Was the same true of students who were gifted in both math and verbal skills? Members of SMPY Cohort 2 completed two visuospatial subtests of the Differential Aptitude Tests.19 The results mirrored those from Project Talent. The students’ visuospatial skills interacted with their verbal and math scores, with high skills pushing them toward STEM and low scores pushing them toward the humanities or social science majors.20
What I’ve just reported applied to Cohort 2, who were in the top 0.5 percent in intellectual ability. Surely (or so it would intuitively seem), this phenomenon has to fade at some level of ability when it doesn’t make any difference which skill set is higher—even your “weaker” skill is fabulously high. But apparently not, or so SMPY’s Cohort 3 seems to tell us.
SMPY’s Cohort 3 is the largest database of profoundly gifted persons ever assembled for systematic longitudinal research: 253 males and 67 females who as 12-year-olds were in the top 0.01 percent of intellectual ability, with an estimated mean IQ of 186.21 The top 1 in 10,000. The phenomenon they had identified in Cohort 2—even highly gifted people gravitate to the fields where they have the greatest comparative advantage—also applied to Cohort 3. Among those whose SAT math scores were at least a standard deviation above their SAT verbal scores (the “high-math” group”), 69 percent had an undergraduate major in math or an inorganic science, compared to 29 percent of those whose SAT verbal scores were at least a standard deviation above their SAT math scores (the “high-verbal” group). Forty-two percent of the high-verbal group had undergraduate degrees in the humanities or arts, compared to 23 percent of the high-math group.22 The sample size of women is too small to draw strong conclusions from Cohort 3 as a single study, but its consistency with the results for the larger sample in Cohort 2 and the nationally representative Project Talent data make it worth noting: The male-to-female disparity in attraction to STEM even among the profoundly gifted suggests that we may be looking at nothing more complicated than people’s attraction to doing what they do best no matter how extraordinarily gifted they are in the things they don’t do best.
Even Among the Exceptionally Gifted Who Are Attracted to STEM, Males and Females Gravitate to Different Types of STEM
Hidden within the overall figures is another intriguing story: Even gifted women who are attracted to STEM gravitate toward the life sciences (People-oriented), not math and the physical sciences (Things-oriented). It was not a subtle tendency. Proportionally, males outnumbered females by almost two to one on the Things-oriented sciences, and females outnumbered males by almost two to one on the People-oriented sciences.[23] The implication: Women who are so gifted that they can deal with any intellectually demanding field are not scared off by science per se. They instead tend to prefer those fields that deal with living things rather than nonliving things.
The results from the smaller sample of profoundly gifted woman in Cohort 3 indicate a similar sex difference in interests, except that the female tilt toward STEM disciplines that deal with living things was even stronger among the women of the top 0.01 percent in math than it was for those who were “merely” in the top 1 percent—and it’s definitely not because the women of Cohort 3 weren’t smart enough to do physics or pure mathematics if they felt like it.24
Even Among the Exceptionally Gifted, Women Have Different Life Priorities and Work Priorities Than Men That Affect Their Career Trajectories and Achievements
The SMPY women were similar to the general population of college-educated women in matters of marriage and children. Among the SMPY women at ages 45–49, 75 percent were married, compared to 72 percent of all college-educated women of the same age. Seventy-four percent of the SMPY women had borne children, compared to 80 percent of their college-educated peers. Fifty-seven percent of the SMPY women had more than one child, compared to 61 percent of their college-educated peers.25
In light of this, it is not surprising that the list of sex differences in work preferences and life values that began this chapter included several that directly or indirectly involved children, particularly with regard to women’s markedly greater unwillingness than men to work long hours. Long hours at work compete with that responsibility. But motherhood doesn’t appear to be the whole story. Remember that the women of Cohort 2 were ages 45–49 when they were surveyed in 2012–13. They weren’t being asked about how many hours they were willing to work outside the home when they had preschool children at home, but how much they were prepared to work in their late 40s and the years ahead. In that context, the results of one survey question were especially intriguing. How much would they be willing to work, at most, if given their ideal job? Thirty percent of these extraordinarily talented women, few of whom had small children to care for, were unwilling to work more than 40 hours per week even if they were given their ideal job, compared to only 7 percent of the men.26 Add in the other responses indicating the priorities SMPY women put on community service, having time to socialize, not working outside the home, having a meaningful spiritual life, and being available for family and friends, and we are once again faced with a sex difference in profiles. Individually, most of the effect sizes ranged from small to medium. As in the case of personality differences, these effect sizes are found in indicators that are conceptually related but functionally distinct, and I would argue that aggregating them would give us a more accurate picture of the magnitude of the sex differences involved. But, even lacking the data to calculate Mahalanobis D, it is safe to conclude that in middle age the SMPY men and women had importantly different life priorities—which makes the next set of results all the more interesting.
