Few will dispute the enduring reality that women are underrepresented in many fields of science, technology, engineering, and mathematics (STEM). Gender discrimination in hiring at various academic levels is often put forward as a major reason, but recent studies have sparked controversy by suggesting that gender bias in male-dominated fields, in fact, favors women. Among the latest evidence going in that direction is a new study published in Science this week showing that, in France, the fewer female academics there are in a given field, the better chance women have of being hired as teachers in that subject. The hiring bias also seems to work in the other direction, though to a lesser extent, with aspiring male teachers being favored in fields traditionally dominated by women. However, the relevance of these results to academia and their generalizability to hiring processes in other countries are subject to some debate.
The hiring of teachers in France offers a unique, real-life setting to investigate the role of gender in skill evaluation. Every year, national subject-specific competitions are held at two levels—for teacher candidates to work at a middle or high school, or to gain the highest-level agrégation qualification to work in a high school, college, or university—to determine who will be offered a teaching position. Candidates are evaluated first by written exams, then oral ones. The identities of the candidates, and thus their genders, are only disclosed during the oral phase. Therefore, by measuring how candidates rise or fall in the rankings between the written and oral exams, the researchers could assess how knowing candidates’ genders affected their scores.
One of the main conclusions of the paper, co-authored by Thomas Breda and Mélina Hillion of the Paris School of Economics, is that in fields as diverse as mathematics, physics, economics, and literature, there is no evidence of discrimination against the underrepresented gender. “We rather find that the gender in minority is increasingly favored” during teacher hiring as the level of underrepresentation in academic faculty positions increases, the authors write in a joint email to Science Careers.
In male-dominated fields such as math, physics, and philosophy, the bias worked in favor of women: The hiring advantage gained by female candidates between the written and oral exams was equivalent to an average of 10% of female candidates overtaking all the men. Meanwhile, there was a subtler yet opposite bias at play in female-dominated fields such as literature and foreign languages, equivalent to 2 to 6% of the male candidates overtaking all the women between the written and oral phases. In most fields close to gender equity, including history and literature, no bias against either gender was found. The highest-level biology exam is however an exception, as female candidates faced a ranking loss of 4% compared with males, even though women represent around 46% of academic faculty in this field in France.
The findings are most directly relevant to aspiring middle and high school teachers, but they are also applicable to those pursuing academic research careers, the authors argue. In France, about a quarter of the candidates who pass the highest-level agrégation exam join colleges and universities, where they are hired primarily to teach but are also allowed to conduct research or prepare Ph.D. theses, the authors explain in their email. This is not the traditional entry route for French assistant professors, but a significant number of Ph.D.-holders take the exam to secure full-time teaching positions at universities while waiting to win assistant professorships. Having an agrégation can also improve aspiring academics’ career prospects in some fields, which makes the results more broadly relevant to academia, the authors argue in their paper. “The main message for female students is that they can enroll in the academic tracks traditionally dominated by men without strong fear of discrimination,” they write in their email.
Some however feel that the results aren't as clearly applicable to academia as the authors are making them out to be. “[T]he authors did a lot of serious number-crunching and came up with some intriguing findings,” says psychologist Virginia Valian of Hunter College in New York City. But the hiring processes and prestige of high school teachers and university faculty members in France are not the same, and a stronger distinction should have been made throughout the article, Valian adds. The fact that only a minority of agrégation-holders go on to an academic career is “a serious limitation to the study.”
Social scientist Kim Weeden, director of the Center for the Study of Inequality at Cornell University, sees another limitation: generalizability to other countries. “In the United States, for example, hiring into professorial positions is based on very local and non-standardized evaluations of candidates’ publications and scholarly potential, teaching record and teaching potential, ‘fit’ with the department, and so forth. Patterns of test scores on standardized exams in France just can’t tell us much about hiring in these systems,” she writes in an email to Science Careers. Even so, Weeden praises the study’s “very rigorous, creative analysis of a unique and high-quality data set.”
