Correlational Studies of Hormones and Behavior
The Case for Nature
People with various kinds of genetic and hormonal abnormalities provide scientists with important information about hormones and gender. Studies of normal human variations in sex hormones—particularly variations in testosterone—also provide valuable information.
Do prenatal differences in testosterone levels lead to sex differences in behavior?
Testosterone and Human Behavior
High levels of testosterone, in both humans and lower animals, are associated with aggressiveness (Archer, 1991; Benton, 1992; Book, Starzyk, & Quinsey, 2001). In one study of more than 700 male prison inmates, psychologist James Dabbs and his colleagues found that high-testosterone inmates were more likely to have committed violent crimes than low-testosterone inmates. They were also more likely to have broken prison rules and to have acquired the reputation of being tough and mean (Dabbs, Carr, Frady, &Riad, 1995).
A U.S. government study provided additional information about aggression and testosterone when it assessed more than 4,000 Vietnam veterans on a host of psychological and physiological measures, included serum (blood) testosterone levels. Analyses subsequently showed that veterans with high levels of testosterone were considerably more likely to report delinquent behaviors as children. They also reported higher adult rates of drug and alcohol use, greater numbers of sexual partners, and more participation in active combat during the Vietnam conflict (Dabbs & Morris, 1990).
Another study showed that college fraternities whose members had lower average testosterone levels were more civilized and polite to female experimenters, whereas fraternities whose members were high on testosterone were more likely to live up to the Animal House caricature of being crude, rude, and coarse (Dabbs, Hargrove, & Heusel, 1996). A recent study from Germany found that higher testosterone levels were associated with higher levels of aggressive and delinquent behaviors in at-risk adolescent boys (Maras et al., 2003).
Men with high levels of testosterone are less likely to get married, and when they do marry, they are more likely to have unhappy marriages that end in divorce (Booth & Dabbs, 1993). They tend also to be less successful in their jobs than men with lower levels of testosterone, and this may result from their impatience, impulsiveness, and aggressiveness. James Dabbs (2000), a psychologist who has spent years studying the effects of testosterone, noted that high testosterone can have paradoxical effects on occupational success. On the one hand, it can foster dominance, risk-taking, and bravado, which help in some occupations such as acting, professional athletics, trial law, or military combat. On the other hand, it can lead to imprudent, reckless, and just plain obnoxious behavior, which interferes with performance in more sedate occupations. Dabbs (2000) noted:
High levels of testosterone evolved when the human race was young and people needed the skills of youth. High testosterone helped them compete, but it also led them to take risks, fight, get injured, and die young—and now it interferes with many modern activities. High-testosterone individuals are energetic but impatient; they do poorly in school and end up with fewer years of education; they can dominate others in face-to-face meetings, but they have trouble handling the complexities of business; they lean toward harsh and competitive activities and away from subdued and thoughtful ones. High testosterone is a drawback when careful planning, reliable work habits, and patience are needed, or when workers must attend to the needs of others. Except for a few of the top jobs in sports and acting, high testosterone, to my knowledge, does not contribute to financial success.
Interestingly, high-testosterone men display different sorts of nonverbal behaviors than low-testosterone men. They smile less in general and appear meaner, harder, and more threatening than low testosterone individuals (Dabbs, 1997). When they do smile, high-testosterone men seem less warm and sincere, in part because they smile with their mouths but not with their eyes. One study found similar effects for women (Cashdan, 1995). The voice pitch of high-testosterone men tends to be lower than that of low-testosterone men, but this effect has not been observed in women (Dabbs & Malinger, 1999).
There is evidence that human testosterone levels are related to cognitive abilities as well as social behaviors. Some studies find that high testosterone levels are linked to better visual-spatial abilities (Silverman, Kastuk, Choi, & Phillips, 1999). According to other studies, however, the relationship between testosterone and spatial abilities follows an inverted U pattern, with both very low and high levels of testosterone associated with lower spatial ability and moderate levels associated with higher abilities (Nyborg, 1983; Tan & Tan, 1998). Because women have much lower average levels of testosterone than men, this implies that testosterone may be positively associated with women's spatial abilities, but negatively associated with men's (Gouchie & Kirnura. 1991; Ostatnikova, Laznibatova, & Dohnanyiova, 1996; Petersen, 1976; Shute, Pellegrino, Hubert, & Reynolds, 1983). Exactly this pattern was found in a study that correlated prenatal testosterone levels with girls' and boys' mental rotation scores at age 7 years, that is, prenatal testosterone was positively associated with boys' mental rotation scores, but negatively associated with girls' scores (Grimshaw, Sitarenios, & Finegan, 1995). Doreen Kimura (1999), a prominent researcher on the effects of sex hormones on cognitive abilities, concluded that peak spatial abilities are shown by people whose testosterone levels are in the low male range.
