Solomon Hsiang sends along this from Corinne Moss-Racusin, John Dovidio, Victoria Brescoll, Mark Graham, and Jo Handelsman:
Despite efforts to recruit and retain more women, a stark gender disparity persists within academic science. . . . In a randomized double-blind study . . . science faculty from research-intensive universities rated the application materials of a student—who was randomly assigned either a male or female name—for a laboratory manager position. Faculty participants rated the male applicant as significantly more competent and hireable than the (identical) female applicant. These participants also selected a higher starting salary and offered more career mentoring to the male applicant. . . .
I hate to talk about things like this since presumably I’m a beneficiary. But now that I’ve climbed the ladder myself I suppose I’m not at any risk. I don’t know anything much about lab manager positions—that’s more something you’d see in a biology department—but I do know that I’ve hired more men than women as postdocs. If I were forced to hire in equal numbers it would be annoying but I suppose I could do it.
If you’re hiring post-docs from a pool of mostly male post-docs, it seems reasonable that most are men. I would hate to know I was hired to fill some “female scientist” quota and not based on my skills. I think the solution needs to be more upstream – getting children (boys and girls) excited about science throughout school and undergraduate degrees.
Or do both! They’re complementary: if more girls get excited about STEM, the pool of post-docs and faculty candidates will eventually be more gender-balanced; conversely, if there are more female role models in STEM, that makes it easier to get girls excited about STEM.
But the point here is that even with equal qualifications women are still being passed over, and, when they are selected, they are being compensated poorly relative to equally qualified males. What we need is the explicit recognition that all academics view women as less qualified, even when they are equally qualified. Adding more women to the pool of applicants isn’t neccessarily going to change that. It requires consciousness of the fact that people simply view the women as less qualified, for whatever reason, and that they tend to undervalue their service relative to men.
Another possibility: if the evaluators are mostly male, then we have the problem that people similar to you (along multiple dimensions) are regarded as more competent.
I’m reminded of the rumors that the graduate program I was in favored ectomorphs (tall, thin body type). And in the grad intramural volleyball league, there were 6 teams from our program and only 0,1,2 from other areas. The person who had done research in this area was (surprise, surprise) an ectomorph.
I see from the abstract that this bias was true of BOTH male and female faculty members. There are at least two ways to interpret this. ONE way is to note that tenured female faculty may have gone through a lot of socialization which leads to their being more “male” in attitude. (I think there’s some research supporting this, but can’t easily find the cite).
This is political slight of hand. Look over here at this bright shiny other issue (getting young girls excited about science). Sure, but lets not be distracted from the issue at hand.
Which should we rather be hired for: an unconscious bias towards ourselves over an equal/better candidate or an explicit attempt to correct an unconscious bias towards a equal/worse candidate.
That is a really good point – I didn’t mean that encouraging better representation of both sexes in science would cure the sexism. I think that forcing employers to hire equal numbers of men and women out of a pool of mostly men is unreasonable and might create more problems than it fixes. But how to actually address the unconscious bias…you’re right, I have no idea where to even start. Reading things like this at least increases my awareness.
What about forcing hiring proportional to representation in the pool? That way you ensure female representation, but you aren’t forcing unreasonable expectations on the employer, as long as the hiring pool is large enough.
What’s the sex ratio of among the applicants for your post-doc positions? In the corporate workd, I’ve been involved in the hire of a lot of consumer research scientists and technicians, and both the applicants and hires are dominated by women. ( A psych degree is a pre-req for the scientists) The technicians are all women, as extreme tactile sensitivity is a prerequisite. Among the statisticans, there are somewhat more men than women, but the ratio is not much different than the pool of applicants. Among our biologists, most of the recent hires have been women. Coincidentally, woman make up more than half of the recent biology Ph.D.s.
I think quotas are a bad solution to hiring bias, if it exists. And other factors like abilities, interests, and desire to work come into play too.
Years ago, one read about how women enrolled in colleges at a lower rate than men, because of discrimination. Now, women far outnumber men and the gap is increasing, but no one really mentions it or asks for money, training programs, or laws to solve the problem. (In private colleges, women outnumber men 60-40, and that despite men of college age outnumbering women 51-49.) Now we read how women are discriminated against in the sciences. No mention of how girls have better communication/language/writing skills than boys — a gap that continues into High School and may be reflected in boys’ lower High School graduation rates..
