SEX AND SMARTS: UNDER-REPRESENTATION OF WOMEN AMONG SCIENCE AND ENGINEERING FACULTY
I am in the process of becoming a scientist.
A female scientist.
I didn’t think that my gender was relevant to my career aspirations. I have plenty of female company (and competition) as an undergraduate science student at UBC. The ratio of women to men in my classes must be close enough to 50/50 that I have never noticed a gender imbalance (nevermind an advantage in heterosexual dating odds for women in my discipline.) Ditto for the distribution of intimidatingly-smart students; the prof-stumping questions come from women just as often as from men. I have worked in research labs that had more female than male grad students. Moreover, I have been taught by female lecturers and never considered the female presence behind the lectern to be noteworthy.
Maybe I should have paid closer attention to the ranks of my female lecturers, to see how many of them were sessional instructors as opposed to tenure track or full professors. While the proportion of doctoral degrees awarded to women over the past three decades has increased steadily, the trend has not been matched by a commensurate increase in the proportion of women winning tenured faculty positions in the sciences. The pool of qualified candidates contains more women than ever before, but they are not advancing into the highest ranks of academia in a proportional way.
In January of 2005, the then-president of Harvard University, economist Larry Summers, made a speech at a conference about diversifying the science and engineering workforce. He outlined his ideas regarding the causes of female under-representation in tenured faculty positions at top-tier research universities, with the intention of provoking discussion. He succeeded on that front; his comments ignited a firestorm of controversy in academia that was intensively reported in the popular press, and ultimately contributed to Summers’ resignation from his post as president.
Summers outlined three hypotheses, also found in the scientific literature, to explain why there are so few tenured female faculty in science and engineering. He ranked the hypotheses according to his evaluation of their relevance, with the first two accorded much greater importance. These hypotheses were, in order:
I. High-powered job hypothesis: Traditionally, only males have been willing and able to choose high-status jobs that require absolute dedication, as evidenced by eighty-hour workweeks. Women continue to choose alternate careers with less status but also lesser time demands.
II. Variability hypothesis: There are more highly intelligent males than females, which accounts for the greater proportion of males in the highest ranks of academia.
III. Discrimination hypothesis: There is discrimination against women by predominantly-male hiring committees who prefer to pick candidates like themselves.
Summers faced criticism for all three hypotheses, but the majority of it focused on his assertion that innate differences in intellectual ability between men and women exist, and contribute more to the under-representation of women than does discrimination. His comments regarding innate differences were certainly the most disturbing to me, because there would be no way for me to remedy an innate deficiency.
Summers referred to his most contentious hypothesis as “ different availability of aptitude at the high end,” advanced as the variability hypothesis by other proponents. It is based on the observation that many studies of particular aspects of human cognition, presumably relevant to scientific ability, find greater variation in the male population than in the female population. In these cases, the male bell curve is a bit squashed, with a shorter peak and more men at the extreme tails than in the corresponding female curve. Summers reasoned that because many studies find more variation among males in some cognitive traits that there must be more variation in scientific ability as well. He also reiterated the belief that the “intellectual elite” who become tenured professors in the sciences must come from the extreme tails of the distribution of scientific ability. If there really are more men than women at the extremes of scientific ability, this could explain why there are disproportionately more men in high-ranking positions in academia.
Summers’ conclusions were based on common myths and incorrect assumptions, but he is not the only academic to espouse those beliefs. At least partly in response to the furor Summers stirred up, in 2006 the National Academies Committee on Science, Engineering and Public Policy released its evaluation of the underlying causes of female under-representation among tenured science faculty. Their evaluation was based on meta-analysis of the literature, and the report included a thorough debunking of the misapprehensions that led Summers astray.
In reference to the variability hypothesis in particular, the National Academies report identified research demonstrating two important points:
1) Differences in variability have diminished over time, in studies of traits considered important to scientific ability, such as mathematical reasoning. As an example, the SAT-Math exam has been administered to precocious middle school students for decades, in order to identify students who would benefit from enriched academic programs. In 1983, the ratio of boys to girls who scored at the extremely high end was 12 to 1. By 2005, the ratio had dropped to 3 to 1. It appears the male population is not much more variable. More importantly, for such changes to occur over the short span of a decade or two, the source of the differences must be social rather than biological.
2) The scientific elite do not necessarily represent the extreme tails of the distribution of ability. To use SAT-Math scores as an example again, analysis shows that the SAT-Math scores of students who later obtained at least a bachelors degree in science or engineering are not terribly exceptional; they come from the top 40% of the test score distribution, not the top few percent. More than one third of future science and engineering grads had SAT-Math scores below that of the average humanities degree-holder. Admittedly, the SAT-Math scores of people who obtain graduate degrees might be more exceptional. However, at every level of SAT-Math achievement, women are half as likely to pursue a science career as men, which suggests that inclination matters more than ability in career choice.
In contrast to Summers’ assessment, the National Academies of Science report identified discrimination and institutional disadvantages to women as the major causes of the under-representation of women among tenured science faculty. The report outlines study after study that show unconscious bias at work. It appears that women candidates are often held to a higher standard than their male competitors. A study of postdoctoral fellowship applications in Sweden in 1997 compared men and women who were judged equally competent, and found that the women required significantly more publications to obtain the same designation of competence.
As I said earlier, I didn’t think that my gender had any relevance to my career aspirations, but I had no idea that women in science were faced with these additional challenges. While I’m relieved that research shows that the obstacles to female success in the highest echelons of academic science aren’t innate, I am also terribly intimidated by the apparent ubiquity of unintentional discrimination and the magnitude of its effects. The path to full professorship is difficult enough, without having to publish twice as many papers in order to compete with male peers.
Fortunately, discrimination can be identified and reduced, and institutions can be restructured. At UBC, the Equity Office tracks the advancement of women in all faculties, in service of the explicit institutional goal of matching the proportion of female faculty hired to their proportion in the pool of qualified candidates. The proportion is currently set at 35%, to match the percentage of doctoral degrees awarded to women across all disciplines at Canadian universities. That goal has been met several times since 1987, and has undoubtedly improved hiring practices in sciences and engineering at UBC.
Similar changes have been underway at institutions around the world for many years. The situation for women academics in the sciences will surely improve, helped along by the spectacular foot-in-mouth performances of academic leaders to spur debate and vigilance.
References
Annual Report 2003, UBC Equity Office
Beyond Bias and Barriers: Fulfilling the potential of women in academic science and engineering. National Academies Committee on Science, Engineering, and Public Policy. National Academies Press, Washington. (2006)
Brody and Mills, Talent search research: what have we learned? High Ability Studies 16(1): 97-111 (2005)
Summers, Lawrence H. Remarks at National Bureau of Economic Research Conference on Diversifying the Science and Engineering Workforce. Cambridge, Mass. January 14, 2005. Office of the President.
Weinberger, C. Is the science and engineering workforce drawn from the far upper tail of the math ability distribution? Cited in: Beyond Bias and Barriers, 2006.
Wenneras, C. and A. Wold. Nepotism and sexism in peer-review. Nature Commentary 387: 341-343 (1997)