Scientists, Your Gender Bias Is Showing

Nobody who is familiar with the literature on this will be surprised, but it’s good to accumulate new evidence and also to keep the issue in the public eye: academic scientists are, on average, biased against women. I know it’s fun to change the subject and talk about bell curves and intrinsic ability, but hopefully we can all agree that people with the same ability should be treated equally. And they are not.

That’s the conclusion of a new study in PNAS by Corinne Moss-Racusin and collaborators at Yale. (Hat tip Dan Vergano.) To test scientist’s reactions to men and women with precisely equal qualifications, the researchers did a randomized double-blind study in which academic scientists were given application materials from a student applying for a lab manager position. The substance of the applications were all identical, but sometimes a male name was attached, and sometimes a female name.

Results: female applicants were rated lower than men on the measured scales of competence, hireability, and mentoring (whether the scientist would be willing to mentor this student). Both male and female scientists rated the female applicants lower.

This lurking bias has clear real-world implications. When asked what kind of starting salaries they might be willing to offer the applicants, the ones offered to women were lower.

I have no reason to think that scientists are more sexist than people in other professions in the US, but this is my profession, and I’d like to see it do better. Admitting that the problem exists is a good start.

This entry was posted in Women in Science. Bookmark the permalink.

235 Responses to Scientists, Your Gender Bias Is Showing

  1. TW says:

    I am not convinced about your interpretation.

    I would argue that experienced researchers use all information available, and sex is additional information in two ways:

    1) The woman on average worked harder to get the same qualification, leaving a man with a greater potential for growth.

    As mentioned before, women are more conscientiousness. Across my student years, many just got better marks, because they did homework well and studied more regularly. Even though some got better marks than myself for example, I always felt they were closer to their limits.

    I recently had a class reunion where I discussed with a female school friend who was the No 1 math student why she never did math at university and “just” became a middle school teacher. I told her: Why did you never do it? You were better than me! She said: No I was not better than you, but I worked so much harder and regularly. I felt my limits. But you were just totally lazy, disorganized and de-focused and still passed!

    2) Women get pregnant. This is a real disadvantage and risk for any project leader. I witnessed myself that a project leader hired a woman with all good intentions, but she got pregnant just after, promised to keep working, but then left. His project was delayed significantly and he said “never again”.

    So given the same qualifications, I would rationally go for the man.

    Having said all this, people are not hired by CV but in direct interviews. Either they connect or not.

  2. NG says:

    “[L]eaving a man with a greater potential for growth” seems like a pretty bold assertion to me. Is there a some secret cap to one’s potential for professional growth that might be fully used by the time one is applying to become a lab manager?

  3. Saurs says:

    NG, according to TW that secret cap would be called a uterus, apparently.

  4. Statistical Discrimination says:

    “I know it’s fun to change the subject and talk about bell curves and intrinsic ability, but hopefully we can all agree that people with the same ability should be treated equally.”

    I’m not sure that the issues can be so completely separated.

    Suppose that you applied the same test, taking a CV and switching out a feature, but instead of sex you used PhD program, e.g. Harvard vs University of Florida. In each case you see the same publications, but you might give more benefit of the doubt to the Harvard PhD, since the Harvard PhDs have higher performance on average and there is noise in the data. You might also favor the Florida PhD for achieving more with less resources. Using your prior information about the distribution of ability for Harvard and Florida PhDs gives more accurate estimates than ignoring it, but penalizes folk from one school or the other.

    If it is really the case that aggregate performance for female applicants (or male applicants, as seems to be the case for undergraduate students at the median) lags, then an accurate estimate of a female applicant’s quality will have to take that prior information into account, in addition to the other evidence in the CV.

    There are then two questions to address regarding hiring bias. First, are the stereotypes accurate regarding real applicants in the world (there is a large literature on stereotype accuracy indicating that stereotypes usually are pretty accurate: If not, then the bias needs to be countered by means such as blinding hiring committees to the sex of applicants, and hire quality can be increased at the same time as sex disparities are reduced.

    If the stereotypes are accurate then we need to consider distributional consequences. Using the stereotype will improve the overall quality of hires, but a given woman will suffer relative to the sex-blind policy absent a countervailing hiring preference for women, and it will make sex and gender salient. So we may want to take a hit in average hire quality to promote equality of outcome between the sexes by sex-blinding hiring (although sex-blinding interferes with affirmative pro-female preferences). But we shouldn’t pretend that treating people with equal non-sex indicators of ability equally is necessarily the same as treating people of equal actual ability equally if there are group ability differences, since our measures of ability are noisy.

  5. female says:

    @TW: Seriously!?! I hope you never make a hiring committee! It’s because of people with an attitude like yours that the problem exists in the first place…

  6. AL says:

    TW, I see absolutely no flaws in your argument. Be sure to prominently note on your CV that you regularly worked below your potential. I’ve heard employers love that.

