Yay! Sexism in science is over!

The New York Times has declared that Academic Science Isn’t Sexist. What a relief! The authors are reporting the results of a broad study of many different parameters of the career pipeline, and are happy to report that there are no problems in academia. None at all, no sir.

Our analysis reveals that the experiences of young and midcareer women in math-intensive fields are, for the most part, similar to those of their male counterparts: They are more likely to receive hiring offers, are paid roughly the same (in 14 of 16 comparisons across the eight fields), are generally tenured and promoted at the same rate (except in economics), remain in their fields at roughly the same rate, have their grants funded and articles accepted as often and are about as satisfied with their jobs. Articles published by women are cited as often as those by men. In sum, with a few exceptions, the world of academic science in math-based fields today reflects gender fairness, rather than gender bias.

Yay! But wait…I looked at their paper, and it was weird. Despite total equality, somehow there are still fewer women entering the academic workforce, still fewer getting tenure, women are getting fewer publications, etc. It’s right there in the data they present. Somehow it all vanishes in the analysis, though. So I look at the data, I look at their interpretations, and there’s some magic that goes on between the two that makes the differences disappear. I’m not sure how — it’s a long paper and it gallops all over the place, so it’s going to take more scrutiny than I want to spend on it to figure out how they’re doing that magic trick. But here’s a clue, one among many:

The cause of this is not that women applicants are not being hired, but rather that they are choosing to opt out of academic science. [my emphasis]

Back to the NY Times article: we see the same sleight of hand.

Our analyses show that women can and do prosper in math-based fields of science, if they choose to enter these fields in the first place.

See? It doesn’t count against academia if the women choose to leave, rather than not being hired or, I don’t know, forced out at gunpoint. So, for instance, if women were facing a hostile work environment in academia, or are discouraged by being told or seeing that they’ll have to work much harder than their male colleagues to succeed, or learn that they have to make a choice between family and work (a choice the men rarely face), and just decide “screw it, this isn’t worth it”, that’s not a problem! Well, in this paper it’s not a problem, it’s waved away. In the real world, it’s a problem.

Here’s another summary of why women leave academia from a completely different source that doesn’t dance away from the difficulties.

Men and women show radically different developments regarding their intended future careers. At the beginning of their studies, 72% of women express an intention to pursue careers as researchers, either in industry or academia. Among men, 61% express the same intention.

By the third year, the proportion of men planning careers in research had dropped from 61% to 59%. But for the women, the number had plummeted from 72% in the first year to 37% as they finish their studies.

If we tease apart those who want to work as researchers in industry from those who want to work as researchers in academia, the third year numbers are alarming: 12% of the women and 21% of the men see academia as their preferred choice.

So women find academia a far less pleasant prospect than men do, and abandon it in droves. But it’s by choice, so Williams and Ceci have an excuse to ignore this hemorrhage. And why do they find it less pleasant?

Women more than men see great sacrifice as a prerequisite for success in academia. This comes in part from their perception of women who have succeeded, from the nature of the available role models. Successful female professors are perceived by female PhD candidates as displaying masculine characteristics, such as aggression and competitiveness, and they were often childless.

As if all this were not enough, women PhD candidates had one experience that men never have. They were told that they would encounter problems along the way simply because they are women. They are told, in other words, that their gender will work against them.

But this is not a problem. Because they choose to leave. In the same way, I suppose we could argue that ISIS is not a problem in Iraq, because all those refugees chose to flee their homes.

But wait! You know this story could be worse. Williams and Ceci could make some argument built around intrinsic differences, you know, the old “boys play with trucks, girls play with dolls” stuff…uh. Oh. Uh-oh. Crap.

As children, girls tend to show more interest in living things (such as people and animals), while boys tend to prefer playing with machines and building things. As adolescents, girls express less interest in careers like engineering and computer science. Despite earning higher grades throughout schooling in all subjects — including math and science — girls are less likely to take math-intensive advanced-placement courses like calculus and physics.

We live in a culture with pervasive gender roles — we all learn early on that if we want to fit in (and as social animals, most of us do want to fit in), you adopt the roles that will make those around you happy, and often that means we can be personally happier that way. But let’s not pretend that these aren’t shaping women to fit into less well rewarded positions, or in many cases compelling them to abandon life choices to which they would be best suited (and likewise, this can afflict men as well). So here they even admit that girls perform better at science and math, yet somehow they end up not following through to enter science and math careers.

But this is not a problem. Because it is their choice.

