The Case for Biological Influences: Sex Differences in Aggression, Visual-Spatial Ability, and Sexual Behavior
The Case for Nature
Sex differences in rough-and-tumble play (i.e., mock aggression) and actual aggression occur during children's third year of life, as early as groups of children can be observed in social settings (Maccoby & Jacklin, 1980; Parke & Slaby, 1983). Boys and girls differ not only in actual aggression but also in fantasy aggression as well. One study collected some 500 stories made up by preschoolers. Aggressive and violent themes were present in 87% of the boys' stories, but only in 17% of the girls' stories (Nicolopoulou, 1997); this replicated findings from previous studies (Libby & Aries, 1989; Nicolopoulou, Scales, & Weintraub, 1994). Sex differences in aggressive fantasies are present in older people too. One recent study asked 300 college students to report recent homicidal fantasies (Crabb, 2000). Almost twice as many men (60%) as women (32%) reported fantasies about killing others. As you will recall, meta-analyses have shown that sex differences in aggression decrease with age (Archer, in press; Hyde, 1986; Knight, Guthrie, Page, & Fabes, 2002). All of these findings suggest that there probably is a biological predisposition leading boys to be more physically aggressive than girls, and, if anything, these differences are dampened with age and socialization.
Across cultures men are generally more aggressive than women (D'Andrade, 1966). Meta-analyses of sex differences in aggression have tended to focus on aggression in laboratory settings. However, when studies employ measures that better assess real-life kinds of aggression (e.g., measures such as direct observations, peer reports, teacher reports, self-reports), sex differences tend to be larger (Archer & Mehdikhani, 2004). And perhaps even more true-to-life evidence comes from social statistics. On virtually any measure of real-life aggression-rates of violent crimes, murders, assaults, sexual violence, participation in warfare, and suicide (which can be viewed as self-directed aggression)—men are much more aggressive than women (Kenrick, 1987). In the United States, men are about six times more likely to commit murder than women. The ratio of men's to women's same-sex homicides is remarkably consistent across cultures: about 9-10:1 (Daly & Wilson, 1988). In short, men murder much more than women do, and they mostly murder other men.
Although absolute rates of aggression vary considerably across cultures, sex differences in aggression appear to be relatively invariant across cultures (Archer, in press; Archer & McDaniel, 1995). They are also relatively constant over historical time. European statistics for the past several centuries show consistently that men are up to four times more likely than women to commit violent crimes (Ellis & Coontz, 1990). Finally, despite dramatic changes in gender roles in recent years, sex differences in aggression seem not to have decreased as a result (Knight, Fabes, & Higgins, 1996; Knight, Guthrie, Page, & Fabes, 2002).
Higher levels ol male aggression occur not only in humans but also in other primates. Male primates generally show more rough-and-tumble play, mock aggression, and actual aggression than female primates, and these sex differences appear at an early age (Fagen, 2002; Meany, Stewart, & Beatty, 1985; Lovejoy & Wallen, 1988; Mover, 1976). As is true for humans, aggression displayed by male primates is directed against other males more than females. Thus male primate aggression seems to be related to male-male competition, dominance, and access to mates, all of which are molded by biological evolution. Although sex differences in primate aggression can be influenced by rearing (e.g., rearing in same-sex versus mixed-sex groups), the genera! finding remains that male primates usually are more aggressive than female primates.
Males generally are more aggressive than females in many other mammals as well. One interesting exception to this general pattern is the spotted hyena, whose females are more aggressive and dominant than males (Yalcinkaya et al., 1993), However, unlike females in most mammalian, species, the female hyena has higher testosterone levels than the male. Are testosterone levels in other species related to aggressiveness, and could typically higher male levels of testosterone help explain sex differences in aggression? As described earlier, many experiments show that eliminating testosterone, either through physical or chemical castration, reduces aggressiveness and dominance in male animals. Conversely, providing or increasing testosterone increases dominance and aggressiveness in both female and male animals (Moyer, 1976).