The findings are nonetheless in line with recent controversial studies based in the United States. Among them was a 2015 study that found, based on faculty member evaluations of fictitious job candidates, that women are favored 2 to 1 for tenure-track positions in engineering, biology, and psychology. One of the authors of that paper, developmental psychologist Stephen Ceci of Cornell University, thinks the new findings are encouraging for women pursuing STEM. “Young people, particularly young women, should be heartened by these findings because it goes against the common narrative that says the deck is stacked against them from being hired as entry-level professors,” he writes in an email to Science Careers. (Ceci gave Breda and Hillion feedback on their manuscript before submission but otherwise does not collaborate with them.) “If anything, the data reveal that the fields in which they are most underrepresented are the very ones that desire them most. If bias is at work, it is bias in their favor, not against them.”
Valian sees it somewhat differently. “Do their findings suggest that women are going to benefit by being badly represented in a field? Hardly. The authors have to wonder why women continue not to be well represented in math, physics, and philosophy, in universities and research institutes. If the exams are picking those women out at higher levels, why don't we see them succeeding more?”
The new data don’t provide much insight into the underlying factors behind the skewed hiring. It could be, for example, that rather than reacting to gender underrepresentation across different fields in academia—which, the authors argue, reflects societal stereotypes—examiners were simply trying to redress gender imbalances they saw in the applicant pool, or even in past exam winners. This is unlikely to be the case, however, the authors argue in their paper, because many of the exams were close to parity both among candidates and past winners. The most plausible explanation, they continue, is that examiners were trying to counteract gender stereotypes.
Weeden however thinks that the authors “are a bit too hasty” in concluding an evaluator bias, as “the various checks that they offer against other interpretations aren’t as convincing as the general pattern of results,” she says. In particular, the oral exams seem to be designed to test pedagogical skills as well as subject knowledge, she notes, so one alternative explanation could be that “the very select group of women who pass the written tests in math are better than the very select group of men who pass those tests at explaining complex math concepts, [a scenario in which] they [would] also rank higher on the oral exams in math.”
To help address the continued underrepresentation of women in many fields in academia, in their paper Breda and Hillion call for implementing policies to counteract stereotypes and discrimination at early stages before educational choices are made, and spreading the message to female students that they have equal or even better chances to succeed as teachers and academics in fields where women are the minority—which is an approach that Ceci agrees is the right one. “By forcing the argument back to earlier in development, Breda and Hillion's results demonstrating that women are not discriminated against in exam scoring and, in fact[,] are advantaged … suggest that hiring-point policies—such as mandating gender sensitivity training for members of hiring committees, or a certain portion of members must be women—are missing the point,” he writes.
But Weeden is only partially convinced. She agrees with the recommendation to counteract stereotypes at early ages, which is consistent with a substantial body of sociology research, including some of her own work, “show[ing] that by early high school, significantly more young men than young women plan to enter engineering, math, and other male-dominated STEM occupations,” she says. But “even among students who show an early commitment to science-related careers, who score well on standardized math exams, and who take high-level math and science courses, young women are much more likely than men to drop out—or be pushed out—of educational pathways that would lead to high-level science careers,” she continues. “Understanding the educational decisions that kids make early in their lives is important, but it’s certainly not the only life stage at which social processes can lead to women’s underrepresentation in science-related occupations.”
Curt Rice, president of Oslo and Akershus University College and head of Norway’s Committee on Gender Balance and Diversity in Research, takes a harder line. On the one hand, he welcomes this large-scale study from France, noting that it is “important to counter the US-dominance of subject pools in this kind of research,” he writes in an email to Science Careers. Also, “[s]tudies which look at significant quantities of ‘real data’ instead of simulations are important and interesting, even though they necessarily will have more variables that have not been controlled for.” But the authors’ recommendation “to bright young women … to study fields in which women are underrepresented … [is] vapid,” he continues. “The test results they discuss tell us nothing about the experience women have in the classroom or workplace when they are severely underrepresented. Of course, I want bright young women to choose STEM subjects, too, but the nuances of the French exam grading system seem to get overplayed in this advice.”