interactions produce a brain structure that is system that can take one of only two forms), and (b) that the multimorphic rather than dimorphic. effects of sex on other systems (e.g., the brain, gender Similar sex-by-environment interactions have been dem- identity) are characterized by a dimorphic outcome (e.g., onstrated in rodent research for many brain regions (e.g., male vs. female brain, male vs. female gender identity). cortex, amygdala, cerebellum), for many brain features Dimorphism can be observed at the level of a single element (e.g., neuronal density, dendritic morphology, neurotrans- of a system—for example, the gonads are generally dimor- mitter systems), and following many different types of phic: in most cases, they differentiate into just two forms, environmental conditions (e.g., rearing conditions, exposure ovaries or testes. Dimorphism can also characterize systems to drugs) at different points throughout development (from that comprise several dimorphic elements—for example, in utero to adulthood, reviewed by Joel, 2011, 2012). Thus, the internal genitals are generally a dimorphic system be- it is unlikely that brains are internally consistent and dimor- cause, in most cases, humans are born with either a uterus, phic; rather, each brain comprises a unique mosaic of fea- cervix, fallopian tubes, and vagina, or with seminal vesicles, tures, some more common in females and others more vas deferens, epididymis, and prostate. For a system to show common in males (Joel, 2011, 2012). sexual dimorphism, each of its elements should be dimor- A recent study that assessed, for the first time, sexual phic, that is, should exist in only two different forms or dimorphism in the human brain in terms of the degree of categories, one typical of females and the other typical of internal consistency found it to be rare. Specifically, Joel males, and all the elements within an individual should be and colleagues (2015) analyzed different structural mea- internally consistent, that is, either all in the form typical of sures, such as volume, cortical thickness, and connectivity, females or all in the form typical of males (Joel, 2011, 2012, using MRI of over 1,400 human brains from four data sets. 2014). In each dataset, they assessed internal consistency in 7 to 12 features chosen because they showed the largest differences The Human Brain and the Gender Binary (i.e., least overlap) between women and men. For each such Although many studies have reported differences between feature, the researchers defined the forms that were more women and men in brain structure (e.g., Lenroot & Giedd, common in women compared with men (female-end form), 2010; Ruigrok et al., 2014), these differences are not sexu- the forms that were more common in men compared with ally dimorphic; rather, there is considerable overlap be- women (male-end form), and the forms that were similarly tween the distributions of women and men. This is true even common in both women and men (intermediate form). For for regions showing the largest sex differences known to example, for a brain region that was larger, on average, in date. For example, the intermediate nucleus of the human women, the female-end form was defined as volumes in the hypothalamus is about twice as large, on average, in men top 33% for women, and the male-end form was defined as compared with women, yet in approximately 30% of men, volumes in the lowest 33% for men. Next, the researchers  174 HYDE, BIGLER, JOEL, TATE, AND VAN ANDERS The Challenge From Behavioral Neuroendocrinology Belief in a gender binary is found not only in neurosci- ence but also in behavioral neuroendocrinology. This belief involves two assumptions: (1) that gonadal hormones are dimorphic (i.e., that there are “female” hormones, such as estrogen and progesterone, and “male” hormones, such as testosterone), and (2) that levels of these hormones are genetically determined and fixed. Current research in be- havioral neuroendocrinology and, in particular, social neu- roendocrinology, challenges both of these assumptions. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. This document is copyrighted by the American Psychological Association or one of its allied publishers. Androgens, Estrogens, and the Gender Binary Many people mistakenly assume that there are male hor- mones and female hormones, but this idea is challenged by the presence of estrogens (e.g., estradiol) and androgens (e.g., testosterone) in both women and men as well as in gender-diverse (e.g., nonbinary) people, because these hor- Daphna Joel mones, as well as progesterone, are produced by both ova- ries and testes as well as the adrenal glands and through peripheral conversion in fatty tissue; these sources are pres- ent in all bodies (for a review, see Gillies & McArthur, tested, for each brain, whether it was