When a CNN article titled “The Myth about Women in Science” came crawling across my feed, I have to admit that I wasn’t optimistic. I wondered what could possibly count as “THE Myth about Women in Science”. Maybe that women and girls have lesser skills in mathematics and spatial reasoning? That is truly a myth about women in science but I couldn’t see why it would be news as it’s been widely disputed.
A quick skim of the article resulted in a briskly raised eye brow. The myth apparently is this: women are less likely to be hired than equally qualified men when they apply for tenure track position. The authors (Wendy Williams and Stephen Ceci, both of Cornell University) claim that this misunderstanding is the major cause of women’s underrepresentation in scientific careers.
“The prevailing wisdom is that sexist hiring in academic science roadblocks women’s careers before they even start.” It is?
“Many female graduate students worry that hiring bias is inevitable.” They do?
Certainly, hiring bias is an issue discussed in relation to women’s careers in STEM areas, but it is nowhere near the top of the list of barriers and obstacles that come up frequently. Bullying, aggression, withholding of resources, stereotype threat and imposter syndrome: sure. Hiring bias, “meh, not so much”, to put it colloquially. To put this in context, Anthropologist Jessica Brinkworth and I recently solicited personal essays of women’s experiences in science. We asked all of the potential authors to specifically describe obstacles they had faced and explain how they tried to address them to move forward or move on. This included women who persevered on the tenure track and those who chose other scientific lives. We received over a hundred submissions and not a single one was about hiring bias. There were essays about mental health, about struggling to raise a family, about sexual harassment and assault, about bullying and even having a job offer rescinded due to pregnancy. But not one women mentioned even being worried about hiring bias.[i] I’m not saying it doesn’t exist or isn’t an issue, but to say that it’s the “prevailing wisdom” on why women are underrepresented in science is a big stretch.
The study they are describing was published this week in Proceedings of National Academy of Sciences. In it, Williams and Ceci contacted several hundred science professors (faculty members in biology, economics, psychology, engineering) and asked them to rank three made-up candidates for a hypothetical job in their department. The professors were asked to make that decision based on narrative descriptions of the candidates’ research accomplishments and the impressions they made on the hiring committee. The candidates always included: two top candidates (one male, one female) and a third slightly weaker candidate (always male). The two top candidates were paired in 20 difference combinations to experimentally test whether factors such as their marital status would make a difference. Based on professors’ rankings, Ceci and Ryan suggest that there is actually a preference for women candidates and that it can be as strong as 2:1.
As you can probably see, this design presents some real challenges. Ranking candidates on a narrative description does not replicate how actual hiring happens (although the authors do try to address this with a small follow-up study based on applicant CVs). And hypothetical hiring is not real hiring. It’s easy to say that you would make what a potentially socially acceptable choice, but when asked to really devote resources to a candidate and make them your department colleague, many other factors (including biases) would come into play. (Physicist Dr. Skyskull addresses this with some wit in what he calls “A one-act play about a study in hiring practices in STEM”). These, and several other methodological issues, have been discussed by others such as sociologist Zuleyka Zevallos, marine scientist Claire Griffin, and philosopher Helen De Cruz. **update: Psychologist David Miller has written a post responding to several of these criticisms of the methods, specifically those highlighted by Zevallos. This response was also posted as a comment on Zevallos post, do see her response there.
But from my corner of the world as a science educator, the way the study is justified and how its implications are explained are what interest me the most. The way that issues about women in science are discussed in public spaces, such as in schools, around the dinner table and in media outlets is one very important factor in women and girls choosing and then persisting in scientific careers. We are all influenced by what our parents, friends, teachers, role models and colleagues say about what it means to be women in science. Presenting findings in a prominent mass audience media space like CNN and saying that “The myth” of women in science is untrue is bound to influence those conversations.
So, the part I want to dig into is the support they provide for the major claims they make about why this study is important. The reasoning goes like this: this experiment found that women are ranked higher in hypothetical hiring, therefore any differences in women’s representation in research positions are due to women choosing not to apply for those jobs in the first place. In particular they say if women were less afraid of hiring bias, they would apply for tenure track jobs in greater numbers and this part of the problem would be solved. From the CNN article, they conclude “The low numbers of women in math-based fields of science do not result from sexist hiring, but rather from women’s lower rates of choosing to enter math-based fields in the first place, due to sex differences in preferred careers and perhaps to lack of female role models and mentors.” This reasoning is not confined to the media summary though and is presented in the published study as well. They are actually even more blunt there in identifying hiring bias as the major issue: “The underrepresentation of women in academic science is typically attributed, both in scientific literature and in the media, to sexist hiring” (p. 1). And they conclude the paper with this: “We hope the discovery of an overall 2:1 preference for hiring women over otherwise identical men will help counter self-handicapping and opting-out by talented women at the point of entry to the STEM professoriate, and suggest that female” (p. 6).
My problem is the evidence they provide for these claims, which is weak at best.