As noted earlier, correlational studies have shown significant links between human aggressiveness (as measured by personality scales, actual aggressive behaviors, or participation in violent crime) and testosterone levels (Dabbs, 2000; Moyer, 1976; Olweus, 1986). Canadian psychologists Angela Book, Katherine Starzyk, and Vernon Quinsey (2001) conducted a meta-analysis of 45 studies on the relationship between testosterone and aggression in humans. Overall, they found a weak but significant relationship. Because measurements of people's testosterone levels tend to be more stable and reliable when taken later in the day than in the morning, these researchers broke down studies by the time of day when testosterone was measured. They found that testosterone correlated more strongly with aggression when it was measured in the afternoon (r = 0.45) or evening (r = 0.38) than in the morning (r = 0.20). (Correlations can range from 0.00, which would represent no relationship between two variables, up to 1.00, which would represent a perfect, straight-line relationship. Expressed in terms of d values, the strength of these testosterone-aggression associations was d = 1.01 for afternoon measurements, d = 0.82 for evening measurements, and d = 0.41 for morning measurements, which classify as moderate to large effects).
Furthermore, the correlation between testosterone and aggression tended to be larger in studies of younger people (age 13 to 35 years) than in studies of older people (over 35 years). This may reflect the fact that testosterone levels decline with age, which may help explain the particularly high levels of aggression observed in young men. Some studies have suggested that high testosterone levels lead human males to be aggressive, particularly when they are provoked by, for example, insults or physical attacks (Christiansen & Knussman, 1987; Olweus, Mattsson, Schalling, & Low, 1980). The fact that there are situational triggers that work in concert with testosterone should not obscure the fact, however, that high levels of testosterone increase the likelihood of male aggression.
To summarize, sex differences in aggression (a) appear early in human development, (b) are consistent across cultures and over time, and (c) are consistent across species. In addition, human aggression is related to testosterone levels, which are much higher in men than in women. All of these pieces of evidence, taken together, suggest that biological factors play a role in producing sex differences in human aggression.
On average, men exceed women on certain kinds of visual-spatial ability (Silverman & Phillips, 1998; see Chapter 1), and some of these differences are large. Men perform particularly well on spatial tasks that require them to mentally transform three-dimensional objects, navigate three-dimensional space, or throw and target moving objects through space (Geary, 1998). Women perform particularly well at spatial tasks that require landmark learning or remembering where objects are located in complex arrays (Choi & Silverman, 2003; Silverman & Eals, 1992). When learning routes, men are more likely than women to use distances and directions ("Go 2 miles down the highway, get off at High Street, and head east"), whereas women are more likely than men to use landmarks and relative directions ("Go down the highway for a while and pass the big white church on your right; when you come to the red brick ti rehouse, turn left") (Choi & Silverman, 2003; Dabbs, Chang, Strong, & Milun, 1998; Joshi, MacLean, & Carter, 1999).
Some studies have found sex differences in preschoolers spatia! abilities (Levine, Huttenlocher, Taylor, & Langrock, 1999; Lunn, 1987; McGuiness & Morley, 1991) and in older children as well (Choi & Silverman, 2003; Kerns & Berenbaum, 1991; Merriman, Keating, & List, 1985). However, stable and substantial sex differences are most often found after early adolescence, when puberty triggers dramatic hormonal changes in boys and girls (Burstein, Bank, & Jarvick, 1980; Johnson & Meade, 1987), After puberty, sex differences in spatial ability remain quite stable (Willis & Schaie, 1988), and these differences seem not to have diminished with changing gender roles (Masters & Sanders, 1993; Voyer, Voyer, & Bryden, 1995). Some studies have found that allowing women to practice spatial tasks can reduce or even eliminate sex differences in performance; however, the results of these studies are mixed (Cherney, jagarlamudi, Lawrence, & Shimabuku, 2003; Masters, 1998; Schaeffer & Thomas, 1998). One meta-analysis concluded that spatial training increases the spatial performance of both women and men, but it did not necessarily eliminate the differences between them (Baenninger & Newcombe, 1989). And even when women do improve, they still may use somewhat different (e.g., more verbal) strategies compared with men (who make more use of mental imagery). (See Halpern,  for a review.)