In one study, men with very low levels of testosterone showed impaired spatial abilities but normal verbal abilities (Hier & Crowley, 1982). In another study, when a group of men were given testosterone to improve their sexual functioning, they also showed improved performance on a block-design test as a side effect (Janowsky, Oviatt, & Orwoll, 1994). Female-to-male transsexuals who receive testosterone in preparation for sex reassignment surgery experience improved visual-spatial abilities, but they simultaneously suffer from decreased verbal fluency (van Goozen, Cohen-Kettenis, Gooren, Frijda, & Van de Poll, 1995). A Dutch female-to-male transsexual beginning testosterone therapy provided the following vivid account:
I have problems expressing myself, I stumble over my words. Your use of language becomes less broad, more direct and concise. Your use of words changes, you become more concrete...
The visual is so strong... when walking in the streets I absorb the things around me. I am an artist, but this is so strong. It gives a euphoric feeling. I do miss, however, the overall picture. Now I have to do one thing at a time; I used to be able to do different things simultaneously.
I can't make fine hand movement anymore; I let things fall out of my hands.
(van Goozen, 1994, p. 173)
You may have noticed that behaviors and traits that are linked to testosterone are often behaviors and traits that also show sex differences (see Chapter 1). For example, men and women show, on average, differences in physical aggressiveness, sexual activity levels, dominance, nurturance, visual-spatial ability, and a number or nonverbal behaviors such as smiling. Testosterone research typically studies within-sex relationships between testosterone and behavior, it is an inferential leap to suggest that sex differences in aggression, dominance, nurturance, and nonverbal behaviors are due to sex differences in testosterone levels. Still, it is worth noting that, on average, men have testosterone levels eight or more times those of women.
Although most research has focused on the correlates of men's testosterone levels, research has also demonstrated relationships for women. A study of 84 college women found that high-testosterone women reported being more enterprising, impulsive, and uninhibited, but less anxious, kind, mature, and warm than low-testosterone women (Baucom, Besch, & Callahan, 1985). And a study of 171 female inmates showed that, like men, women prisoners with high testosterone levels were more likely to have committed violent crimes and were more likely to he rated by prison staff to be aggressive and dominant (Dabbs & Hargrove, 1997; Dabbs, Ruback, Frady, Hopper, & Sgoutas, 1988). In general, findings for women are similar to those for men. Women who have high testosterone levels are characterized as aggressive, tough, competitive, dominant, and risk-taking. In both women and men, testosterone is linked to sensation-seeking and lack of inhibition (Daitzman & Zuckerman, 1980; Daitzman, Zuckerman, Sammelwitz, & Ganjam, 1978).
Most studies on the relation between testosterone and human behavior have measured hormone levels in adults, usually from blood or saliva samples. Biological theories, however, propose that prenatal exposure to sex hormones is critically important in organizing later sex-typed behaviors. Unfortunately, it is very difficult to measure prenatal hormones in humans. However, one recent study indirectly measured female fetuses' exposure to testosterone by measuring testosterone and other chemicals in their mothers' blood during pregnancy (Udry, 2000). The daughters' degree of feminine behavior was later measured in adulthood. Daughters who had been fetally exposed to higher levels of testosterone in the second trimester of pregnancy were more behaviorally masculine as adults. That is, they were more interested in male-typical, high status occupations; they were less interested in caring for children; and they were judged to be more masculine in their demeanor and nonverbal behaviors. They also scored higher on personality scales of instrumentality (i.e., dominance) and lower on scales of expressiveness (nurturance).
