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.
What I like about TW’s logic @26 is this:
1) A woman works her ass off – don’t hire her.
2) A woman doesn’t work when she gets pregnant – don’t hire her.
That’s what I call a lose-lose situation.
Perfect!
Not.
@TW, your initial post stated that you were ‘not convinced about the interpretation’, because people ‘use all the available information, including gender’. This is a ridiculous argument. The study was designed to test the specific hypothesis that gender would explain variation in how CVs were judged in terms of candidate competence. Therefore, the researchers controlled for other sources of variation (e.g. qualifications and so on), and only changed the gender. So the interpretation that gender explained a significant amount of variation in how these CVs were assessed is a totally justified and straightforward interpretation.
Further, it’s also strange to me that you disagree with this interpretation, then go on to make a chain of sexist and unjustifiable remarks, based on anecdote and your own personal biases – which brings in to question your own ability to make a judgement based on scientific data, as well as rather ironically supporting the researchers’ interpretation of their data, since you are openly admitting to exactly the kind of sexism that their research suggests is happening elsewhere.
Your comment that women should be treated as some kind of ‘risk’ factor because they are able to get pregnant is ridiculous. Yes, a woman might take time off work while they are pregnant, but then a person of any gender might need to take time off work for any number of reasons (illness, family bereavement etc.), and it is not the employer’s position to guess how likely these circumstances might be, nor do they have the necessary information to make any kind of reasonable guess.
I am a young female scientist who so far has shown a lot more scientific skill than many of my male cohort. I am also biologically unable to have children. Should I state this at the top of my CV to improve my job prospects??
Pingback: Friday links | Dynamic Ecology
Pingback: Scientists Rate Female Job Applicants Lower
@Fiona: if you are unable to have children, then stating it would increase your probability to get hired because you have a lower risk. but given the current social environment you would be viewed as strange and awkward to put it on your cv, but you can say it at the interview.
@MNb: >> 1) A woman works her ass off – don’t hire her.
absolutely, if i want to have top scientists and not just work slaves.
if i am a sport coach and i see someone winning or equalling games without much training against someone who had trained hard, i would immediately take this person.
@TW: It makes me sick to think that someone with dated and unscientific views like yours might be in a position to judge applicants for a science job, or any job for that matter.
I have no problem with the study perse, but the notion that we must correct for these subconscious biases is really quite unmanageable.
The fact of the matter is you can probably come up with hundreds of similar biases.
Biases against a right handed person vs a left handed person. Bias against styles of handwriting. Bias against redheads. Bias against people with stutters or asymmetrical faces. etc etc
At one point, you just throw your hands up in the air. Human beings are not perfect, and its probably impossible to arbitrarily design a selection algorithm that does much better anyway. Some of these biases indeed have been imprinted into our minds for a reason (evolutionary biology likely favors individuals who can more readily find and create groups)
I am not very surprised by this level of gender bias in science. It is completely ridiculous from the ethical point of view, while at the same time quite expected from the performance point of view. The employer takes into account the potential absence of the applicant due to pregnancy leave, and as a result favors males, or offers less salaries for females, or both. It’s a simple calculation of potential damage control. It is completely unfair to women, but then again business in general is never fair. It’s always about effective exploitation — a man can be exploited more effectively, since he does not get pregnant. Therefore prefer men when hiring. As simple as that.
Lucy (126):
“What if they don’t want a kid, ever? Ever met one of those? Do you think it’s fair to
not let them have the job, even if they’re more qualified?”
I wonder, what if the boss requires the applicant to sign a document that he/she will not ask for a paternity/maternity leave during the lifetime of the contract? How many women would agree to sign such a document in order to be treated equally as men? I am assuming that most men would have no problem signing such a document, if they really want the job.
Also, would such a thing be legal? Can an employer ask the applicant to give up their right to go on a maternity leave, in order to get the job? If the employer actually asks for such a document, would that eliminate the gender bias? Or would it just eliminate women who want to have an open option wrt. pregnancy?
Btw, regarding the legal treatment of pregnancy on the same footing as being sick — it is just ridiculous. Illness is something a person will always avoid if at all possible. Pregnancy is something that can be wanted and planned, several times over. I just cannot believe that pregnancy leave can be treated on the same footing as cancer-treatment leave, or such. US laws always keep amazing me with the level of nonsense in them.
Just to add a different twist to the conversation. The authors aggregate their data and compare treatment of male versus female applicants. That’s fine. It tells a story and allows them to write their grant.