Sex differences in spatial abilities prove to be quite consistent across cultures. They have been documented in England (Lynn, 1992); Scotland (Berry, 1966; Jahoda, 1980); Ghana (Jahoda, 1980);; Sierra Leone (Berry, 1966); Japan (Mann, Sasanuma, Sakuma, & Masaki, 1990); Norway (Nordvik & Amponsah, 1998); and India, South Africa, and Australia (Porteous, 1965). In addition, they have consistently been reported in studies conducted throughout the United States. (See Voyer, Voyer, & Bryden  for a review.)
Nordvik and Amponsah (1998) assessed spatial abilities in Norwegian college students who were majoring either in science/technology or in social sciences. The science/technology students had much more experience with math classes than the social science students, and they generally performed better on all spatial tests. Nonetheless, sex differences in spatial abilities were equally strong for both groups of students. Thus, specialized training in spatial tasks did not reduce the observed sex differences. The Norwegian study is doubly interesting because Scandinavian countries promote egalitarian gender ideologies. This, however, had no effect on the sex differences in spatial ability reported in this study.
Sex differences in spatial abilities have been observed in a number of other species, including voles (a kind of rodent; Gaulin, 1992; Gaulin & Fitzgerald, 1989) and rats (Seymoure, Dou, Juraska, 1996; Williams & Meek. 1991). Voles are typically studied in naturalistic settings as they navigate their home ranges, whereas rats are more likely to be studied in laboratory settings as they learn mazes. In animals, sex differences in spatial abilities are often explained in terms of evolutionary pressures. For example, male voles have better spatial abilities than female voles, particularly in polygynous species (species In which males have multiple mates). In such species, males have to roam over large ranges of territory to locate their mates. A possible evolutionary explanation for human sex differences in spatial abilities is that ancestral males were more involved in hunting and warfare, which required throwing projectiles and tracking prey and enemies across large territories, whereas ancestral females were more involved in foraging, which required good spatial location memory (Silverman & Phillips, 1998).
Brain structures have been identified that are related to sex differences in spatial ability. Recent research shows, for example, that the hippocampus—a region deep inside the brain—is the site of certain kinds of spatial abilities in both humans and animals (Maguire, Frackowiak, & Frith, 1997). The male meadow vole has a larger hippocampus than the female meadow vole, and this may partly explain observed sex differences in meadow voles' spatial abilities (Jacobs, Gaulin, Sherry, & Hoffman, 1990). The ultimate cause of sex differences in the size and structure of the hippocampus seems to be prenatal or perinatal exposure to sex hormones, particularly testosterone.
Are sex hormones linked to human visual-spatial ability? The answer seems to be, yes. As described earlier, women exposed to high levels of prenatal testosterone because of CAH perform better on spatial tests. Other studies show that the absence of sex hormones (as in Turner syndrome women) or insensitivity to androgens (as in androgen insensitive XY individuals) leads to decreased spatial abilities. And finally, normal variations in levels of sex hormones (testosterone and estrogen) are correlated with people's spatial performance.
In sum, a number of spatial abilities show sex differences in humans, particularly after puberty, and some of these sex differences are large. Sex differences in spatial ability are consistent across cultures, and despite changing gender roles they have remained constant. Sex differences in spatial ability are often observed in other species, and they seem to be related to early exposure to sex hormones, which produce sexually dimorphic brain structures. Human spatial abilities are correlated with sex hormone levels, both in individuals who have experienced early hormonal abnormalities and in men and women with normal variations in sex hormone levels. In sum, a variety of evidence suggests that biological factors contribute to sex differences in spatial abilities.
Men's and women's sexual behaviors differ in a number of ways (see Chapter 1). Men are more interested in casual sex than women, and they engage in various sexual ac tivities more than women do. Men tend to rate youth and beauty in a mate more highly than women do, whereas women rate dominance, material resources, and status in a mate more highly than men do. Finally, men are sexually attracted to women on average, and women are sexually attracted to men on average.