In the article, they cite four studies to support their claim that fears about hiring bias are prevalent and a major reason for the underrepresentation of women.
The first is a study by Sheltzer & Smith (2014) published in the same journal and titled “Elite male faculty in the life sciences employ fewer women”. This study uses publicly available data about junior scientists (graduate students and postdocs) employed in elite labs scientific labs. They find a gender gap in the labs but are very clear that bias is only one possible explanation: “Thus, one cause of the leaky pipeline in biomedical research may be the exclusion of women, or their self-selected absence, from certain high-achieving laboratories.” In fact, they cite a wide variety of factors that impact hiring gaps including societal expectations related to which partner should move to benefit the other’s career, self-chosen and socially implied ideals of work-family balance, and biases that preferentially provide access to scientific resources (mentoring, supplies, public visibility). They say explicitly “Notably, our current data do not show conscious bias on the part of male PIs who employ few female graduate students and postdocs. It may be the case that women apply less frequently to laboratories with elite male PIs” (p. 10110) . So, that study is not making the claim that Williams and Ceci say it is, they are merely noting that for many different reasons, women a less represented in elite labs something that likely impacts their future career prospects.
Ok, next up is a report from the American Association of University Women (Hill, Corbett & Rose, 2010) titled Why so Few? Women in Science, Technology, Engineering, and Mathematics.
This report is not about hiring bias. With chapters on topics such as “Spatial skills”, “Self-Assessment” and “The College Student Experience”, it’s a summary report on the immense variety of factors that impact girls and women at all points in the science career journey. When it does come to talk about hiring, it actually quotes a study with the same findings that Ceci and Williams say should come as a shock: “Although recent research found that when women do apply for STEM faculty positions at major research universities they are more likely than men to be hired, smaller percentages of qualified women apply for these positions in the first place (National Research Council, 2009).” Williams and Ceci cite this same NRC report in their conclusion as something that supports their findings. So Hill, Corbett and Rose (2010) does not address hiring bias except to cite a report that Williams and Ceci say shows no anti-women bias. This report is not strong support for their claim that hiring bias is the major factor used to explain gender gaps.
The third piece of evidence they provide is Beyond Bias and Barriers: Fulfilling the Potential of Women in Academic Science and Engineering from the Institute of Medicine and National Academy of Engineering (National Academies Press, 2007). This report does make an unsupported claim in the document summary the women are less likely to be hired. But a deeper look into the documents shows that the chapter on hiring related bias instead looks at gaps between the available talent pool and those who are hired. The report explains those gaps with a long list of challenges that impact productivity (such as access to resources and mentoring) and with the usual list of reasons that women self-select out of tenure-track jobs (from work-life balance to poor working environments). Except for a brief mention in the summary, this document makes no claims that hiring bias is even present let alone that it is a major factor in women’s underrepresentation.
And finally, the fourth piece that they cite to support these claims is the American Association of University Professors Gender Equity Indicators 2006 (West & Curtiss, 2006). This report is mostly a summary of statistical representation of women in various positions. It suggests that there is a mismatch between those who are hired (especially at research universities) and the presumed talent pool of doctoral students. There is no data collected to explain the mismatch. Discrimination is mentioned as a possibility but it is not clear if this means that women leave because of discrimination they experience in the workplace or if they are discriminated against during the hiring process. And again, this is only one small part of a larger analysis of issues related to tenure, promotion and graduate school persistence. So to say that it supports claims that hiring bias is the major explanation seems unrealistic.
And in reference to the impact of hiring bias, absolutely no evidence is provided that hiring bias is a major fear among young female scientists or that it even one of the reasons women choose not to apply for tenure track jobs. None. For the record, studies that do ask them what they fear and what they experience instead highlight bullying, sexual harassment, and hostile environments. (And this is not even to touch on the fact that this study treats fears and experiences of bias as homogenous among all women, including presumably women of colour, women with disabilities, LGBT women,…)
To their credit, the question of whether bias is really prominent as an explanation is something Williams and Ceci address in the supplementary materials published alongside their study. But I don’t think the answer is very compelling because, as above, most of the support they cite isn’t actually about hiring bias or includes hiring bias as only one small factor. For example:
“Psychological research has shown that most people–even those who explicitly and sincerely avow egalitarian views–hold what have been described as implicit biases … There are countless situations in which such mechanisms are triggered: classroom situations, hiring committees, refereeing of papers for journals, distribution of departmental tasks (research, teaching, admin.) etc.” Oct. 2, 2010 at http://www.newappsblog.com/2010/10/implicit-biases-1.html
“It is now recognized that (sex) biases function at many levels within science including funding allocation, employment, publication, and general research directions” (Lortie et al., 2007, p. 1247).