internally consistent— 2010). Another common misunderstanding is that these that is, whether all elements had the female-end form, or all hormones circulate at sexually dimorphic or nonoverlapping had the male-end form, or all had the intermediate form— levels. In actuality, average levels of estradiol and proges- and contrasted internal consistency with mosaicism, that is, terone do not differ between women and men (Liening, having at least one element with the female-end form and at Stanton, Saini, & Schultheiss, 2010; van Anders, 2010). least one element with the male-end form. Regardless of the Moreover, the changes in steroid levels that accompany sample, age, type of imaging, and method of analysis, reproductive phases (e.g., pregnancy, ovulation) highlight mosaicism was much more common than internal consis- the breakdown of the gender binary; for example, nonpreg- tency; mosaicism was seen in 23% to 53% of the brains and nant women have estradiol and progesterone levels more internal consistency in 0.7% to 10.4% of brains, depending similar to men than to pregnant women (Tulchinsky, Hobel, on the sample and specific brain measure. (The remaining Yeager, & Marshall, 1972). Classification on the basis of brains had either combined female-end and intermediate steroid levels (estradiol and progesterone) would make for a features or male-end and intermediate features.) Accord- very different kind of binary, one between pregnant women ingly, sex differences in the human brain do not add up to and everyone else (nonpregnant women, men, and gender- create two types of brain, a male brain and a female brain. diverse individuals). Instead, most brains are gender/sex mosaics. Differences in the levels of these hormones vary across the life span, with no differences during the prenatal period except for one brief span corresponding to genital sexual Summary differentiation, and no differences from birth to adolescence except one period during the first year of life. In short, The division of humans into two categories, females and fetuses and prepubertal children cannot be categorized into males, on the basis of the form of their genitalia is often a gender binary on the basis of androgens and estrogens. accompanied by the assumption that males and females During adolescence, testosterone levels increase in both belong to two distinct categories in other domains; however, boys and girls, but at a much higher average rate for boys current scientific evidence refutes this assumption for the (Gillies & McArthur, 2010). However, the size of this brain. The distributions for men and women on different difference has been mischaracterized; although testosterone brain features are overlapping, and internal consistency levels are higher in men than women, on average, the across features within individuals is rare. Thus, human difference is much smaller than widely believed and the brains are not internally consistent for male-typical and distributions show considerable overlap (Granger, Shirtcliff, female-typical features. Instead, most human brains are a Booth, Kivlighan, & Schwartz, 2004; Liening et al., 2010; mosaic of these features. Overpeck, Colson, Hohmann, Applestine, & Reilly, 1978;  FIVE CHALLENGES TO THE GENDER BINARY 175 fixed characteristic of individuals— even though they may show some trait-like patterns— but instead are a set of changing and interdependent parameters (Wagner, 2006). Evidence for this assertion comes from a large body of research, including research described in the next section. Hormones such as estradiol and progesterone are already understood to show variability within the context of repro- ductive phases in women (e.g., menstrual cycle, pregnancy, menopause). Research on the social modulation of these hormones demonstrates that they vary in response to social context and behaviors as well, often in ways that counter gendered stereotypes of these hormones. For example, en- This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. gaging in dominance contests can increase estradiol and This document is copyrighted by the American Psychological Association or one of its allied publishers. progesterone (Oxford, Tiedtke, Ossmann, Özbe, & Schul- theiss, 2017; Stanton & Schultheiss, 2007). Social closeness also increases progesterone (S. L. Brown et al., 2009), as might social rejection (Duffy, Harris, Chartrand, & Stanton, 2017). Because testosterone is often seen as a biological account Charlotte of differences between men and women, understanding Chucky Tate whether it is innate and fixed, or socially influenced and malleable, is important. Testosterone is understood in many Sapienza, Zingales, & Maestripieri, 2009; van Anders, ways to be the biological essence of sex, accounting for 2010). In fact, the unquestioned belief in the gender binary female–male differences, in