There is considerable evidence that biological factors contribute to all three of these sex differences; however, the focus here is mostly on the last difference (sexual orientation). Because sex differences in sexual behavior do not generally emerge until puberty, the age at which sex differences emerge will generally not be an important piece of evidence for sexual behavior. However, the other three kinds of evidence remain relevant: cross-cultural consistencies, cross-species consistencies, and the relationship between biological factors (brain structures, sex hormones) and sexual behaviors.
Many sex differences in human mate preferences show substantial cross-cuitural consistency, and this suggests that biological factors are at work. University of Texas psychologist David Buss (1989, 1994) conducted a landmark study in which he assessed more than 10,000 people from 37 cultures scattered across six continents. Some of the cultures he studied were preindustrial; others were highly developed (countries like the United States and Canada). Some of the cultures had strong gender roles (e.g., in various Latin American countries); others had more egalitarian gender roles (e.g., Scandinavian countries). Participants from some cultures practiced polygyny (men allowed to have more than one legal mate); others practiced monogamy. Despite all of these variations, however, sex differences in human mate preferences were often quite consistent across cultures. For example, women valued a marriage partner's financial prospects about twice as much as men did, regardless of culture. Men universally preferred mates who were younger than them, and they rated a mate's physical attractiveness to be more important than women did (Buss, 1989; Buss & Schmitt, 1993).
Greater male sexual activity has been documented repeatedly by sex surveys in modern industrialized countries (Oliver & Hyde, 1993; see Chapter 1). Cross-culturally, polygyny is a much more common practice than polyandry (females having multiple mates) (Daly & Wilson, 1983; Symons, 1979). Men seek sexual stimulation through pornography much more than women do (Byrne & Osland, 2000), and men seek sex for pay (from prostitutes) much more than women do (Burley & Symanski, 1981; Kinsey, Pomeroy, & Martin, 1948,1953). All of these male tendencies seem to be true cross-culturally. Sex hormones, particularly testosterone, are related to sex drive and sexual activity levels, both in animals and in men and women (Alexander et al., 1997; Sherwin, 1988). All of this evidence points to the conclusion that biological factors likely contribute to differences in men's and women's sexual interest and activity levels.
Because a person's degree of sexual attraction to men and to women shows such powerful sex differences in humans, it is worth analyzing in detail the evidence for biological influences on this aspect of human sexuality, Sexual object choice (i.e., attraction to men or to women) not only shows huge sex differences, but also is linked to individual differences in masculinity and femininity within each sex (see Chapter 2). Thus sexual orientation is linked to each of the two faces of gender discussed in this book.
Although many sexual behaviors do not emerge until puberty, there are childhood behaviors that predict adult sexual orientation (Bailey & Zucker, 1995). On average, boys who grow up to be gay men are more likely to display feminine behaviors as children. They avoid rough-and-tumble activities, physical aggression, and competitive sports. They like playing with girls, and they often possess the social reputation of being "sissies." In contrast, boys who grow up to be heterosexual show more male-typical interests; they like rough-and-tumble aggressive play and competitive team sports. They also show a more masculine demeanor and prefer to play in all-male groups.
Girls who grow up to be lesbians are more likely to display masculine behaviors as children. They tend to like the kinds of activities that pre-gay boys dislike, and they have the reputation of being tomboys. In contrast, girls who grow up to be heterosexual tend to show more female-typical behaviors; they often like to play with dolls, to play house, and to wear feminine clothes with other girls. The fact that early masculine and feminine behaviors are linked to later sexual orientation suggests that there may be common biological factors that underlie both childhood sex-typed behavior and adult sexual orientation (Bailey, Dunne, & Martin, 2000; Loehlin & McFadden, 2003).
The presence of opposite-sex attraction is universal across cultures. Indeed, this consistency is so taken for granted that social scientists have not much studied it. The incidence of homosexuality (same-sex sexual and romantic attraction) is harder to assess across cultures. Tolerance for homosexual behaviors clearly varies across cultures, and studies of preindustrial cultures suggest that about two thirds have at least some form of accepted or institutionalized homosexual behavior (Ford & Beach, 1951). However, such behavior is usually shown only by a minority of individuals.