  7. chris says:

    for all those who suggest double-blind hiring/reviewing etc. is the solution: it is not. the scientific community of a given subfield is typically so narrow that you know whom you have to review. everything is “networked” which makes it a lot easier for all sorts of biases to creep in.

  8. AsianAstronomer says:

    Really? Only females are discriminated against? What a simple world that would be!

    What about the dozens of submissions from Asia to American and European journals that get nowhere, owing only to the biases of the referees and editors? Why are they able to get away with reasoning that is flawed at best, when rejecting papers? Why not talk about that?

  9. TW says:

    @Al: It is not that I did not do stuff but I did not do the stuff that people, my parents or the university told me to do such as programming, trying to hack computer games, philosophy, long discussion on useless stuff like consciousness or political stuff, or try to optimise PI computation. I did what I was interested in. So I did learn a lot but not what I was supposed to learn. As a result my grades were not exactly brilliant but sufficient to pass through the system. Judging only by degrees (or system obedience) is a bad proxy for success as all these things are NOT on my CV.

    @female: I am interested in the bottom line and not in political or ideological obedience. Any market would price the risk of pregnancy to a project. I know one project that got severely delayed with consequences to other people including females. Our society sacrifices common good for individual rights.

    @NG: I did not say that men in general have a higher growth per se but for equal grades and qualification!

  10. DrMobs says:

    We don’t seem to have this problem.

    The last jobs we advertised (research assistants) we had 90 applications, 34 of whom were eligible on qualifications.
    Ranking on experience only, we gave interviews to the top ten. Nine were female; the one male didn’t show up.

  11. bad Jim says:

    TW: one of my most talented employees did leave abruptly after her second pregnancy. I’ve also had to fire two male employees personally for gross incompetence (and I hope you never have to do anything like that, because it’s a miserable gut-wrenching experience). I’m old enough that most of the women I’ve worked with either already had kids or had opted out (not always successfully, oddly enough) and they were condescendingly more dependable than my male colleagues.

    A boy’s club is no longer a viable model for any sort of enterprise, whether it be a company, a university, or an advocacy group. Sausage fests should be restricted to wiener roasts.

  12. APEER survey says:

    It would be good to see a similar study on how women and men are chosen as peer reviewers. Are selection criteria applied fairly? ( Becoming a peer reviewer confers some of the advantages – recognition and a chance to shape science – that Sunil D’Monte mentioned above.

  13. AI says:

    A rational agent should factor risk of pregnancy into it’s decision.

  14. cory says:

    @AI & TW

    Please, let me introduce you to federal laws. Following them will avoid the next woman you don’t hire suing you for it.
    There’s this one:

    Title VII of the Civil Rights Act of 1964 (Title VII), which prohibits employment discrimination based on race, color, religion, sex, or national origin;

    Title VII prohibits not only intentional discrimination, but also practices that have the effect of discriminating against individuals because of their race, color, national origin, religion, or sex.

    Title VII’s broad prohibitions against sex discrimination specifically cover:

    Pregnancy Based Discrimination – Pregnancy, childbirth, and related medical conditions must be treated in the same way as other temporary illnesses or conditions.


    A rational employment system would ensure that the compensation for work performed was enough so that the financial gain from returning to work was actually that, a gain. Many women with intentions to return to work find a net loss when factoring in the cost of day care to (see above) already lower-than-male budgets. Perhaps if women were paid equally the $12,000 yearly cost in child care wouldn’t matter (as much) in their calculations. As I look at it, if a woman starts $5,000 behind, and then has $12,000 a year in child care costs, she has a compensation gap that will only get larger as she starts being shunted aside due to “low productivity”, usually because supervisors see more personal breaks being taken for nursing, pediatrician appointments, etc, all of which are most evident and necessary in the first year of a baby’s life than any other, exactly when employers are judging the woman’s competence and trying to decide whether “she’s still focused on her work”.

    A rational employment system would provide on site day care so that female employees could return to work sooner without risking their child’s health by having to pump breastmilk in bathrooms for infants at sub-par care facilities.

    A rational science field would allow for work-life balance because healthy, well-adjusted employees and scientists are freed to focus on the task at hand. We don’t have a rational system by any means. The laws attempt to correct for that.

  15. TW says:

    I do not advocate discriminating any one per se. And I feel concerned as unlike women I have a real physical handicap. I strongly feel that we should look at the individual and see whether it’s the best choice for YOU. But if my handicap makes me likely to perform less well, i just have to accept that the equally qualified not handicapped should get the job.

    The point is just that hiring a woman with equal qualification is likely a worse choice due to what I explained above. Not because of the mere fact that they are females, but because of REAL disadvantages such as pregnancy risk and lower growth potential.