Another way to look at their own data (Table 1) is to separate the hiring faculty by gender and see how they view male and female applicants. One way to do this is to calculate a female:male ratio using the mean values supplied. Thus, on competence, male faculty rate females at 3.32 and males at 4.01 for a ratio of 82.7….female faculty rate females at 3.33 and males at 4.1, for a ratio of 81.2%. If we do this for all four measures, we get:
Competence: 82.7% (male faculty)….81.2% (female faculty)…one could argue this is a wash, but I didn’t have time to do formal stats on this.
Hireability: 79.1 % (male faculty)…72.4% (female faculty)….no way this is a wash.
Mentoring: 84.4% (male faculty)…82.6% (female faculty)….also tough to say if this is even. Maybe, maybe not.
Salary: 88.8% (male faculty)…85.2% (female faculty)….most likely not a wash.
It is interesting how across all measures, male faculty seem to judge female applicants less harshly (relative to male applicants) than their female faculty counterparts. Why is this so? Any sociologists care to chime in? And this is not just the ratios that are talking, but the raw data as well. Female faculty seem to judge female applicants lower than male faculty…and….what I found most paradoxical…is that female faculty tend to judge male applicants EVEN HIGHER than their male counterparts. Indeed, the hirebility difference above is mostly due to the fact that a female faculty will judge the same male applicant to be more hireable than her male faculty peer (3.92 versus 3.74)!!! And male faculty are willing to give a much higher starting salary (27,111.11) to a female applicant than a female faculty is (25,000).
So, what’s another take home message here, besides the obvious “yes, there is a gender gap”? One might also conclude that a woman applicant might be better off having a male faculty judge her candidacy (at least relative to equally-qualified-on-paper males) because she’s more likely to be offered more money by the male faculty, and less likely to be judged as “less hireable” than a male applicant by the male faculty.
As women are underrepresented in the faculty ranks (indeed, the ratio seems to be about 3:1 in this sample), one could also envision how the situation would be even worse for women if the faculty split were 50:50 because there would be more women faculty to judge the female candidates more harshly…and offer them less money! I’m NOT…repeat NOT…suggesting that hiring more women into faculty is bad. I am suggesting that there are other issues at play beyond the obvious gender gap. And as irresponsible as it is for people to claim that there is no gender bias, it is equally irresponsible (based on this data) to reduce the argument to this must be another case of the uterine-free oppressing the uterine-endowed.
Discuss.
Pingback: Global Newsflash – Breaking News, US, World, Technology, Enertainment, Science | Science Confirms The Obvious: Science Faculty Think Female Students Are Less Competent
I’d like to address my fellow female scientists who are commenting that they don’t intend to have children so they should not be discriminated against based on the pregnancy issue:
I appreciate the point that you are trying to make, but your comments make me a bit uncomfortable because at some level they validate the argument that a woman being pregnant/having a child makes her undesirable as a scientist. It seems clear to me that the “pregnancy catastrophe” is a specious argument attempting to put a veneer of logic/pragmatism on what is really just regular old sexism.
TW has no data, just what he thinks are logical conclusions, but are really hunches. I wouldn’t worry about him. TW is not the scientist you want to be working for.
The problem is that there seems indeed to be no data on how many female scientists leave work permanently due to pregnancy. I would certainly guess it is less than the female workforce generally. Without that, there is not even a statistical justification for hiring bias, especially if there is no data on male disability-related absences.
The whining about figure 2 is indeed hilarious.
I don’t know how anyone could look at this study and say it doesn’t show a big problem. It hasn’t been a problem for me personally (female scientist) but there’s too much money in my field at the moment , and too few skilled people, to allow for that. I am fortunate enough to be in demand, and to have been raised to be assertive more than likable. but I know it’s fortune. I preferentially hire women, because someone has to balance this stuff out.
Does PNAS also include social sciences ??…I don’t understand why this shitty article was published in PNAS…..my physics paper which had original experimental observation for the first time was recommended for society level journal and here they are publishing this kind of crap…I can’t believe it….
#138 funny, I can totally believe it.
Statistical Discrimination made the most interesting comment. This is a bit of conceit on my part, as I made almost the same comment when I heard about this study. Here’s a simple example of how this can happen using real facts.
The average IQ (or g factor) for men and women is equal, but men have significantly higher standard deviations. If you condition a sample from these two distributions conditional on some minimum intelligence, the resulting distributions will have men with higher intelligence on average.