general, and maleness, in par- has hampered the study of these hormones because many ticular (Fine, 2017; van Anders, 2013). As such, when researchers have studied “male” hormones (e.g., testoster- testosterone is studied, it is generally studied as the sole one) only in men and “female” hormones (e.g., estradiol and cause of gender/sex differences, and social factors are ex- progesterone) only in women (van Anders, 2013). It is only cluded. In addition, it has traditionally been studied only as recently that researchers have turned to closely examining a cause and not as an outcome of behavior. These biologi- actual empirical data on gender/sex variation and overlap in cally deterministic approaches run counter to scientific ev- adult testosterone levels. idence concerning testosterone, as described next. Is testosterone genetically determined? Testosterone has a relatively high heritability, implying that genetics Androgens and Estrogens: Genetically account for a relatively high proportion of variability in Determined and Fixed? testosterone levels (Harris, Vernon, & Boomsma, 1998; In addition to gonadal hormones being viewed as dimor- Kuijper et al., 2007). But heritability estimates of testoster- phic, this presumed dimorphism is generally assumed to be one also demonstrate a relatively large role for nongenetic genetically determined and fixed. Innate fixedness is an factors, including the environment (e.g., time of day) and important feature of people’s beliefs in the gender binary, social factors (Harris et al., 1998; Kuijper et al., 2007; van and gonadal hormones are presumed to underlie the gender Anders, 2013). Moreover, the genetic contribution is likely binary and contribute to its stability. Research, however, overestimated because it reflects samples taken from people demonstrates that malleability is an important characteristic in the same social and behavioral contexts; that is, if people of biological phenomena, as exemplified by research on were sampled across a wider range of environmental cir- neuroplasticity and epigenetics (Pittenger & Duman, 2008; cumstances, environment would likely account for more Weaver et al., 2004). However, when it comes to gender/sex variance, and heritability estimates would be lower (van and hormones, plasticity and malleability are still largely Anders et al., 2014). Thus, the idea that testosterone re- ignored or presented in ways that fix biology in new critical search supports an innate gender binary because of its periods or early life programming (Pitts-Taylor, 2010; Rich- heritability is challenged by evidence showing that testos- ardson et al., 2014). terone levels are influenced to a considerable degree by The assumption that levels of gonadal hormones are in- nongenetic factors, and by more recent research demonstrat- nate and fixed is challenged by evidence that their levels ing that they dynamically respond to social context. vary widely within individuals and are socially modulated Is testosterone fixed? Research on hormones chal- (e.g., Nyby, 2008; van Anders, Goldey, & Bell, 2014; van lenges the notion of the gender binary as fixed by showing Anders & Watson, 2006). That is, hormone levels are not a that social factors influence testosterone. Social neuroendo-  176 HYDE, BIGLER, JOEL, TATE, AND VAN ANDERS al., 2012). This research not only counters the notion of tes- tosterone as fixed but also expands notions of which behavioral contexts are meaningful for investigations with androgens and challenges their assumed link to masculinity. Indeed, testosterone responses can be parsed more meaning- fully into decreases related to nurturance (involving warm, close, supportive, and/or loving contact) and increases related to competition (involving acquisition of resources, broadly defined), providing a theoretical basis that is grounded in both sociocultural and evolutionary understandings as well as other endocrine systems (via the steroid/peptide theory of social bonds; van Anders et al., 2011). This matters because testos- This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. terone—like other hormones— has evolved to respond to This document is copyrighted by the American Psychological Association or one of its allied publishers. some, but not other, behavioral contexts. The empirical evi- dence that breaks down fixed views of testosterone shows that it is not only socially modulated but also socially situated, with social modulation of testosterone itself grounded in sociocul- tural processes and experiences. Indeed, some social neuroendocrine research has begun to Sari M. van examine how gendered experiences themselves might modu- Anders late hormones, with evidence suggesting that testosterone can respond to gender socialization and gender norms (van Anders, crinology attends to this “reverse relationship” (effects of Steiger, & Goldey, 