    Laws are not rational but the product of political forces driven by ideology. The law current tries to make equal what is not equal.

  16. TW says:

    @Cory: Have you considered in your calculation that men die earlier by three years or so, and therefore have reduced the pension contribution for male scientists?

  17. JMW says:

    TW: You cite as evidence one woman who got higher marks than you and admitted to not being as good as you, and generalize that to the entire gender? Frankly, I wish I was not as committed to civil discourse as I am.

    Let’s spell this out. The study presented individuals with one CV and asked that individual to rate the candidate. For some individuals, the CV had a male name attached; for other individuals it had a female name attached. On the aggregate, the individuals rated the female singificantly lower despite all the other information on the CV being identical.

    Adam: although this has been answered, there was no screening for specific qualifications. There was one (1) CV that each person in the study was asked to rate, and the CVs were all identical except for the female/male name at the top. Stop trying to apologize for sexism.

  18. Peter Erwin says:

    Jodi @ 17:
    This study would suggest a similar bias also exists against female authors. Double blind peer review would even the playing field

    As chris (@32) pointed out, double-blind peer review is, in practice, very difficult to achieve, since it’s often rather easy to figure out who the authors are. This Inside Higher Ed article mentions several economics journals giving up on double-blind peer reviewing for a number of reasons, including the infeasibility of double-blind review; the editor of one of the journals mentioned that “[his] journal did an experiment typing in the titles of 20 recently submitted papers and was able to correctly link almost all of them to authors, who post working papers, talks given at meetings or information about their research on various websites.”

    In addition, I think the available current evidence is that there isn’t any significant bias against female authors in peer reviewing at scientific journals (at least within biology); see the meta-analysis in this PNAS study from 2011. (Of course, this leaves open the possibility that bias can still exist in some fields or sub-fields; most of the studies discussed in the meta-analysis were in biology.)

  19. TW says:

    @JMW: I just gave you one example. I spent more than 10 years in this environment. I saw it over and over again. Grades and ability to do good science do not 100% correlate. Women just get better grades by studying more and more regularly but that doesn’t mean the men didn’t learn other stuff in the mean time.

    The CVs were not the same. Sex is an additional piece of information. And we are humans and are all biased as we take a lot of unconscious decisions based on our experience. So many might have unconsciously discounted for pregnancy risk and potential.

    I would be interested in whether just men had this bias or also female reviewers.

    And the study is somewat unrealistic, because NO-ONE is hired by CV but by interview.

  20. cory says:

    @ TW

    The same laws that apply to discrimination against women apply to discrimination against handicaps, as you describe your condition.

    As per your statement, you are saying that you would just accept not being hired if you knew for a certainty that the only difference in the decision was your physical handicap? Even knowing that that was an illegal practice? As long as women don’t get any “favors” either?

    I am inclined to believe that you are not that lacking in self-interest.

  21. Pingback: Useful new study « Feminist Philosophers

  22. TW says:


    1) no two candidates are exactly the same. People always have a preference for someone. And if it’s just that s/he looks like your favourite niece. It makes you as a boss feel better so why not take him/her.

    2) if s/he doesn’t want me just due to my handicap, then I don’t want him as a boss anyway.

    3) if s/he is concerned that my handicap has real disadvantages (as pregnancy has!), the law would forbid him to discuss it openly with me and say “I would but I don’t because you have this disadvantage”. I would not have the chance to give him counter arguments.

    4) If I had the choice, I would rather take my clone without my handicap.

  23. AI says:

    Cory I am only talking about logic – pregnancy is a factor for females and so a rational decision process should take it into account. The things you cite in your post only support this.

    Now if a society decides, for whatever reason, that equally qualified females should have equal opportunity for hiring this effect has to be countered somehow, for example by enacting laws penalizing discrimination similar to ones you cite or rewarding hiring of females.

  24. Drake Sullivan says:

    “Please, let me introduce you to federal laws. Following them will avoid the next woman you don’t hire suing you for it….”

    Exactly. Legal threats. So when I see identical resumes from, say, Bob and Fiona, I know that Fiona has achieved all her prior positions with the implied threat of legal action if she doesn’t get the job. Bob has achieved his without this benefit. So, if all their achievements are equal, Bob’s likely to be that little bit better.

  25. Lily Yang says:

    I sometimes feel less comfortable and able to voice my concerns around female colleagues than male ones. Sometimes it seems to me that my female mentors and colleagues, more so than the male ones, have expectations for how I should make career and family choices and view me more critically for not choosing according to their idea of correctness. This is particularly difficult because they are perceived “on my side.”

    I have drastically fewer data points for female mentors and colleagues than male ones, though, so I suppose that impression is not reliable.