In other words, the average male and female intelligence are equal, but the average intelligence of men with at least 120 IQ is higher than that of women with at least 120 IQ. If you assume that people applying to jobs in STEM are of above average intelligence, when you work out the math you actually get a correlation between gender and intelligence. This could be affecting people’s perceptions even subconsciously, given that our brains are designed to notice patterns and trends.
Of course, IQ is hardly a perfect measure of job performance, and so on. It would be interesting to look at reasonable measures of grad school success (i.e. number of papers published or sum of impact factors during PhD) and compare them between men and women in the same field, to see if people in academia have reasons to make a connection between gender and competence.
Oh, and a number of people here posted their opinions on how they think women are more competent (blanket statement). Funny how that blanket statement is ok, but the reverse is not. Here are my experiences from a decade of being in university:
1) I work with math, physics, algorithms, etc. Abstract intelligence is valuable and usually quickly inferred in these fields. I find that men are far more likely (even proportionately speaking) to be brilliant problem solvers that have the intuition to do the hardest theoretical problems than women (consistent with my statistical argument above). On the other hand, even in a discipline like theoretical physics, once you get out of classes and into research this is FAR from the most important thing. A good work ethic, persistence, working well with others can all be more important. In those other categories, I find women to be at least the equal of men on average.
2) Women are not as assertive as men. I think this mostly hurts women not because this makes them less efficient managers, but because they don’t stand up for themselves and take credit for their great work. This is really a shame. My advisor is female and I feel like she consistently doesn’t get as much credit for her work as she deserves, because she is not brazen enough about telling people how great it is.
Those are the two main differences I’ve noticed, and both of them (one correctly, one falsely) might condition people to think that a male would do a better job in the same role as a female.
@Erin: the fact that you admit that you preferentially hire women is atrocious. Justify it to yourself however you want, it doesn’t make it one iota more acceptable. It’s just as bad (no better) than the people preferentially hiring men.
#134: “And as irresponsible as it is for people to claim that there is no gender bias, it is equally irresponsible (based on this data) to reduce the argument to this must be another case of the uterine-free oppressing the uterine-endowed. Discuss.”
Andrew, there is nothing to discuss. You apparently lack basic reading skills. The authors of the study wrote: “Both male and female scientists rated the female applicants lower.” This has been pointed out several times in this comment thread. No one is interested in discussing your straw-man argument.
@Nir I think it’s atrocious that you think men are inherently better at math. You may find it to be so in your experience, but people find what they expect to find. You justify one set of generalization (men are better at STEM) with another (women work harder), and you have no data for either one. I’m not apologizing, because that type of sloppy thinking perpetuates this situation, and I want the situation to change. I’m changing it.
It probably doesn’t help the cause to say it outright, because the knuckle-draggers among you will point to me as an example of why it’s ok to discriminate against women. But as they’re doing it anyway, I’m not too worried.
In my post, I said I was giving my personal experiences, as many others did. If you read what I wrote in the previous post, you will see that I don’t think men are better at math. The averages are pretty comparable and the average is the only thing that’s meaningful to talk about when you use the simple word “better”.
There are a lot of people out there though that still strongly believe that males display greater variance in intelligence (I would say it’s a majority, but I don’t know the field well enough. It might well be a majority however). It’s controversial and there are people who don’t agree, but believe me for every paper you find that claims the variance is the same I’ll find one that claims it’s larger.
A larger variance means that men are over-represented amongst people with high intelligence, as they are over-represented amongst people with low intelligence.
All this was already posted. You can disagree with my claims or reasoning, but to say I have no data when I bring up a very standard claim in the testing world for which dozens of studies have been performed and agree is ridiculous.
Nothing you do would justify discriminating against women. Justification has nothing to do with it because the people in the study are not (at least, as far as we know) doing it consciously. That’s the whole point; we want to try to figure out what and why that is, and ideally if practical eliminate it. But you can’t accuse someone of being immoral for having an unconscious bias, everyone has those. But someone who willfully displays a bias is acting immorally.
PS Just to give you an idea, a quick google search on “gender variance intelligence” yielded the following hit: http://www.sciencedirect.com/science/article/pii/S0191886906000420. In the abstract, the over-representation of men in the upper extremes of intelligence is taken as a GIVEN. Then data is given supporting the reason for it is greater variance.
PPS I like how you use the word “inherent”. A difference between two groups of people can only be from two sources: their genetic code, or the collection of events that happened to them since conception. So really what you meant was “biological”. The nice thing about inherent is that it’s sloppier and more volatile.