What’s really ironic is that the conclusion of their paper is that they have swept away all the attempts to reduce the causes of the attrition of women in STEM to a single “culprit” — it’s multifactorial and complicated, they proudly announce! Yeah, we already knew that. I don’t know anyone who has tried to explain it away as a consequence of a single factor…except those people who try to argue that girls just don’t like machines and building things. Or blue things. Or stuff that requires active involvement. Don’t you all understand that men do things, women have things done to them?

You’re going to see a lot more about this paper in the near future — all of the gnathostomes are dropping their jaws at the denial of the obvious in that op-ed. Galileo's Pendulum sees a big omission.

The basis for dismissing sexism seems to be a small study of faculty hiring practices, comparing the percentage of male and female applicants who successfully landed academic physics positions. They didn’t look at retention — the problem that many assistant professors don’t achieve tenure or are slow to be otherwise promoted — and they seem to ignore all of the factors that decide whether women feel welcome in the profession. That seems to be a significant problem, not one that should be dismissed as “anecdotal”.

Emily Willingham has done a more thorough reading of the paper than I have — she points out that the data in their figures is elided in the text to an amazing degree.

Check out Figure 15. Go ahead. Just for fun. And scroll on down to Figure 16. Look at the salary values on Table 4. Look at Figure 18. See the job satisfaction results in Figure 19. Take a gander at Figure 5. Figure 4. I don’t understand how they wrote the paper or the op-ed they did while looking at the same results I see in their paper. Nothing about these data says, “OK, folks. Our work in the academy is done. Let’s focus on those kindergartners.”

And evidently, the implications weren’t manifest to them, either. Even as these authors say there’s no sexism in the science academy, they write:

… we actually found a greater exodus of women from non-math-intensive fields in which they are already well represented as professors (like psychology and biology, where 45 to 65 percent of new professors are women) than from fields in which they are underrepresented (like engineering, computer science and physics, where only 25 to 30 percent of new professors are women). Our analyses show that women can and do prosper in math-based fields of science, if they choose to enter these fields in the first place.

Emphasis mine. See, the problem here is that women don’t choose to enter these fields in the first place. But that’s not because academia’s not unwelcoming to them or anything.

Willingham seems to have had the same reaction to the paper I did — it’s bizarrely jarring, in that the data just scream “ACADEMIA HAS A PROBLEM HERE!”, while the text chortles happily and says “Academia has no problem at all here.”

Ceci SJ, Ginther DK, Kahn S, Williams WM (2014) Women in Academic Science: A Changing Landscape. Psychological Science in the Public Interest 15(3):75–141.


  1. Al Dente says

    It would appear that Ceci et al didn’t look at their own data before reaching their conclusion.

  2. nyarlathotep says

    I continue to find only one perpetual truth: there are lies, damn lies, and statistics.

  3. Rock Doc says

    There’s no sexism in science because women choose not to be scientists…But women choose not to be scientists because science is sexist.

    My head hurts.

  4. dreikin says

    I’m glad – but not particularly surprised – that math & computer science tends to be about even or favoring females in the stats I looked at in the article.

    However, I suspect that may be from math lifting up computer science. I wish they hadn’t grouped them together, and I wonder if they had ulterior motives for doing so, given the reputation the field has with regard to its treatment of women.

  5. dreikin says

    I should correct something: that’s for the stats not dealing with proportion female. I.e., there’s a lot less female than male, but those that are there seem to be paid the same and such. I should probably go through and tabulate that to check if my impression is correct..

  6. tuibguy says

    I took Developmental Psych from Ceci at the University of North Dakota the year before he moved to Cornell. His article makes me wonder if everything I learned from him about Piaget is wrong.

  7. microraptor says

    Is it just me, or do these kinds of reports sound more and more like something from a Paranoia campaign every year?

  8. fmitchell says

    This problem concerns some people in IT (information technology, not the Stephen King novel). Not enough women enter IT, and the many who leave cite crappy maternity leave policies, pay which doesn’t even cover day care costs, and comments that stays just under the threshold for sexual harassment.

    Others, e.g. many commenters on Slashdot, say “women are interested in different things” or even “women just don’t want to work in IT, why are you still talking about it?” At the risk of immodesty, here’s a summary of an article I submitted and the resulting grousing about feminists.