I wish we’d see gender and education addressed in an even-handed manner. There are severe issues on both sides of the aisle, as it were.
Two previous high-profile papers with much larger sample sizes (N>1000 in both, vs N<130 in this PNAS study) found slight discrimination against MALES (Bertrand & Mullainathan, 2004; Milkman, Akinola, & Chugh, 2012); the latter involved 6000+ professors as subjects.
Neither study is cited in the paper (BM2004 is cited in the supplement).
Much less important, but relevant for this blog, the PNAS article includes a gem of a quote:
"it is important to note that we obtained the necessary power an representativeness to generalize from our results while purposefully avoiding an unnecessarily large sample that could have biased our results toward a false-positive type I error (48)."
It caught my eye because (48) is a paper we wrote (Simmons et al, 2011) and we certainly did not make the fallacious claim that large samples have higher type-I errors! What gives?
Bertrand, M., & Mullainathan, S. (2004). Are Emily and Greg More Employable Than Lakisha and Jamal? A Field Experiment on Labor Market Discrimination. The American Economic Review,
94(4), 991-1013.
Milkman, K. L., Akinola, M., & Chugh, D. (2012). Temporal Distance and Discrimination an Audit Study in Academia. Psychological Science, 23(7), 710-717.
Simmons, J. P., Nelson, L. D., & Simonsohn, U. (2011). False-Positive Psychology: Undisclosed Flexibility in Data Collection and Analysis Allows Presenting Anything as Significant. Psychological Science, 22(11), 1359-1366.
I don’t understand this comment Uri.
The Milkman paper shows a similar (maybe more nuanced) finding than the one AG refers to? Can’t see how one can argue that this study shows discrimation against males, cf. the quote below:
“Through a field experiment set in academia (with a sample of 6,548 professors), we found that decisions about distant-future events were more likely to generate discrimination against women and minorities (relative to Caucasian males) than were decisions about near-future events.”
I think the wording and graphs can be a bit confusing in that paper because of the fact that white-males are the omitted category. If one compares just gender overall, or gender by race it becomes easier to notice that there is a general, if small, preference for females.
For instance, of requests for meeting today, about 70% of females get a response, compared to about 66% of males. Similarly, about 41% of females get a meeting accepted compared to about 36% of males, so there is a pro-females bias of about 4%-5% percentage points.
The discrimination against males is attenuated for future meetings where females enjoy just a +-1% percentage point *higher* response rate.
Broken down by ethnic group: White, Black, Indian and Hispanic women do slightly better than their male counterparts, only among Asians do women fare worse than men.
I don’t think that Milkman paper is fully relevant here. And some of its findings are so puzzling (e.g. why is it that the group with highest “today” response and acceptance ratings is African American females?) that one has to wonder if there’s more going on than meets the eye.
Bertrand paper is from a totally different area, it looked at help-wanted ads “in the sales, administrative support, clerical and customer services job categories” (naturally female-dominated fields).
I also noted the quote Uri copied above, and puzzled about it and the reference to Simmons et al (2011). It does become a little less fuzzy (but admittedly still fuzzy) if you include the preceding paragraph:
“Additionally, in keeping with recommended practices, we conducted an a priori power analysis before beginning data collection to determine the optimal sample size needed to detect effects without biasing results toward obtaining significance (SI Materials and Methods: Subjects and Recruitment Strategy) (48). Thus, although our sample size may appear small to some readers, it is important to note that we obtained the necessary power and representativeness to generalize from our results while purposefully avoiding an unnecessarily large sample that could have biased our results toward a false-positive type I error (48).”
However, I have a gripe about both the Moss-Racusin et al and the Simmons et al papers — namely, the neglect/dismissal of considering family-wise Type I error rate. I strongly disagree with Simmons et al’s decision not to recommend accounting for multiple testing, particularly since they give a simulation example showing how multiple testing can inflate Type I error rate. Yes, there is no one good way to take multiple testing into account, but that’s no excuse to ignore the problem.
