Men, Women, and Graduation Statistics

There was a great big New York Times article on women in science this week, which prompted no end of discussion. (I also highly recommend Bee’s response at Backreaction.) It’s built around the personal story of the author, Eileen Pollack, a physics major at Yale who decided not to go to grad school, and her story is compellingly told, providing a nice frame to her investigation of the question of why there continue to be so few women in the sciences.

Pollack comes out very much in favor of the notion that many women choose not to go to graduate school in the sciences because they don’t receive sufficient encouragement, writing that “The most powerful determinant of whether a woman goes on in science might be whether anyone encourages her to go on.” She also quotes Yale prof Meg Urry agreeing with this, saying “Women need more positive reinforcement, and men need more negative reinforcement. Men wildly overestimate their learning abilities, their earning abilities.”

This explanation certainly fits with her personal story, but there are two numbers that make me think this might be a little too pat. Those are, from the American Institute of Physics statistics division, the percentage of bachelor’s degrees in physics awarded to women, which has been consistently in the 20-23% range since about 2000, and the percentage of Ph.D.’s in physics awarded to women, which is… right around 20% (the last specific number I saw in quickly scanning the AIP stats was 19%, but that was in 2008).

Those numbers are more or less the same– the uncertainty in them is probably a couple of percent. And those suggest that women go on to graduate school in more or less the same proportion as men do. This is further supported by their “Focus on First Year Graduate Students” report, which gives the average of the 2007 and 2009 entering classes as 20% female, during years when the percentage of bachelors degrees awarded to women was around 21%. And once in grad school, they go on to complete the Ph.D. in about the same proportion as men do. (And as previously noted are if anything very slightly overrepresented in faculty ranks.)

Now, it might be that this is a recent development– Pollack’s personal story dates from 1978, after all, and there have been numerous efforts to improve support systems for women over the intervening years. Maybe the support Pollack was lacking is now there, giving women the encouragement they need to go on to graduate study. You wouldn’t get that from the article, though.

It’s also possible that total graduation rate statistics are just too coarse to catch the real problem. You could maybe come up with an explanation involving overly confident men going off to make shitloads of money in finance, while the second-rate male students go on to grad school, or something. Again, though, that’s not something you get from the article, nor most of the online discussion I’ve seen.

Going from the available numbers, though, I am skeptical that the “not enough encouragement” explanation is the complete story, at least at the college/ grad school level. I suspect that the confidence levels aren’t quite as mismatched as the article would have it, and there are more timid men and bold women out there than are accounted for in the anecdotes provided.

Please note that I’m not trying to gainsay anybody’s personal anecdotes or claim that there is absolutely nothing wrong. Some of the stories that are reported in the times piece are absolutely appalling, and should’ve cost people jobs. I have no reason or desire to deny the validity of the experiences of those women.

But at the same time, the numbers suggest that even the sexist culture suggested by those anecdotes can’t be the whole story. For every woman who decides not to continue on to grad school or leaves grad school before the Ph.D., four men must also be leaving, for reasons that presumably don’t have to do with gender bias directed at them. And it’s hard to imagine that women are immune to whatever’s pushing the men out (I suspect that a lot of sexist bullshit is actually part of a more general sociopathy that makes life suck for everyone).

And, as always, the most important take-home is that this problem starts early. Which was driven home again recently– SteelyKid has said for a long time that she wants to be a scientist, but at one point this week said “I’m not going to be a scientist. Girls can’t be scientists.” I jumped on that as quickly and as hard as I could (and as she tends to do, she tried to pass it off as a joke– “I was just making that up”), but it’s incredibly depressing to realize that we already need to deal with this nonsense. I’m trying to think of an excuse to bring her over to campus and take her around to meet the many female scientists and engineers I work with at Union…

17 thoughts on “Men, Women, and Graduation Statistics

  1. Thanks for the link 🙂 Good you throw some numbers into the discussion.

    Everybody I know who left the academic track in physics (after finishing their masters or equivalent) did so primarily because they didn’t like the present and future options (or absence thereof). One the one hand, it’s the energetic, creative and independent people who decide to leave because they see that they can’t do what they want anyway and will have to jump between short-term contract for years before they might, if lucky, get a decent job with responsibility. That’s unfortunate and bad for the field generally. On the other hand there’s those who leave because they want stability in their job situation and that tends to be more important for women than for men at the same age. At least that’s been my impression.