The SMPY Men and Women Saw Themselves as Having Equally Satisfying Lives
As part of their survey, the members of Cohort 2 were given some standard instruments for measuring subjective well-being.27 As a group, their evaluations of their lives as of their late 40s were extremely positive—in the top decile, according to the norms for those instruments. The differences between women and men were uniformly minuscule.28 The authors’ appraisal of these findings seems exactly right:
In short, marked sex differences in how participants allocated their time and structured their lives were not accompanied by corresponding sex differences in how they viewed their career accomplishments and close relationships, or in their positive outlook on life. One interpretation of the lack of appreciable differences between the sexes across these indicators is that there are multiple ways to construct a meaningful, productive, and satisfying life.29
I have given so much space to the SMPY cohorts for two reasons. The first is to emphasize that sex differences in ability profiles are not the whole explanation for differences in educational and vocational choices. As I discussed in chapter 3, sex differences exist in the male and female profiles of abilities, and they do indeed have implications for performance in certain occupations. But the data from the SMPY cohorts convincingly document that STEM sex differences persist at virtually the same levels when women are easily capable of having successful STEM careers if they want them.
Second, the SMPY results pose a challenge to which defenders of social-construct explanations must respond: Suppose we grant that socialization discourages girls from STEM fields. But socialization to avoid STEM fields is not something in the water that all little girls drink in similar quantities. It is fostered by specific inputs from parents, teachers, peers, and the media, among other agents. Different little girls get different amounts.
As noted earlier, we have good reason to think that the SMPY women disproportionately grew up with gender-neutral toys, had mothers who were in professional careers, had parents both of whom proactively told their daughters to transcend gender stereotypes, were educated in progressive upper-middle-class schools, and had peer groups consisting of other girls raised in similar circumstances.
Vulnerability to socialization into traditional female roles also varies by a girl’s personal characteristics. We know for sure that the SMPY girls were all extremely smart and knew
they were smart. It seems reasonable to assume that they were also disproportionately confident, with keen critical faculties, and resistant to propaganda.
In short, the nature of the SMPY sample tells us that the SMPY females got a smaller dose of socialization to traditional female roles than average and had higher levels of resistance than average. If socialization is the whole explanation for differences in attraction to STEM vocations, how is it possible that the girls of Cohort 2, in the top half percent of academic ability, show the same ratio to the SMPY men that the general population of college-educated girls shows to the general population of college-educated men? The SMPY results imply sex differences that transcend socialization.
GOING BEHIND THE NUMBERS
For a three-dimensional understanding of the priorities of highly accomplished women, you should read Susan Pinker’s The Sexual Paradox: Men, Women, and the Real Gender Gap (2008). In addition to a still-relevant account of the science of sex differences, the book contains a number of in-depth profiles of women who were extremely successful professionally but made choices in their careers that were different from those men usually make. Pinker’s narrative is in effect the SMPY results brought to life.
Three final thoughts before leaving the SMPY women:
First, it’s time for another reminder that all the results I have described amount to statistical tendencies. The SMPY cohorts included women who were as professionally driven as stereotypical men and men who were as involved in family and community as stereotypical women. But it’s still true that the sex differences in profiles were unmistakable.
Second, where is it written that STEM majors or careers are objectively more valuable than non-STEM majors or careers? Lubinski and Benbow make an important point that should be obvious but isn’t: “Given the ever-increasing importance of quantitative and scientific reasoning skills in modern cultures, when mathematically gifted individuals choose to pursue careers outside engineering and the physical sciences, it should be seen as a contribution to society, not a loss of talent.”30
Third, where is it written that spending years of seven-day, 80-hour workweeks is more fulfilling or more fun than combining a less intense career with richer family life? These gifted women, applying the mature judgment of their late 40s, took great satisfaction from their time and engagement with their children, spouses, and communities. If you try to argue that these women were duped into accepting traditional female roles, you run into a problem: Chances are that the women who made those judgments are a lot smarter than you are.
Extrapolating from the SMPY Experience: The Breadth-Based Model
Psychologists Stephen Ceci, Wendy Williams, Jeffrey Valla, and their colleagues have pulled these strands together in a theoretical explanation of sex differences in attraction to STEM. It is known as the breadth-based model. The argument runs like this: It has been demonstrated that the decision to pursue a STEM career depends on two things: high math ability (hardly anyone without high math ability is attracted to STEM) and lower verbal ability.31 The public and academic debate over female underrepresentation in STEM has focused on the first of those two, but the evidence indicates that the second may be even more important. People with equally high or higher verbal skills (male or female) have an array of attractive alternatives to STEM—the arts, social sciences, law, and business, for example—and many with high math skills and high verbal skills choose those alternatives.
Now recall the asymmetry noted in the SMPY sample: Gifted boys tend to be gifted in those skills that go into superior performance in STEM occupations but not in skills that go into superior performance in non-STEM occupations, while gifted girls tend more often than boys to have a choice—they are capable of superior performance in just about any cognitively demanding field.