2015). Thus, hormone research is showing behaviors on hormones) and also focuses on recursive links the ways that gendered expectations and lived experiences can between hormones and social behaviors, all while attending actually shape the very hormones thought to underlie the to social context (van Anders, Goldey, & Kuo, 2011; van essence of femaleness and maleness, again challenging basic Anders & Watson, 2006). A growing number of researchers tenets of the gender binary. use social neuroendocrine framings that take social context into account (e.g., Gettler, McDade, Feranil, & Kuzawa, Summary 2011; Hamilton, Carré, Mehta, Olmstead, & Whitaker, 2015). For example, sexual thoughts increase testosterone Social neuroendocrine research challenges the gender binary levels in women (Goldey & van Anders, 2011); testosterone in multiple ways. Androgens and estrogens are not two distinct responses to sexual thoughts are correlated with the type of sets of sex hormones— one set for women and one set for fantasy content in men (Goldey, Avery, & van Anders, men— but rather hormones that are found in all humans. That 2014); parenting behaviors decrease testosterone but only is, human bodies produce hormones like estradiol, testosterone, when they are nurturant (van Anders, Tolman, & Volling, and progesterone regardless of gender/sex, and levels of estra- 2012); and relationship transitions are recursively linked diol and progesterone are similar in men and women. More- with testosterone (Dibble, Goldey, & van Anders, 2017). over, levels of these hormones are not fixed, but are dynamic A large and growing body of research documents the ways and can be influenced by gendered social experiences. In that social behavioral contexts modulate testosterone levels. humans, for example, testosterone baseline levels across gen- For example, recent research with large samples over multiple der/sex change as a function of multiple environmental factors waves shows strong evidence for sexual and relational modu- (e.g., time of day, season) and social factors (e.g., relationship lation of testosterone and very little for androgenic modulation status, mood; van Anders, et al., 2014). Thus, the idea that of these phenomena (Das & Sawin, 2016). Other research testosterone— or any hormone—is the biological basis of the shows evidence for both directions, as when the presence of gender binary is belied by the scientific research, and chal- many family members in one’s social network is associated lenges to the biology of the gender binary are challenges to the with lower testosterone at later points, and testosterone nega- gender binary itself. tively influences perceived later social support (Das, 2017). Some research shows iterative associations. For example, The Challenge From Psychological Research: many fathers to-be show decreases in testosterone over the Gender Differences and Similarities duration of their partners’ pregnancies, and these decreases in testosterone can predict parenting behaviors (Gettler et al., As noted earlier, for more than a century, psychologists have 2011; Saxbe et al., 2017). But nurturant parenting behaviors, in devoted themselves to research on psychological gender/sex particular, seem to decrease testosterone in men (van Anders et differences. That research rests on an assumption that there are  FIVE CHALLENGES TO THE GENDER BINARY 177 just two categories of people: females and males. Often, results tions, and whether the magnitude of the gender difference are translated into statements about psychological dimor- varies according to factors such as age, ethnicity, and na- phisms, as if there was no overlap in the female and male tionality. Because of the dominance of the gender binary in distributions for the behavior being measured. For example, psychological research, meta-analyses to date have synthe- findings of (often small) average gender differences are trans- sized research examining differences between just two gen- lated into statements such as “Girls and boys play differently. der categories: women and men. They learn differently. They fight differently... They hear The magnitude of a gender difference is typically as- differently” (Sax, 2005, p. 28). In the sections that follow, sessed using the statistic d ⫽ (MM – MF)/sw, where MM is results based on two different methods for synthesizing data on the mean score for males, MF is the mean score for females, psychological gender differences are presented. One is a “mo- and sw is the pooled within-groups standard deviation. Ac- saic” analysis like the one described in the section on neuro- cording to this formula, a positive value of d means that science; the other uses meta-analysis. males scored higher on the measure, and a negative value This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. indicates that females