“Research has pointed to (sex) bias in peer review and hiring. For example, a female postdoctoral applicant had to…publish at least three more papers in a prestigious science journal or an additional 20 papers in lesser-known specialty journals to be judged as productive as a male applicant…The systematic underrating of female applicants could help explain the lower success rate of female scientists in achieving high academic ranks” (American Association of University Women: Hill, Corbett, & Rose, 2010, p. 24). Huh, this example is particularly interesting. The way this is quoted in the supplementary materials excludes the important information that this quote is describing a study of bias in peer review during funding applications not hiring. The actual quote is “Research has also pointed to bias in peer review (Wenneras & Wold, 1997) and hiring (Steinpreis et al., 1999; Trix & Psenka, 2003). For example, Wenneras and Wold found that a female postdoctoral applicant had to be significantly more productive…” Wenneras and Wold were looking and postdoctoral funding applications to the Swedish Medical Research Council. And for the record, Trix and Psenka (2003) is a study of the discrepancies in the reference letters that male and female scientists receive. It too is not a study of the impact of hiring bias of the type that Williams and Ceci have examined.
A few of the quotes they provide do refer to one study of hiring bias that they directly refute as inappropriate because it is not about faculty hiring (Moss-Racusin et al., 2012). It’s about hiring a hypothetical laboratory manager. And that’s a fair point about that study, but the authors of that study are clear that the impact is on students and not faculty hiring, an appropriate interpretation of the data I think. But saying that some people have cited a single study that you disagree with isn’t enough evidence to say to CNN that the prevailing wisdom is that hiring bias in the main factor in underrepresentation.
Okay, at this point, it’s probably important to ask if this level of scrutiny is really valuable. How much does it matter that they haven’t supported this argument very well. It’s the data they collected that matters, right? Well, no. The data is interesting (even if a bit flawed), and it’s something that could be used to help focus attention on more salient issues. That’s important and I’m glad someone is doing experimental research on hiring bias. The problem is the way they are talking about and promoting the study, especially in public venues. As I said earlier, a huge influence on girls and women in science (and all people in science really) is what the important people in their lives think about their participation in science. If family and friends value science and express confidence in girls’ and women’s abilities, they are far more likely to persevere. The social support they receive during times of difficulty is essential. Social support can include very specific actions, but it can also be built in the casual conversations the people have with each other about women in science. This study is framed to be provocative and to have an influence on those conversations. The authors wouldn’t have published at CNN if they didn’t want this to be a dinner table, playground, and soccer team conversation piece. Yet, without much evidence that hiring bias is the major obstacle, this study adds a strong voice to the public conversation about science that says: “Guess what, no bias! Just choose to apply to tenure track jobs!” How might the CNN piece (and even the study) sound around the water cooler?: “That thing about women in science struggling to get jobs, totally a myth.” If there were compelling evidence that bias was the cause of underrepresentation and now it was solved, I would be strongly in favour of changing these conversations. But this is not the case.
The underlying implication then is that women are at fault if they experience bias or discrimination or bullying or harassment or withholding of resources. I think, and I’ll be clear that I’m speculating here, that promoting studies like this by saying that there are simple and clear reasons for women’s underrepresentation and suggesting that they’ve been solved will actually have the opposite effect from what I think Williams and Ceci intend. I think they hope to encourage women to apply for prestigious tenure track jobs. By influencing social conversations about women in science, this message might instead cause women to blame themselves when they experience bias and discrimination. Blaming themselves and receiving less support from family and friends who may now doubt the challenges they face could turn more women aware from science at critical junctures such as job application.
So I would ask us all to take care in how we share this story and how we talk about it. Potentially interesting, sure, but also potentially damaging when presented as something that it isn’t.
Further reading and references:
Bevan, V., & Learmonth, M. (2012). ” I Wouldn’t Say it’s Sexism, Except That… It’s All These Little Subtle Things”: Healthcare Scientists’ Accounts of Gender in Healthcare Science Laboratories. Social studies of science, 0306312712460606.
Clancy, K. B., Nelson, R. G., Rutherford, J. N., & Hinde, K. (2014). Survey of Academic Field Experiences (SAFE): trainees report harassment and assault. PloS one, 9(7), e102172.
Johnson-Bailey, J. (2014). Academic Incivility and Bullying as a Gendered and Racialized Phenomena. Adult Learning, 1045159514558414.
Moss-Racusin, C. A., Dovidio, J. F., Brescoll, V. L., Graham, M. J., & Handelsman, J. (2012). Science faculty’s subtle gender biases favor male students. Proceedings of the National Academy of Sciences, 109(41), 16474-16479.
Simpson, R., & Cohen, C. (2004). Dangerous work: The gendered nature of bullying in the context of higher education. Gender, Work & Organization, 11(2), 163-186.
Wenneras, C., & Wold, A. (2001). Nepotism and sexism in peer-review. Women, science, and technology. Routledge, 46-52.
[i] Jessica recently presented a summary of the project motivations and submissions: Brinkworth, J.F., & Shanahan, M.-C. (2015, March). Surviving the Sexodus Project: How STEM Women Approach Career Challenges. Presentation at the American Association of Anatomists, Boston, MA. And we are embarking on a related research project: more on that soon!