@Nir I have better things to do than explain myself to you. I don’t accept your judgment of me. If I cared about you or your opinion, I’d argue it. Bye now.
I’m not buying this. Funding is tight, job market is tight, and people are vying for whatever edge they can get to land what they need to support themselves and their family. Because of these, everyone imagines some form of discrimination (including reverse discrimination) if they take a blow to their ego and don’t get what they were hoping for.
Equality is here and we’ve all seen it…sure we’ve seen isolated cases of discrimination and we’ve also seen cases be taken to court. Everyone has the opportunity at this point to apply, be considered, and even fight the outcome by legal means if they feel it is unjust.
This is an example of using a personal interest and common fear to try and gain an edge in the marketplace (in this case it is likely funding seeking or continuance of funding). This study is despicable and very hard to control for ambiguities and experimenter bias.
@Erin: so you call me names when I present a view that doesn’t agree with your perspective that men and women apparently have to be equally good at everything, then when I present more evidence you tell me you don’t care about my opinion and leave the discussion?
You didn’t have better things to do but to post in this blog several times, but suddenly you are too busy to respond?
Clearly your objectivity as a scientist comes second to your need to hold certain views as correct. I am happy to say that your approach is not at all characteristic of the many women in science that I respect, women that would be able to try to view the data and reach a conclusion whether or not that conclusion was convenient.
I dearly hope you aren’t a physicist.
@Nir What name did I call you again? Yes, I have a job that must limit the time I spend being goaded by anonymous blog commenters, sorry about that. It doesn’t make me less of a scientist for not wanting to talk with you (personally) on this subject further.
@Anon
I’m quite happy with my reading skills, thank you. Yes, the authors do state that both male and female scientists rated the female applicants lower. And I do not deny that this is true. Indeed, my post states several times that the conclusion that a gender gap/bias exists is “obvious”. I invite you to reread my post.
What I am merely pointing out is that if you separate male faculty behavior from female faculty behavior (rather than treating the faculty as one large group), the gender gap/bias in the way male faculty consider the female applicants (relative to equal male applicants) is smaller than the way female faculty consider the female applicants (relative to equal male applicants). In other words, there is a greater disparity between how a female faculty judges a female applicant versus a male applicant (at least for the metrics of hirability and salary) and the way a male applicant judges the very same applicants. Look at the data yourself. The female faculty are: 1. harsher on the female applicants than the male faculty…2. kinder to the male applicants than the male faculty…and 3. offer the female applicants less money than the male faculty (8% difference, versus 4% difference).
This isn’t creating any strawmen. I am looking at the very same data everyone else is. Yes, it is still true that both faculty genders still give the male applicant higher score across the board. I don’t deny that. The gender gap is real. But it is also impossible to deny that relative gender gap is smaller for male faculty than female faculty (again, at least when hirability and salary are concerned). If there is some “constant” or “standard” bias against female applicants, one might expect male and female faculty to prefer males over females to the same extent. But they do not. The extent to which male faculty are biased towards male applicants (over female applicants) appears to be smaller than the extent to which female faculty are biased to male applicants (over female applicants). Again, this is most obvious in the metric of hireability and salary. This has been noted before in this thread (#19 Fergal) but by putting numbers to it, I thought I might rekindle that point.
So I merely want to know why? Why might women, whom one might expect to be more “sympathetic” or “understanding” (or use your own adjective here) to the inherent barriers facing women, contribute more to such barriers? I am not insinuating anything. I am merely asking for someone more qualified than me, who might be able to speak intelligently (perhaps because they work in this field) as to why this might be the case. Along those lines, I’m suggesting that issues of gender bias, while very much real, are more complicated than the typical patriarchal-stereotype arguments.
But clearly I won’t get that sort of insight from you. I invite you to check out post #94. At least TeaHag has an interesting idea to explain the salary phenomenon, suggesting that women’s experience with receiving lower salaries might precondition them to offer lower salaries in turn. But if that’s the case, one might expect a female to discount the male salary by the same amount…yet female faculty offer male applicants only ~4% less money than male faculty, but they offer female applicants ~8% less than male faculty.
Cheers.
@Corinne Moss-Racusin @Meowmeows
I think it’s an important study and I don’t doubt the results. I was merely suggesting that it was not an optimal representation. When axes cross, it looks like a “0” point. When bars are used for plotting, it is natural to compare the lengths. These features could be improved upon.
I thought I was commenting on a less contentious topic than the gender issue itself, but apparently not! 🙂
BTW, “I am a researcher at Caltech” gives out an “argument from authority” vibe.