    It’s very hard for me to tell how much ‘general sociopathy’ played a role in these decisions, but is it really so bad? I mean, look, there are assholes everywhere, what about that is so specific to academic science? In other words: where’s the control group for the complaints? I’m wondering if not all this complaining just creates a self-fulfilling prophecy. The more we complain about alleged sociopaths, the more attractive it becomes to the sociopaths and the less attractive to the rest.

    If you can’t take your daughter along, maybe tell her some stories of famous female scientists?

  2. The 20 % to 20 % equivalence is pretty poor analysis. You are assuming the only difference in the male/female populations is gender. The fact only 20 % of the undergraduate majors are female suggests a selection is going on at the early undergraduate level. If there is a selection than likely the male & female populations are not equivalent in terms of make-up.

    If you control for other factors would you predict 20 % of the graduate level would be made up of women?

    Maybe you would predict 10 % would be women in which case 20 % actually enrolled suggests at the undergraduate level you are doing very well encouraging women into graduate school, and should be celebrated. If the prediction was 30 % than it is problematic. If 20 % is what you would predict than the focus really needs to be on the climate leading up to choosing an undergraduate major.

    Being superficial in your analysis isn’t productive.

  3. The bulk of the selection occurs pre-college, from what I can tell. The eventual 20% graduation figure is not hugely reduced from the fraction of women we see in the introductory calculus-based physics course that all engineering and physics majors take. My vague recollection is that those tend to run around 30% women; I couldn’t find a number for that on the AIP site, though, and didn’t want to spend the time needed to get a figure from old class rosters.

    I agree that there’s a selection process taking place to winnow that number down, but one of the points that I think gets lost in these discussions is that there’s also a selection process for the men. The number of students who get an undergrad degree in physics is a tiny fraction of the number of students who take introductory physics, so neither gender of graduates is a random or representative sample of the general population.

  4. Obviously there is a selection issue pre-college and in college, and their is a selection going into graduate school for everyone.

    The discussion at hand is whether gender bias is contributing to the selection. You are downplaying that possibility in physics at the level of going from undergraduate to graduate school.

    The equivalence (20 % undergrad, 20 % grad school) doesn’t provide the evidence you are suggesting. It requires better controls and in depth analysis. Is the undergraduate female physics major population really the same as the undergraduate male physics population?

    The story told in the NY Times article was to highlight actually research, to humanize it for a wider audience. The research showed having a male name was an advantage in terms of perceived scientific competency by science faculty (physics, chemistry, and biology) . Women faculty were just as guilty as male faculty.

    It fits with other studies that suggest we have a bias towards men in science and begs for a more scientific analyses of what is happening at all levels in our society regarding the dearth of women in many of the natural sciences.

  5. I guess I’m not sure what you’re after that isn’t covered by “graduation statistics are just too coarse to catch the real problem.” As I’m about to be Internet-less for 8-12 hours, minimum, I’ll have to leave it at that.

  6. As a physics undergraduate from ’98 to ’03 my class was about 90% male, 10% female. One of the things I noticed (warning! anecdotal evidence…) was that the female students all were above average students. I concluded that something in the high school period must be discouraging the girls with just average physics potential from signing up for a physics study. Asking around among female friends, in both physics and other exact sciences, as to what this discouragement might be I mainly ran into two explanations:

    1) Girls don’t want to be “that one girl” in an all male field of study, especially not when male physics students have a reputation for being socially awkward nerds (to put it mildly.) A bit of a conundrum, we need more women to get more women.

