Monday, May 05, 2008

Gender Bias at Fermilabs

Sherry Towers has posted a paper preprint to arXiv which look at gender bias in particle physics at Fermilab: A Case Study of Gender Bias at the Postdoctoral Level in Physics, and its Resulting Impact on the Academic Career Advancement of Females. She focused on the 9 female and 48 male post docs who worked on on the Run II Dzero project from 1998 to 2006. Using information in public databases regarding internal publications and presentations, she compared scientific productivity to the number of conference presentations given. As the paper explains:

Conference presentations are important to the career advancement prospects of postdoctoral particle physicists primarily because they give a young physicist much needed positive exposure to future employers (important because postdoctoral positions are inherently temporary). Conference presentations are also important to career advancement because the names of all of the several hundred collaborators are included on the author list of refereed publications. The names appear in alphabetical order only, and there are no first authors. This is because particle physics in theory strives to be an egalitarian field where the contributions of all physicists in the collaboration are given equal merit.
Significantly, while every participant was expected to present publications describing the progress of their work, the selection of who would present the project's data at conferences was determined by a committee behind closed doors.

The results of Tower's analysis is striking:
  • The females in the study were more productive (as measured by internal publications), on average, than the males. Half of the males produced fewer internal papers per year than the least productive female in the sample. There was a much broader distribution for the male post docs: "nearly all the females are highly productive, whereas 1/2 the males produce almost nothing, somewhat half are moderately productive, and a select few are extremely productive." Note that internal publications were used to measure productivity because peer-reviewed journal publications always list all of the project's participants in alphabetical order as authors.
  • Males were much more likely to be alloted conference presentations. The ratio of physics conference presentations to internal physics papers produced for males was double that of females (triple if all presentations and papers were considered).
Since conference presentations are important for career advancement, the relative lack of presentations at conferences puts the female physicists at a disadvantage relative to their male colleagues. While roughly similar percentages of the male and female postdocs moved on to faculty positions, Towers points out that ". . . the females in our cohort have worked significantly harder than their male peers to achieve this “equity” in academic career advancement, and yet some highly competent female physicists nevertheless appear to be prevented from moving on to faculty positions because of the conference allocation gate-keeping mechanism."

In an interview with the journal Nature she explained that she believes this bias selection of conference presenters is likely unconscious:
"I don't think for a second that there is a conscious bias going on," she says. But the committees "are in danger of being prone to patronage and cronyism". Male committee members are more likely to nominate male protégés to receive presentations time, she claims.
University of Washington astronomy professor Julianne Dalcanton has her own take at Cosmic Variance:
. . . I doubt that anyone participated in conscious “Girls can’t do math” types of sexism. I think you get a net effect from a series of small decisions that seem defensible when taken individually, but that add up to a real net bias. We have a natural tendency to recognize talent that looks like our own (ie. people whose strengths lie in careful detailed analyses often find fast-moving creative types to be shallow and showy, and fast-moving creative types are more likely to find careful deliberate scientists to be plodding and dull). Thus, when passing out conference presentations, you naturally want to give them to someone that you think will do a good job, where “a good job” frequently means “do the job the way I would have done it”. This leads to a tendency to give breaks to people who fit into a mold that you already know and respect, and require extra proof of merit from people who lie outside that norm. The net result can be gender biased, even if the decisions that produce the result aren’t intrinsically gendered.
I would think that such biases are especially difficult to counter, since the committee members are likely to feel that they are actually allocating conference slots based on "fair" criteria.

Not surprisingly, the paper has stimulated a lot of discussion. The female physicists that Nature spoke with weren't surprised by the analysis:
"You often see a young guy with an older guy gossiping an dhaving coffee, but never a woman," says Freya Blekman, a physicist on the CMS experiment at CERN. "Im convinced," agrees [CERN physicist Pauline] Gagnon. "There is absolutely no shodow of a doubt in my mind."
Fermilab, on the other hand, appears to be dismissive of Tower's analysis. An internal investigation at Fermilab determined that the project had "followed its policies correctly", and the leaders of the DZero project claim that if there ever was discrimination, there isn't now, because from 2006 to 2007 women "gave 17% of all talks despite making up just 12% of the collaboration." That seems to miss the point of the analysis: Towers isn't suggesting that women be given conference presentation slots on a quota basis relative to their population withing the project. Instead female postdocs should be allotted presentations based on their productivity. If her analysis holds true that the current female postdocs on the project are substantially more productive than many of their male colleagues, I would have expected their percentage of conference slots to be even higher than 17% if they were strictly merit based.

Not surprisingly, this has stimulated a lot of discussion online. It's interesting to read people's own experiences within physics. And yes, there are criticisms of the study, both thoughtful and not so much. Check out:
Tower's own experience as a postdoc in particle physics at Fermilab who found her career derailed because she took maternity leave was covered by the Chronicle of Higher Education. That article is behind a pay wall, but Zuska covered it in her old blog. Towers is currently in the Department of Statistics at Purdue University.

Tags: ,