  9. dreikin says

    First off, the stuff I said I would look at above:

    Females in the group Math and Computer Science, on average and compared to males:

    Work more average hours per week (for Tenured and Tenure-Track Faculty with Children) [Fig. A3]

    Work more average hours per week (for Tenured and Tenure-Track Faculty) [Fig. 15]

    For All Universities: [Table A1]
    ├─ Have fewer publications
    ├─ Have more conference papers
    ├─ Have more patent applications
    └─ Have fewer patents granted

    For Research I Universities: [Table A1]
    ├─ Have more publications
    ├─ Have more conference papers
    ├─ Have more patent applications
    └─ Have fewer patents granted

    As of 2010, have about the same “Percentage of Tenure-Track and Tenured Faculty Reporting to Be Very or Somewhat Satisfied”, though still slightly less than males. Difference in 1997 is much greater. [Fig. 19]

    Have as a proportion of male salaries: (1.00 is parity, greater favors females, lesser favors males) [Fig. 17]
    ├─ Female Assistant Professors: Between 0.95 and 1.00 (and declining from 1995-2010)
    ├─ Female Associate Professors: Greater than (1995) or approximately equal to (2010) 1.00.
    └─ Female Full Professors: Less than 0.90 (1995) to greater than 1.00 (2010)

    For All Institutions: [Table 4]
    ├─ Make less as Assistant or Associate Professor,
    └─ Make more as Full Professor

    For Research I Institutions: [Table 4]
    ├─ Make more at all three levels of professor.
    └─ // This is probably what most lead to me make the previous statement; Math/Comp.Sci is the only field listed where this is true.

    Have slightly fewer publications over the previous 5 years as Assistant Professors both with and without children. [Fig. 16]
    ├─ Males are still slightly higher,
    └─ The difference is larger but still less than one for those with children

    Have fewer publications over the previous 5 years as Assistant Professors in 2008 and 1995; difference decreased. [Fig. 14]

    Have significantly more publications over the previous 5 years as Associate Professors in 2008, and fewer in 1995. [Fig. 14]

    Have slightly more publications over the previous 5 years as Full Professors in 2008, and fewer in 1995. [Fig. 14]

    Have only 2 out of 30 data points where the “Female disadvantage (female publications as percentage of male publications)” is greater than 50%. [Table 2]
    └─ // This one is not Math and Comp.Sci. specific

    So meh. Most of it is at or near parity, somewhat biased towards males, with one notable exception. And while I didn’t list most of them, all the percentage-involved stats effectively say “heavily male dominated”, so again, this is only for those persistent enough to actually get to an academic position, which is not many, and getting smaller [see below].

    Interesting additional stuff that I noticed:

    The graphs and tables are a mess. Some things show historical trends, and others you may want to compare them with don’t. Some things (table 1, I’m looking at you) don’t provide gender comparisons. There’s frequently a question of usefulness, relevance, and/or completeness. Several of the tables are much messier than necessary. It’s not always clear what is being demonstrated. Some related items are separated by others not closely related. It feels a lot like “Oh, pretty! let’s add that one too!”.

    Table 1 is not much use, since it can’t be compared to non-female stats (in the article, anyway)

    More patent applications, but fewer patents granted? Why? [Table A1]

    Some fields have had sharp drop-offs in percentage of PhDs awarded to females in recent years (starts around 2007-2008) [Fig. A1.b: Mathematics and Statistics, Computer Science] [Fig. 3.b: Math and Computer Science]
    ├─ While some other fields show similar size drop-offs over the same span, none of them appear as signficant – they appear like noise in a generally rising graph, but
    └─ This is the only field where it looks like a trend, following a plateau-like period. [same figures]

    Percentage of Bachelor’s degrees awarded to females has generally and obviously decreased since about 1984 for Computer Science [Fig. A1.a] and Math and Computer Science [Fig. 3.a]
    └─ This is the only field this appears to be true for. The closest one is economics, which is still (slightly) generally rising. [same figures]

    Table 5 seems discordant with the rest of the information presented, but the tables and figures are in such a disarray (as mentioned before) it’s a pain to pinpoint how. It may well accord with the other data, and if so it hints at where the biggest problems may lie. For example, for the GEEMP column things get better after the “College Major” row, hinting that the biggest issues occur before then. OTOH, it picks out economics for particular mistreatment in some places, but doesn’t make special note of the unusual decline in females for Math/Comp.Sci. graduates.

    Fig. 9: Interesting correlation, but the causation is as clear as onyx. There are just so many potential factors (including the same things that cause racial disparities in IQ tests) that could confound any easy explanations.

    * Females are heavily filtered out in Math/Comp.Sci., and it’s getting worse.
    * The females who make it through do end up treated about as well according to these stats.
    * The paper is a mess.

  10. chris61 says

    The title of the NYT article doesn’t seem to me to reflect the conclusions of the paper very well at all.