Recent sex surveys in North American and European countries provide relatively stable estimates of the percentage of men who identify themselves as gay (around 3% to 5% of the population) and of the percentage of women who identify themselves as lesbian (around 2% or 3%) (Diamond, 1993). Because homosexuality has been stigmatized in many countries and cultures, it seems likely that surveys may underreport gay and lesbian populations. Nonetheless, it seems almost certain that a very large majority of men and women have heterosexual orientations and a relatively small minority of men and women have bisexual or homosexual orientations. Both sex differences in the incidence of lietero sexuality and homosexuality and the incidence of various sexual orientations within each sex show substantial cross-cultural consistency, and this would suggest that biological factors are operating (Bolton, 1994).
Heterosexuality seems to be a norm across species as well as across human cultures. At the same time, many examples of homosexual behavior can be found in lower animals (Bagemihl, 1999). Bonobo chimps are probably the most famous example of pan-sexual primates who freely engage in both heterosexual and homosexual behaviors (sometimes while hanging from trees!) (Parish, 1994, 1996). Given that reproduction is the central engine of Darwinian natural selection, it seems obvious that biological evolution fostered opposite-sex sexual attraction and mating.
From an evolutionary perspective, homosexuality is the puzzle in need of explanation. Various explanations have been proposed (McKnight, 1997; E. M. Miller, 2000). These are among the most promising:
1. Homosexuality is maintained through kin selection (i.e., it aids the survival of genetic relatives of homosexuals).
2. Genes fostering homosexuality, although decreasing reproductive fitness in one sex, may produce offsetting increases in fitness in the other sex.
3. Genes that, in combination, lead infrequently to homosexuality in some individuals may at the same time foster traits that have offsetting reproductive value for most individuals.
4. Certain kinds of homosexual behavior (like that in bonobo chimps) serve nonsexual functions, such as fostering same-sex coalitions and defusing aggressive encounters.
Regardless of which explanation is correct, few doubt that biological evolution has molded the heterosexual majority's sexual attractions and behavior. (See Buss  for a comprehensive review.)
As described earlier, animal experiments have shown that early exposure to androgens masculinizes sexual behavior and deprivation of androgens feminizes sexual behavior. Critics have noted that male and female sexual behaviors in animals are not the same as human sexual orientations. In animals, early hormones affect stereotyped sexual behaviors and reflexes such as mounting and lordosis, but they do not. necessarily determine the object (male or female) of sexual attraction (Breedlove, 1994). However, unless there is a complete discontinuity between humans and lower animals, it seems likely that the hormonal influences that are powerful in channeling animals' sexuality play a role in human sexuality as well.
Research shows that adult levels of sex hormones are not much related to human sexual orientation. Theoretical speculation has focused instead on the effects of prenatal hormones (Ellis & Ames, 1987; Meyer-Balburg, 1984), As described earlier, women exposed to unusual prenatal hormone environments (such as CAH and DES~exposed women) show higher levels of homosexual and bisexual attraction as adults. And studies of homosexual-heterosexual differences on physical traits such as finger length ratios, otoacoustic emissions, and hip-to-waist ratios also suggest that adult sexual orientation is linked to early androgen exposure (Singh, Vidaurri, Zambarano, & Dabbs, 1999). Indeed, a complex array of physical and behavioral traits are linked to sexual orientation, and it seems more and more unlikely that these associations have purely environmental or social causes (Rahman & Wilson, 2003).
Handedness is one of the many traits that are related to sexual orientation (gay and lesbian individuals show a higher incidence of left handedness than heterosexuals do; Lalumiere, Blanchard, & Zucker, 2000; Lippa, 2003b). Left handedness also shows sex differences (more men than women are left-handed), and this difference may be linked to early exposure to testosterone. Another theory is that handedness and sexual orientation are influenced by developmental instability, which occurs when species-typical prenatal development is perturbed by biological and environmental factors such as infectious agents, diet, uterine environments, food supplies, immunological reactions in the fetus's and mother's bodies, and so on. Whatever the causes of higher rates of left handedness in gay and lesbian individuals, they are unlikely to result from social-environmental factors. Rather, such findings again implicate biological factors in the development of sexual orientation.
