The soon-to-be-completed leadership succession at Xerox from Ann Mulcahy—a woman—to Ursula Burns—also a woman—is one for the record books. It will be the first woman-to-woman transition at a Fortune 500 company, and Burns is the first African-American woman to take the helm of any such corporation.
Mulcahy has balked at the notion that the über-competent Burns needed her help—or anyone else's—in making her way through the ranks. Nonetheless, the paucity of women in senior positions who might in turn mentor young women on their way up the ladder is one of the primary reasons put forth to explain the continued existence of the glass ceiling in corporate America. In science and technology, the situation is even bleaker—women are under-represented at every level, from advanced college classes to the executive suite, making Burns' rise from math major to CEO of tech giant Xerox all the more remarkable.
Of course, female mentorship is only one strand of a complex web of explanations—from aptitude to temperament to societal discrimination—for the gender gap in science and elsewhere. A recent National Bureau of Economic Research Working Paper on gender and academic achievement at the U.S. Air Force Academy, however, finds that the importance of female mentors may be even more powerful than previously thought. The study, by University of California-Davis economists Scott Carrell and Marianne Page and their colleague James West at the Air Force Academy, finds that replacing a male instructor with a female one has such a strong effect on female achievement as to erase the gender gap entirely.
The trio of economists examined the undergraduate careers of 9,481 cadets taught by nearly 250 different science and math instructors at the USAFA during the years 2000-08. Untangling the impact of an instructor's gender from the many other factors that influence student performance has hamstrung most research on the topic: If women taught by women perform better (or worse) is it because the instructors attract better (or worse) female students rather than teach them better? Or is it because women "dumb down" their syllabi or exams to make it easier for lower-performing women to do well? Or is it something else entirely?
For the students at the Air Force Academy, the researchers are blessed with the rigid curriculum and protocol of a military college, ruling out most complications. Cadets face a heavy slate of compulsory first-year courses, including a battery of science, technology, engineering, and math requirements, and within each science-related course, students are randomly assigned to one of several teaching faculty members. So whether a student—male or female—was taught by a woman was a matter of luck rather than any active choice by student or professor. Each class used an identical syllabus and all students took the same exam, which prevented the various instructors from adjusting their courses to cater to better or worse students. Further, since the researchers also had access to students' math SAT scores, they could take account of any differences in quantitative abilities among the students.
The authors found that women on average obtain scores that are 0.15 grade points lower (half the difference between an A and an A-) than their male classmates, even after accounting for students' SAT scores. The gap in performance was widest for women taught by men. When a female instructor was put at the front of the classroom, nearly two-thirds of the grade point gender gap evaporated. (It was also the case that men performed better when taught by other men, but the difference was far less substantial.) The authors persuasively demonstrate that the overall male-female performance difference is due in large part to the fact that men dominate the Air Force Academy science faculty (as is the case in most schools), with only 23 percent of courses taught by women.
The researchers also found that the influence of professor gender was even starker for the set of students who were math whizzes when they entered the Academy (those with math SAT scores above 700). For these students, a female instructor eliminated the gender GPA gap entirely—and solely because high-performing women did well in their classes rather than because high-ability men underperformed.
What's more, having a male instructor didn't just affect female cadets' performance in their first-year classes—ramifications could be seen throughout their undergraduate careers. Not surprisingly, students who did well in their introductory science classes were more likely to go on to obtain science degrees (and presumably go on to science-related professions). Among high-math-SAT students—those most likely to be the ones to go on to obtain science degrees—the authors calculate that having a women-only roster of faculty would create gender parity among science majors.
What is it about a woman instructor that is so important for female pupils? It's unlikely to be simply the sense of empowerment of seeing that women can in fact make it in science. If that were the case, then having all female professors should help their female students catch up to the men and having all male professors should cause the male-female performance gap to widen. Yet the authors found that, while female students perform better on average in classes taught by female professors, there are some male professors under whom there's no achievement gap between male and female students (and also some female professors for whom the gender gap is as big as that of some of their male colleagues). So some men are very good at mentoring women, just not nearly enough of them.
What kind of man makes a good mentor? Is it because, as is sometimes suggested, men with daughters make good mentors, having developed greater empathy for the challenges faced by their female students? Or differences in teaching style? The authors unfortunately don't know much about the Academy's teaching staff, so for now the enormous impact of professor gender remains a bit of a black box.
Regardless of the underlying mechanism at work, the study has wide-ranging implications for what might be done to keep talented women on science career tracks. Most obviously, the findings provide further justification for affirmative action programs to promote women in the sciences, to break the cycle of talented women opting out of science because there are no women in science. At the same time, we might unravel the mystery of what makes people—men or women—better at mentoring their female protégés.
I posed the question of how to create gender equity in science to Stephanie Pfirman, a Barnard College environmental scientist and a member of a Columbia University initiative on women in the sciences. She pointed out that recruiting female mentors and making men into more women-friendly bosses and teachers are both efforts aimed at changing the environment faced by young women. While these are worthy objectives, she suggested developing coping mechanisms to deal with circumstances as they are—for example, realizing that getting an A- or even a B+ in an introductory course doesn't spell the end of your career as a scientist, as many high-achieving young women believe. Yet the results of this study suggest that just by helping more women to overcome the adversities they face in becoming scientists today, we will make science less of a man's world for the female scientists of tomorrow.