The Moss-Racusin et al paper used plain old ANOVA and t-tests (no shrinkage estimators), performing over sixty(!) hypothesis tests on a data set with 127 observational units. This definitely calls for some discussion of family-wise error rate. I tried a Holms procedure, but p-values were not given with enough precision in the paper to determine whether any of the tests were significant at a 0.05 FWER level (but it appears that only the nine listed with “p < 0.001" might be).
Unfortunately, the types of misunderstandings in Moss-Racusin et al are all too common (in fact, seem to be the norm) in much of the literature I've seen in psychology and education.
Sexism, or correct compensation for unobserved variables like being slightly autistic and willing to work stupid-hard for low pay?
Wouldn’t correct compensation for unobserved autism involve giving the applicant higher hireability ratings and lower salary at the same time?
Besides, the article reports that subjects were administered a standard sexism test and that scores on that test were significantly and inversely correlated with ratings those subjects were giving to the female applicant.
> Wouldn’t correct compensation for unobserved autism involve giving the applicant higher hireability ratings and lower salary at the same time?
Under some circumstances, not under others. For example, if this is a general phenomenon then you may want to offer higher salary to compete with other job offers, while women would only be worth hiring if you could get them below their discounted value.
> Besides, the article reports that subjects were administered a standard sexism test and that scores on that test were significantly and inversely correlated with ratings those subjects were giving to the female applicant.
Same objection applies. If you think women won’t work stupid-hard for you and will leave for greener pastures and dare to have families or recreational time, then perhaps you think less of those women. Think of East Asian complaints about ‘gold misses’ who are merely demanding their worth in the marriage market.
> who was randomly assigned either a male or female name
OK, this made me think of a solution; how about simply masking the applicant’s name out and conducting the entirety of the hiring process similarly to a blind study?
Naturally, at some point a connection would be made — verifying the publications listed in resume, personal recommendations (although these could be gender-neutralized), personal interview — however, by the time it gets to the interview stage, there should be enough information about the candidate to at least alleviate any biased preconceptions (since these, by definition, are formed prior to having enough information).
Thoughts?
You should check out the Equal Employment Opportunity Commission’s “Four-Fifth’s Rule” for determining what is suspicious.
http://www.uniformguidelines.com/uniformguidelines.html
“D. Adverse impact and the “four-fifths rule.”
“A selection rate for any race, sex, or ethnic group which is less than four-fifths (4/5) (or eighty percent) of the rate for the group with the highest rate will generally be regarded by the Federal enforcement agencies as evidence of adverse impact, while a greater than four-fifths rate will generally not be regarded by Federal enforcement agencies as evidence of adverse impact. Smaller differences in selection rate may nevertheless constitute adverse impact, where they are significant in both statistical and practical terms or where a user’s actions have discouraged applicants disproportionately on grounds of race, sex, or ethnic group. Greater differences in selection rate may not constitute adverse impact where the differences are based on small numbers and are not statistically significant, or where special recruiting or other programs cause the pool of minority or female candidates to be atypical of the normal pool of applicants from that group. Where the user’s evidence concerning the impact of a selection procedure indicates adverse impact but is based upon numbers which are too small to be reliable, evidence concerning the impact of the procedure over a longer period of time and/or evidence concerning the impact which the selection procedure had when used in the same manner in similar circumstances elsewhere may be considered in determining adverse impact. Where the user has not maintained data on adverse impact as required by the documentation section of applicable guidelines, the Federal enforcement agencies may draw an inference of adverse impact of the selection process from the failure of the user to maintain such data, if the user has an underutilization of a group in the job category, as compared to the group’s representation in the relevant labor market or, in the case of jobs filled from within, the applicable work force.”
Ugh! I guess they have to write it in that convoluted way to handle all the anticipated objections from lawyers…
If they wrote it in plain English, it would be easier to notice how crazy it is.
The point that interests me is that males were offered more career mentoring than females. That is one thing that I would have liked to have received more of.
Pingback: Somewhere else, part 27 | Freakonometrics
Does it occur to the writer and the syncophantic PC commentators that there are some logical evolutionary reasons most males have a greater ‘knack’ for math and science than do most females?