    2) Discouragement from teachers, especially the older generation. While the people with high talent in physics were encouraged, regardless of whether they were male or female, girls who were just ‘okay’ in physics were more likely to get an “are you sure?” response than their male counterparts when they announced their choice of study. As this is done mostly subconsciously, some awareness training of new teachers will go a long way in resolving this I think.

  7. While I agree that the question of encouragement probably isn’t a key factor in going on to graduate school, what we really ought to consider is the % women who received their bachelor’s from US institutions in the incoming PhD classes. At the graduate level, you do have an influx of students who did their undergraduate work internationally.

  8. I am going to be really hated for this post, but it has to be said.

    I see this discussion all the time and something I really can’t quite understand is why this should be such a big problem. Ultimately, what matters is that the knowledge of humanity as a whole increases. Whether that is accomplished by an all-male, all-female, or 50-50 evenly split scientific workforce is entirely irrelevant. That women don’t become mathematicians and physicsts would only be a problem if those who don’t would have improved the average quality of researchers. There is no evidence to suggest this is the case and I am not saying this from a sexist perspective, I am saying it because the reality of the situation currently is that there is a severe shortage of research positions relative to the number of people qualified to fill them, i.e. there are already a lot more highly qualified males alone than the system can take.

    The bulk of the selection against women happens at such an early age that it is not really the case that their dreams of becoming scientists are shattered by the hostile sexist environment. In most cases they never really develop such dreams to begin with. Those who do face no worse odds than males in making it, as the 20% bachelor, 20% PhD graduation observation clearly shows. If anything, I have heard a lot of stories of math and physics departments hiring substandard female candidates just so that they can demonstrate how serious they are about eliminating gender bias.

    The constant discussion of gender quality is in fact adding to the problem because all it accomplishes is to successfully drill into everyone’s heads (both male and female, with the corresponding consequences for each) that there is a difference between the two sexes. When I was growing up, a lot of the best students in school were female, and I never really had any sense of an intellectual difference. I was only exposed to that idea by the very people fighting against sexual discrimination. If we want this to become a non-issue, the only way to do that is to eliminate the idea that there is a difference between the intellectual abilities of the two sexes and this will never happen if it is constantly thrown in people’s faces by the crusade to achieve 50/50 representation in science.

    And, of course, there is the simple fact that all else equal, unless she chooses to remain childless, a woman is going to produce less science than a man. It’s just the reality of human biology. All else equal, the main determinant of scientific productivity is how much time and energy you dedicate to work, and there is a very serious asymmetry in the time and resource commitment to raising the offspring of the two sexes is humans, as a result males will always be able to dedicate significantly more of those to science than females (unless the females forgo child bearing)

  9. I (female biochemist) teach AP Chemistry at a high needs high school. We are trying to create a pipeline of AP courses for our strong science students – Chem, Bio, Physics. My class generally has slightly more girls. There is an equal distribution of male and female students who score 4 or 5’s (though a relatively small sample size so Anecdotal evidence alert.) All boys who scored in the highest range and most of the 3’s and 2’s went on to AP Physics. The highest scoring girls did NOT go on to AP physics choosing instead to take 4 other APs – government, calculus, English, psychology in spite of our active and aggressive recruitment. We have a very successful and popular physics program taught by an excellent female teacher. These very bright girls are good at all subjects and profess to be more interested in politics and English. They know they would do well in physics but are more interested in other topics so go there – they just have more options since they are strong in all subjects. The boys who scored well in chemistry scored less well on their history and English junior AP courses and therefore profess more interest in math and science. This seems to be a mechanism of self selection out of science for many of our brightest girls.

  10. The real issue is that physicists are overpaid relative to biologists. There’s a feed-forward loop between salary and male-dominated aspects of a field. So if we just pay biologists more, more men will abandon the boring ship physics for the sexy ship biology, and some women who were snubbed for ‘merely average’ physics ability at the high school level will persist in physics (and the 20% of female physicists in grad school won’t get so fatigued by the asshattery of grad school that they don’t even bother to try for the research university professorship track).