Next, consider the empirical observation that, on average, females are drawn more to the occupations facilitated by high verbal skills. The breadth-based model argues that this is no accident, nor does it draw from a narrow calculation that financial success is more likely by going with one’s best skills. Rather, from an early age, males and females tend to have different interests. Those interests lead to different experiences and acquisition of knowledge, and those in turn lead to different choices of careers. As Valla and Ceci put it in their 2014 article summarizing the breadth-based model, “the ‘nature’ of cognitive sex differences lies not in absolute ability, but in breadth of intrinsic interests—and its downstream developmental effects on interests, abilities, and career choices.”32
Valla and Ceci are referring to the choice between STEM careers and their alternatives. But the sex differences in personality traits discussed in chapter 2 suggest that we need not limit the argument to the small proportion of the population that has high math skills. It’s time to leave the special case of the gifted and look at sex differences in interests that apply to men and women in general.
The Revolution in Women’s Education and Work Since 1960
Any discussion of the trends in vocation and life choices among the general population of women must begin with the transformed opportunities for women since 1960. Sociologist Catherine Hakim has cast these in terms of five specific changes that produced a qualitatively new reality for women:
The invention of new forms of contraception, especially the pill, gave women reliable and independent control over their own fertility.
The equal opportunities revolution gave women better access to all careers and positions.
White-collar occupations, the ones most attractive to women, expanded.
Jobs for secondary earners expanded, making it easier for women to combine childcare and work outside the home.
Freedom of lifestyle choices in liberal modern societies increased.33
All of these changes began in a concentrated period of time during the 1960s. Without hyperbole, the results have been revolutionary.
In 1960, a few years before second-wave feminism took off in the United States, only 41 percent of women ages 25–54 were in the labor force. In 2018, that figure stood at 75 percent.
By 2015, women had a presence in high-status jobs that was inconceivable in 1960. From 1960 to 2018, women went from 1 percent of civil engineers to 17 percent; from 5 percent of attorneys to 35 percent; from 8 percent of physicians to 42 percent.
Not a single woman was the CEO of a Fortune 500 company in 1960, nor would there be any until 1972.34 In 2018, 25 women were Fortune 500 CEOs, among them the chief executives of General Motors, IBM, PepsiCo, Lockheed Martin, Oracle, and General Dynamics.
In 1960, there was one woman in the U.S. Senate. After the 2018 election, there were 25. In the 1960 House of Representatives, there were 19 women. After the 2018 election, there were 102.
Female students from elementary school through college have long had higher mean grade point averages than males in most subjects (including math).35 But in 1960 women were nonetheless a minority of entering college students (46 percent), and the gap grew during the undergraduate years. Almost two males got a bachelor’s degree for every woman who did. In 1982, the number of women getting bachelor’s degrees surpassed the number of men. The gap continued to widen subsequently. By 2016, 1,082,669 women got bachelor’s degrees compared to 812,669 men—a 33 percent difference.[36]
The change in professional degrees was even more dramatic. The figure below shows the number of PhDs, medical degrees, law degrees, dental degrees, and others ordinarily requiring at least three years of graduate work.
Source: National Center for Education Statistics, Digest of Education Statistics, 1995 edition: Table 236, and 2017 edition: Table 318.
In 1960, 20 men got a professional degree for every woman who did. By 1970, the ratio was less than 10 to 1. By 1980, it was less than 3 to 1. In 2005, women caught up with men. Since then, more women have gotten more professional degrees than men in every year. As of 2016, 93,778 women got a professional degree compared to 84,089 men.[37]
Nothing about women’s abilities had ch
anged in the interim. Opportunities formerly denied them had opened up, and women were exercising talents and capacities they had possessed all along. That is the overriding headline about gender and occupational changes during second-wave feminism.
It is in this context of progress for women over the course of half a century that I now show you the data on current sex differences in vocational interests and life.
Sex Differences in Vocational Interests and Life Choices in the General Population
The argument here amounts to two mini-propositions: that the vocational preferences of men and women as measured by tests differ in ways that correspond to differences on the People-Things dimension, and that the job choices that people actually make correspond to their preferences.
Men and Women Are Attracted to Different Vocations, and Those Differences Correspond to Differences on the People-Things Dimension
In 1959, psychologist John Holland published an article titled “A Theory of Vocational Choice” that has shaped vocational counseling ever since.38 At the heart of his theory, which was created without regard to sex differences, was the premise that an adult comes to a vocational choice with “a hierarchy of habitual or preferred methods for dealing with environmental tasks.… The person making a vocational choice in a sense ‘searches’ for situations which satisfy his hierarchy of adjustive orientations.”39 In its final form, Holland’s theory posited six clusters of these orientations:
Human Diversity Page 10