The ability of males to count, compute, follow maps, and evaluate other scientific phenomena was a great survival advantage in seeking food and killing off competitors while the women reproduced and tended the kids.
Government regulations and lawyers had nothing to do with it.
BigEd:
Good point. Back when my sister worked at McDonald’s, they had a rule that girls could do the fries but only boys could make the burgers. Back in the 70s, the bosses at McDonald’s hadn’t yet forgotten the ancient wisdom that grilling was inherently a male skill.
Andrew: I got a kick out of your response. Whenever I would tell Marines stories about academia, both male and female Marines would often express the opinion that all grad students/professors were females. While I wouldn’t claim as much myself, I could see were they were coming from. Academia and Science resembles child rearing far more than it does combat.
I think the reason males are found more often in math and the “hard” sciences of physics and chemistry (and related engineering disciplines) is due more to innate male aggressiveness than innate male ability. Math is hard — for everybody — and it involves a lot of abstruse reasoning and notation that take a big commitment to master. It also involves an enormous amount of competitiveness, which — ultimately — is about who gets to call himself (or herself) the smartest. A lot of people survey what is required to play the game and decide that it isn’t worth it, and they may very well be right based on an eminently rational set of career goals (i.e., achieving maximum compensation, recognition and efficacy at a minimum of effort). Many people — men and women — have the ability to do well if they would apply themselves, but they become discouraged and convince themselves that they “can’t do math.” It is only those who are really driven to succeed — people who refuse to believe that there might be something out there which they cannot understand, for whom the competitiveness of math and science are motivation enough to spend an inordinate amount of time on difficult and often mundane topics without immediate and clear benefits for society or one’s career — who do well enough to become professionals. And so, although there is absolutely nothing preventing more women from becoming great mathematicians and scientists (and I know many women who are those things), human nature being what it is, it is more often men who answer the call, for reasons that have more to do with testosterone than with smarts. It turns out that your Marine buddies are not actually correct: academia — especially math and science — is really more like combat than it is like child rearing. It is mastered by those who, for better or worse, lack the sense to call it quits when all the evidence says that they should.
So mothers facing death in child birth, plagues, famine, brutal males, and high infant mortality wouldn’t have needed perseverance in the face of hopelessness? And furthermore this quality wouldn’t have been strongly selected by evolution?
That’s not what I said. Persevering through childbirth seems like a very reasonable thing to do — through probability theory or real analysis, perhaps not so much.
Look, for better and worse, men are more aggressive — in a “damn the torpedoes” sense — on average, than women. I think that’s pretty uncontroversial. It’s part of the reason why there are more male entrepreneurs and more male violence. I am suggesting that it is perhaps also part of the reason why there are more male physical scientists. Women now outnumber men on college campuses by a factor of almost 2:1, but men still outnumber women in math and the physical sciences even at the undergraduate level. There are plenty of women who are capable of doing the work. Many of them do and go on to perform at the highest levels, but many of them choose to do something else. Why is that? I don’t argue that cultural factors aren’t part of the story, but I think we’re kidding ourselves if we don’t admit that men and women are — on average — different in some ways, and that can lead to differences — on average — in their interests. The idea, posited above, that men are innately more capable than women in math is pernicious and at the very least not proven, but its belief is widespread. It’s time to look for a different explanation.
To be absolutely clear, I think that there are many bold women and prudent men, and that everyone should be considered as an individual. But what we are talking about here is why there are more men than women in physical science. If I were forced to hire equal numbers of women and men as postdocs, I’d be more than a little bit annoyed. There are in fact already significant incentives for hiring women, and I support those as a reasonable response to lingering cultural understandings of math and science as a man’s endeavor and to the discrimination which undoubtedly still exists. But I do not support ‘Title IX’ in STEM fields; I think that would be deleterious for the affected disciplines and patently unfair to men.
Wow, that’s the most tepid response to institutional sexism I think I’ve ever read. Heaven forbid you should inconvenience yourself trying to make the world a more just and equal place. . .
Pingback: [research] Sexual discrimination in science: why we must act now | feimineach.com