scored higher. With meta-analysis, a This document is copyrighted by the American Psychological Association or one of its allied publishers. d value, or effect size, is computed for each study and then The Psychological Gender Mosaic a weighted average d is computed across all studies. Ac- Do the few behavioral and psychological variables that cording to conventions established by Cohen (1988), d ⫽ show large gender/sex differences add up to create two 0.20 is a small difference, d ⫽ 0.50 is a moderate difference, genders, each with its own dimorphic set of psychological and d ⫽ 0.80 is a large difference. Hyde (2005) added the and behavioral characteristics? This question can be an- interpretation that a d value ⱕ0.10 is trivial. swered by applying an analysis of internal consistency to Figure 1 shows four possible alternatives for the distri- these variables, just as was done to brain regions showing bution of males’ and females’ scores on a trait, which could large gender/sex differences in the neuroscience section of be anything from hippocampus size to mathematics perfor- this article. Joel and colleagues (2015) did exactly that and mance. Panel A shows a very large gender/sex difference found that internal consistency in personality traits, atti- (d ⫽ 5.0)—a dimorphism, in the language of biology, tudes, interests, and behaviors is extremely rare. In contrast, because this difference is so large that there is virtually no most humans possess both feminine (i.e., more common in overlap between the two distributions. Panels B, C, and D women than men) and masculine (i.e., more common in show the overlap of male and female distributions when d ⫽ men than women) psychological characteristics (see also 0.20, 0.50, and 0.80, respectively. Even for d ⫽ 0.80, there Koestner & Aube, 1995; Spence, 1993). Of special interest is still substantial overlap in distributions, that is, the trait is was a dataset of 10 highly gender-stereotyped behaviors in not dimorphic. U.S. college students (boxing, construction, playing golf, Today, numerous meta-analyses of research on psycho- playing video games, scrapbooking, taking a bath, talking logical gender differences are available, and many of them on the phone, watching porn, watching talk shows, and produce surprising results. The three meta-analyses re- using cosmetics). Even for these highly gendered behaviors, viewed here are illustrative. which showed very large gender differences (1.0 ⬍ d ⬍ Mathematics performance. Mathematics is stereo- 2.03), less than 1% of the students exhibited only feminine typed as an area of male superiority, and the implicit asso- or only masculine behaviors, whereas over 55% showed ciation test shows that people associate males and math some combination of both feminine and masculine behav- more closely than they do females and math (Nosek, Banaji, iors (feminine and masculine were defined here as the scores & Greenwald, 2002). Meta-analyses, however, challenge characteristic of the most extreme 33% of women and men, this stereotype. respectively). Thus, although stereotypes of women and One meta-analysis synthesized data from state assess- men clearly exist, individuals who consistently match these ments of U.S. children’s math performance from Grades 2 stereotypes are rare. through 11, based on the testing of more than 7 million children (Hyde, Lindberg, Linn, Ellis, & Williams, 2008). Across grades, d ranged between ⫺0.02 and 0.06. That is, Meta-Analyses of Psychological all the differences were trivial or nonexistent, and the au- Gender Differences thors concluded that girls had reached parity with boys in The statistical technique of meta-analysis allows a much mathematics. A second meta-analysis accumulated data more powerful and nuanced analysis of research on gender from 242 studies, representing the testing of more than 1.2 differences than individual studies (for an overview of million people (Lindberg, Hyde, Petersen, & Linn, 2010). methods, see Borenstein, Hedges, Higgins, & Rothstein, Overall, d ⫽ 0.05, again indicating no gender difference. 2009). Meta-analysis can synthesize dozens or hundreds of Depression. Depression is stereotyped as a female dis- studies and tell us how large a gender difference is, how order, and the expression of depressive symptoms, such as much overlap there is between male and female distribu- sadness and tearfulness, violates male gender role stereo-  178 HYDE, BIGLER, JOEL, TATE, AND VAN ANDERS group, and then declined into the 20s and remained rela- tively stable around 1.8 after that. Even with such a gender disparity, though, an OR of approximately 2.0 means that roughly one third of all depressed people are males, a point that is taken up again in a later section. Sexuality. Evolutionary psychologists have argued that there are large gender differences in sexual attitudes and