Jo Handelsman and Corinne Moss-Racusin like to shed light on the unseen and unnoticed. Though Handelsman, a molecular biologist at Yale University, primarily researches bacteria, lately she teamed up with Moss-Racusin, a psychology post-doc also at Yale, to study something equally invisible: unconscious biases in the scientific community.
In a recent paper published in PNAS, they and their co-authors find that many university faculty, regardless of their own gender or age, exhibit bias against female students. Science professors at major research universities in the U.S. gave very different evaluations for a hypothetical lab manager prospect based on a single difference – the candidate’s first name. The male candidate “John” was more likely to receive a job offer, a higher salary, and mentoring than his counterpart “Jennifer.” Scienceline spoke with Handelsman and Moss-Racusin about the paper and their thoughts on the struggle to increase the presence of women in science.
How did you become interested in this issue?
HANDELSMAN: I run a science teaching group where we look at the factors that increase or decrease success of students in science and technology majors. I had given a lot of talks on unconscious bias and in every single talk someone would say afterwards, “but we’re scientists, we are trained to be objective so we wouldn’t have these biases.” And that of course misses the point that these biases are unconscious so they’re not something you can logically talk yourself out of […] So that just didn’t ring true, but I never had an answer to it.
Why publish now?
HANDELSMAN: Scientists are going to continue to say “this doesn’t apply to us” until we have evidence to the contrary. We are about getting the evidence.
Just last year, a paper in PNAS said gender discrimination in the sciences was a thing of the past. How does your study respond?
MOSS-RACUSIN: What their study is able to say is that when women do reach the point of applying for faculty jobs, the correlational evidence shows they do well. They’re likely to be hired. They’re likely to be invited for interviews. What we’re interested in is why are so few women reaching that point in the first place. So to really answer that question, we needed to do a controlled experiment to determine, all else being equal, if you have equivalently qualified male and female students, are reactions to those students identical?
Can this study be replicated or is the cover blown?
MOSS-RACUSIN: There are obviously complications with doing more of this type of study because this one, I think, has gotten so much attention, so anytime a faculty member gets a survey in the mail, they’re probably going to be a little bit suspicious. With the faculty participant pool, I genuinely wonder whether it can be replicated now. You would need a clever alternate design with a different methodology or cover story. I do think that it could probably be replicated with laypeople because these stereotypes about the male gender-typing of science are pretty robust. We know from other work that if you ask very young children to describe scientists, they tend to describe an older white man.
What about the results was most surprising to you?
HANDELSMAN: The specific result that discouraged me and really had me thinking differently, perhaps, about the biases against women undergrads, was the response to the mentoring question, “Would you mentor this student?” That the faculty would be more likely to mentor a man than a woman, I found profoundly disturbing.
What kinds of bias have you experienced in your own career(s)?
HANDELSMAN: It’s very difficult to assign any one decision or outcome to bias because we don’t have a controlled experiment, […] it can always be explained away by some other variable. I’ve had a series of papers over a 20 year period – papers on what I thought was a novel or important finding – rejected from one of the particularly high profile journals, then had the experience of within a few weeks or a few months seeing a paper by a man or group of men that either showed the same thing that we were told wasn’t of general interest, or used the same method that we were told was flawed. And for each one my initial reaction was “oh I guess it just wan’t good enough.” But in retrospect, after this happing over and over and over, I do have to wonder, if there was a gender component to all this rejection.
What do we do with this information?
MOSS-RACUSIN: Now we’re focusing on intervention work to see how to mitigate this bias and improve faculty members’ attitudes toward diverse students. But in the meantime there are things departments can do to help guard against the influence of these biases. The way mentoring happens in academia does kind of invite the influence of subtle gender bias because it happens in such a one-on-one private way between individual mentors and their students. There’s no system in place to check how the mentoring relationship is progressing and to ensure it’s progressing equally for different types of students. Simple changes can make that relationship more transparent. Establish beforehand how many times mentors should meet with their different graduate students and maybe assign secondary mentors as well. We know from the literature that having multiple role models is one of the key predictors of success, especially for female students and students of color.
Besides recognizing bias, what can the scientific community do to increase the presence of women?
HANDELSMAN: The issues with women are not just bias, but also the challenges of combining family and career, which does represent certain challenges that men may not have. They are not always exclusively challenges for women and men are not always exempt from them. Addressing the very practical challenges that women face in terms of timing of careers and childcare access […] is absolutely essential to enhancing the representation of women.