Recent research shows a relationship between birth order and male homosexuality, with gay men more likely to have older brothers than heterosexual men (Blanchard, 1997; Blanchard, Zucker, Siegelman, Dickey, & Klassen, 1998; Bogaert, 2002, 2003a). The most likely explanation for this finding is again biological, immunological reactions between mothers and their male fetuses are more likely with each succeeding male fetus, and these reactions probably affect prenatal hormone levels or other biological factors that lead to sex-typed behaviors. The older-brother-and-homosexuality link is seen particularly in gay men who have relatively short stature, and this again suggests that there is something biological going on (Bogaert, 2003b).
Behavior genetic studies show that homosexuality is partly heritable. One study found that when one identical twin was gay, there was a 52% chance that the other twin was also gay. For fraternal twins, however, there was only a 22% chance that the second twin was gay (Bailey & Pillard, 1991). Although the results of other studies vary, most find some degree of heritability for homosexuality (Bailey, Dunne, & Martin, 2000; King & McDonald, 1992). Another study assessed lesbian and bisexual women with twins or adoptive sisters (Bailey, Pillard, Neale, & Agyei, 1993). When one identical twin was lesbian or bisexual, there was a 48% chance that the other twin was as well. However, concordance percentages went down to 16% for fraternal twins and to 6% for adoptive sisters.
Family pedigree studies have shown that homosexuality runs in families (Bailey & Bell, 1993; Bailey & Benishay, 1993; Bailey, et al., 1999; Pattatucci & Harrier, 1995; Pillard & Weinrich, 1986). Some research suggests that the pattern of inheritance shows maternal linkage; that is, families with gay male children show an increased number of gay relatives on the mother's side but not on the father's side of the family. This suggests that the X chromosome may be involved, for this chromosome is passed from mothers to sons but not from fathers to sons. Recent molecular genetic studies specify a specific locus on the X chromosome that may be related to sexual orientation (Hamer, Hu, Magnuson, Hu, & Pattatuci, 1993). However, this research awaits replication.
In a much publicized study, neuroscientist Simon LeVay (1991) found that regions of the preoptic area of the hypothalamus (a small structure deep in the brain, next to the pituitary gland) show significant size differences in gay and heterosexual men. These regions also show sex differences, and gay men have preoptic areas more like those of women. Animal studies corroborate the human findings, showing that regions of the preoptic area show sex differences. In addition, these areas are influenced by prenatal hormones, and they are linked to sexual behaviors.
There is a final kind of research that, indirectly, has implications for biological theories of sexual orientation, namely, research on the influence (or lack of influence) that gay and lesbian parents have on their children. Children reared by gay and lesbian parents tend, on average, to be as well adjusted as children reared by heterosexual parents; in addition, studies suggest that these children do not differ much from other children in the likelihood that they will grow up to be gay, lesbian, or heterosexual themselves (Anderssen, Amlie, & Ytteroy. 2002; Patterson, 2002, 2004). However, they may be more open to the possibility of same-sex attraction, and they may be less sex-typed in their levels of sexual activity (Stacey & Biblarz, 2001). Given that children of gay and lesbian parents have quite different sorts of role models (at least when it comes to sexual orientation) than do children of heterosexual parents, the finding that parents' sexual orientation does not have much of an influence on their children's sexual orientation seems to counter simple social learning and role-modeling theories of how sexual orientation develops. (Similarly, it is worth noting that most gay and lesbian individuals grow up in families with heterosexual parents and siblings, who often strongly desire them to be heterosexual.) The elimination of simple social learning explanations for the development of sexual orientation leaves biological theories increasingly in the running.
Taken together, the threads of evidence just summarized suggest that biological factors play a significant role in determining human sexual orientation.