    GM- what the heck is wrong with your scientific ability that you would conclude that “from lots of stories” you can tell anything about the reality of the situation??? Why would men and women’s abilities be PERCEIVED equally when we’ve seen resume studies that demonstrate that a mere name results in a competency hit? Why should you (or, in fairness, I) be exempt from these perceptual psychological glitches that are endemic in the society we are in? Why should the people who are telling you these stories be exempt from those glitches? And even if the stories are 100% accurate, why should we assume anybody will *remember* all the stories about sub-par male scientists, or view them as sexism instead of other forms of favoritism (neopotism, clanism about shared identity, ect)?

    And, of course, there is the simple fact that all else equal, even if she chooses to have children, a woman will live longer than a man, and thus will produce more science. It’s just a reality of human biology, that extended lifespan more than makes up for a couple years out of commission here and there for spawning.
    There is a fair question as to whether “two years at this point in your career” is as productive as “two years at that point in your career” and why physicists, and mathematicians even more so, skew to pedophilia in their fields. I don’t have an answer to that, but I suspect it is a direct effect of their moral failure to be biologists, who do not peak so young, but who are more productive throughout their lives.

  11. @GM: Chad works at a liberal arts college. Teaching is a far larger part of his job than research. I seriously doubt he agrees that what matters most is that the knowledge of humanity as a whole is increased.

    Okay – let me try to explain what I think PonderingFool meant.

    This is only anecdotal and also highly colored by my own perception, but when I was a first year grad student (in math), I noticed that, on average, the female grad students were far more highly motivated. There were quite a few male grad students (myself included) who ended up in grad school because they were good at math and did not have a clear idea what else to do with their life. On the other hand, all the female grad students had consciously decided they really wanted to do math as a career. Apparently, all the female students who were not highly motivated had already been discouraged.

    If entering motivation is an important factor in whether a graduate student finishes a Ph.D. or not (and, judging from the male students, it seems to be), then one would expect that female students would be more likely to complete a Ph.D. than male students.

    If female grad students complete their degrees at the same rate as male students despite their entering characteristics predicting a higher rate of success, then it seems likely there is some gender-specific factor at play.

  12. @quasihumanist

    Neither you nor Becca really addresses what I said.

    Of course I fully agree with the position that nobody should ever be denied the opportunity to do science based on anything other than their intellectual ability and motivation to work.

    However, from that and the fact that we have so few women in science it does not all follow that we have a problem. We have great gender imbalance in the military, nobody complains. We have great gender imbalance among nurses, nobody complains either. We have no male rhythmic gymnasts or synchronous swimmers at the Olympics, yet we added female boxing, wrestling and weight-lifting, and nobody complains about that either (more on that later).

    Again, this is only a problem if the level of science as a whole suffers as a result. And, again, there is no evidence to suggest this is the case.

    The fact is that we are already turning down most people that are capable of doing science at a high level without any regard for their gender. It’s a handful of top universities when one really has the opportunity (in terms of available collaborators, start-up funding, general intellectual environment, quality of graduate students, etc.) to fully blossom scientifically and implement one’s intellectual program. There is a very very small group of people who are at such a high level they have a close to 100% chance of ending up at those, but for most who could successfully fill up those positions, it’s a completely stochastic sampling based on things like who got lucky enough to be assigned a project that would turn into multiple C/N/S papers, who has the academic pedigree to push him through the hiring process, and other factors that don’t necessarily have much to do with one’s scientific ability. With the end result being that most capable people are left out of the system, and their number is significantly larger than the number of women who entered science as undergraduates, had the capability but eventually dropped out because they didn’t have enough support.