A British study of 337 preschool girls and 343 preschool boys round that early levels of testosterone, as assessed prenataily from mothers' blood samples, correlated with girls' but not boys' sex-typed behaviors; that is, higher testosterone levels were associated with more masculine play and activities in girls, as rated by their parents (Hines et al., 2002). Why were the findings present for girls but not boys? One possibility is that virtually all boys have a high enough dose of endogenous (self-produced) prenatal testosterone that variations in their mothers' levels do not make that much of a difference.
Physical Characteristics Related to Prenatal Testosterone
Research has tried to infer people's exposure to prenatal testosterone levels indirectly by measuring body characteristics thought to be related to prenatal testosterone. One such characteristic is the ratio of the lengths of the second and fourth digits of the hand (i.e., the index finger and the ring finger; Manning et al., 2000). Women tend to have shorter ring fingers relative to their index fingers, whereas men tend to have longer ring fingers relative to their index fingers. Index-to-ring-finger length ratios correlate with people's occupational choices, fertility levels, dominance, and sexual orientations (Lippa, 2003a; Manning, Scutt, Wilson, & Lewis-Jones, 1998; Williams et al., 2000). These findings suggest that prenatal testosterone levels are linked to adult gender-related behaviors.
Otoacoustic emissions provide another route to inferring prenatal testosterone exposure. Otoacoustic emissions are very faint sounds produced spontaneously by the inner ear or produced in response to faint clicks. Women tend to show more otoacoustic emissions than men, and this sex difference—like most physical sex differences that occur early in life—is thought to result from prenatal exposure to testosterone (McFadden, 1998). Otoacoustic emissions are related to adult sexual orientation (McFadden & Pasanen, 1998, 1999). Once again, the implication is that prenatal testosterone levels influence adult gender-related behaviors. Both finger lengths and otoacousic emissions seem to be fixed early in life—probably prenatally—and thus they are not likely to be influenced by socialization and cultural learning.
Other Hormones and Gender-Related Behaviors
Testosterone is only one of many sex hormones. Before leaving the topic of hormones and human behavior, let us briefly consider the effects of estrogen and other steroid hormones. Research on Turner syndrome suggests that a certain minimum level of estrogen may be necessary for normal cognitive development in women. But are normal variations in estrogens related to various kinds of human behavior? Studying the effects of estrogen levels is complicated by the fact that they change dramatically over the course of women's menstrual cycles. Nonetheless, recent evidence suggests that estrogen, like testosterone, is linked to human behavior.
Some studies have investigated whether variations in women's estrogen levels are related to their performance on cognitive tasks. Several studies show that when estrogens are high (in mid-menstrual cycle, around when ovulation occurs) women tend to show better speech articulation, verbal fluency, and manual dexterity. In contrast, when estrogens are low (around the time of menstruation), women tend to do better on visual-spatial tasks (Hampson, 1990a, 1990b; Hampson & Kimura, 1988; Phillips & Silverman, 1997). The relationship between estrogen levels and cognitive performance is not necessarily large, particularly considering that women's estrogen levels can vary by a factor of 20 over the course of their menstrual cycle. The largest association tends to be between women's cycling estrogen levels and their mental rotation abilities (approximate d = 0.65), with lower estrogen associated with higher ability (Hampson & Moffat, 2004). Nonetheless, studies of changes in cognitive abilities over the course of women's menstrual cycles make the interesting theoretical point that estrogen levels in adults can display activating effects on specific cognitive abilities.
Human data on the effects of estrogen on cognitive performance is bolstered by animal experiments (Hampson & Moffat, 2004). In many rodents, males learn to navigate the spatial layouts of mazes better than females do, despite the fact that females, on average, explore the mazes more actively. Female performance is negatively correlated with their naturally cycling estrogen levels; that is, higher estrogen is associated with worse performance and lower estrogen with better performance (Galea et al., 2000; Warren & Juraska, 1997). At low levels of estrogen, females are similar to males in their level of maze performance. Such research shows that measuring average sex differences in cognitive abilities maybe misleading because females are moving targets in terms of their cycling estrogen levels and related cognitive performance.
As noted before, some estrogen may be necessary for optimal mental performance. Postmenopausal women show improved performance on mental ability tests after taking estrogen supplements (Duff & Hampson, 2000; Resnik, Metter, & Zonderman, 1997), and estrogen therapy may reduce symptoms of Alzheimer's disease in elderly women (Jacobs et al, 1998; Tierney & Luine, 1998).