    The great filter happens much much earlier than the point of entering college (which I think pretty much everyone agrees on), and the people who constitute that filter have in most cases absolutely nothing to do with science. After that the reality is that the number of highly scientifically capable males is considerably larger than that of females, and is therefore counterproductive to try to fix the gender imbalance at the postgraduate level. If you do that, you will produce worse science than if you completely exclude all females (and, once again, I am by no means suggesting doing that). It has to be addressed very early on (especially given that by far the greatest predictor of how far you will develop your mathematical and scientific abilities is how early in life you start learning the subjects) and the best, and pretty much only, way to do that is to eliminate the idea that there are intellectual differences between the sexes. Which kids do not naturally have and which they (from both sexes) very successfully develop by learning from adults. But the idea will never disappear as long as we keep fretting about gender imbalance.

    To address what Becca said regarding productivity. Yes, women live longer than men on average in the general population. But is that also true among academics who generally live healthier and safer lifestyles than the general population? I don’t know the answer but I would be curious to see. And even if it’s the same difference, if’s largely irrelevant – it’s a 3 year difference, but it’s the difference between 76 and 79 or something like that. for the general population (and I would expect the numbers increase among academics). At that age most people are retired and/or have begun to develop some sort of neurodegenerative disease. Also, there is no evidence to suggest that even if all else is equal women would be able to work into their 80s. I don’t really know of female scientists who were productive at that age, I know of plenty of males, and yes, I am aware of the fact that those generation didn’t have many women scientists to begin with – it will be interesting to see what the situation is 20 years from now. Meanwhile what I said about child bearing in female physiology is very real, and happens in the most productive years of one’s life, which also impact what happens after that (because you are not only doing science, but actively learning science when you’re around 30). From this it follows that if one claims that there should be a 50/50 gender balance at the top of science, those women should be allowed to have at least 2 kids and spend all the time with them necessary for properly raising them, and anything else is bad for science, she is actually guilty of reverse gender discrimination because the only way that could be a proper reflection of reality is if women were so much smarter than men that that difference is more than enough to compensate all the time and energy expended on the reproduction and raising the offspring. That’s a lot like what happened in the Olympics – all sports and disciplines that were male-only in the past, now have female versions, but there is still not male rhythmic gymnastics and synchronous swimming. Is that fair?

  13. @GM – Only the first paragraph of my comment was meant for you. My only point is that we train scientists for purposes other than advancing science, and for many of those purposes, having a gender balance is helpful.

  14. quasihumanist is getting at the problem with your analysis.

    The women who are graduating in Physics at the undergrad level maybe enriched for the type of students who are more likely to go to graduate school to begin with. If that is the case then 20 % undergrad leading to 20 % earning PhDs is actually pretty poor.

    Basically the right controls haven’t been done for you to state what you have with the level of certainty that you have. Yes 20 % undergrad and 20 % PhD but what that means is unclear.

  15. I wasn’t trying to make a definitive statement of anything, just arguing that the numbers cast doubt on the simplest interpretation of the story. If I sounded unduly certain about some particular conclusion, that’s my fault; I was a little distracted by the impending TED@NYC trip, and might not have been careful enough in my wording.

    I tried to pull some numbers together that might speak to this, on a local level, and have been waffling about whether to post them. Ultimately, I’ve decided that the sample sizes we have are so small that the uncertainties render any comparison pretty well meaningless. They’re barely “plural anecdotes” let alone “data.” And, of course, there are all sorts of ethical issues weighing against sharing the information that would be most useful.

    So I’ll just say that while I agree that there could be other factors about motivation and so on coming into play, my local anecdotal impression doesn’t really fit with that. But it’s absolutely a question deserving more investigation by somebody with bigger datasets and the statistical chops to tease out the relevant information.

  16. One of the confounding factors in this analysis is that entering grad students in physics are not the same population as graduating seniors in physics. While some of the seniors may be foreign students, ISTM that a significantly larger fraction of the entering grad students are foreign. (By “foreign” I mean students who are not US citizens or permanent residents–these students are ineligible for most forms of undergraduate financial aid, but in grad school the only condition on aid is their ability to get a visa, which they need anyway in order to enroll.) So to do the comparison right, one would have to factor out the foreign students, who might have a different gender ratio from the US citizen/permanent resident population. I don’t know which direction that would skew things.

Comments are closed.