    In sum, depending on the life-course transition point, the cause of early lack of interest in GEEMP subjects and later attrition from GEEMP fields is the result of one or more of a confluence of variables. Attempts to reduce these causes to a single “culprit” (e.g., bias by search committees against female applicants; women’s preference for other fields or lack of math aptitude; publica tion rates; salary differentials) are not supported by the
    full corpus of data and research findings. Granted, one can cherry-pick aberrant examples that seem to suggest bias or aptitude gaps or differences between the sexes in productivity or impact, but the entire scientific corpus reveals that no single cause can account for the dearth of women in GEEMP careers. The most significant implication of our analysis is that failure to acknowledge the nature, complexity, and timing of causes limits progress in increasing women’s representation in math-intensive careers, by directing resources to areas that are not currently major reasons for the dearth of women in math- intensive fields.

    I don’t read that as saying there is no problem but rather as saying there are multiple problems that need to be addressed to increase women’s representation in science.

  11. biogeo says

    What a load of motivated reasoning.

    I think PZ probably blogged about this when it came out, but earlier this year there was a study published in Current Biology that built a descriptive model for the probability that a young investigator will some day manage to make it to principal investigator (i.e., the head of his or her own lab). Science Careers made a little calculator letting you plug in your vital statistics (number of publications, impact factor, etc.) One of the largest effects in their model, sadly, was gender. I’m male, and just finishing up grad school in neuroscience, and depending on how I count things, the calculator indicates that my probability of becoming a PI someday ranges between about 35-47%. If I were female but with the same qualifications, that would be 24-36%. So just by virtue of being male, I have a career boost of 11 percentage points over my equivalently credentialed female colleagues. Or put another way, a female colleague of mine will have to have published literally twice as many first-author papers as I have so far in order to have an equivalent chance of someday running her own lab. This is not a particularly pleasant realization for someone who wants to believe that any success he may have in his career will reflect his merits rather than his gender.

  12. zetopan says

    “It would appear that Ceci et al didn’t look at their own data before reaching their conclusion.”

    Of course not, you silly person. That takes up valuable time and besides it is totally redundant when you start with the conclusion. How could you possibly miss these two key points?

  13. We are Plethora says

    Professor Myers,
    How would you interpret this statement, which we found buried in the paper somewhere? Are they saying that women not only are choosing to “opt out” but that they are making that choice prior to college? And that’s perfectly fine and dandy according to them?

    Are we reading this correctly or have we misunderstood?

    We conclude by suggesting that although in the past, gender discrimination was an important cause of women’s underrepresentation in scientific academic careers, this claim has continued to be invoked after it has ceased being a valid cause of women’s underrepresentation in math-intensive fields. Consequently, current barriers to women’s full participation in mathematically intensive academic science fields are rooted in pre-college factors and the subsequent likelihood of majoring in these fields, and future research should focus on these barriers rather than misdirecting attention toward historical barriers that no longer account for women’s underrepresentation in academic science.

  14. auraboy says

    Whew, glad they got that sorted out then. And how about that progressive world view on display? It’s all about a woman’s right to choose! You ladies can choose to put up with inherent sexism or you can just opt out and go do something else instead! Solved! Next! To follow up we can finally put that whole racist law enforcement myth to bed by noting that it’s just black people being inherently more criminal and geared towards crimes that white police like to investigate…

  15. markd555 says

    As if all this were not enough, women PhD candidates had one experience that men never have. They were told that they would encounter problems along the way simply because they are women. They are told, in other words, that their gender will work against them.

    So friggin sick of this line of “If only women weren’t told about harassment, and scared by it with horror stories, then there wouldn’t be a problem!” Same crap going on in GamerGate right now: “Women are being scared out of game development because of those Feminists! Stop driving women away!” That’s so damn backwards I can’t even start.

    Cover your eyes, the bad man will probably go away!
    If anything happens, don’t talk about it! You will ruin everything! Might startle those other flighty wimmens.

  16. carlie says

    So friggin sick of this line of “If only women weren’t told about harassment, and scared by it with horror stories, then there wouldn’t be a problem!”

    I wonder how they explain how women didn’t exist in these fields back when we didn’t talk openly about harassment.

  17. garnetstar says

    Here are some of my experiences are a woman in academic science (all of the remarks were spontaneous on the speakers’ part, I’d never mentioned that I’d had any problems):

    A male colleague told me that I didn’t look pretty enough in photos, and couldn’t I fix myself up?

    My students asked me to publish papers (on which they were, of course, authors) with only a first intial instead of my obviously-female first name, so that the papers wouldn’t be disregarded or scorned.

    A reviewer wrote that I was “combative” when I answered his comments on my paper.

    My secretary said “They treat you like you’re Madonna”, meaning, as an exotic, different creature, not a colleague.

    My best student, upon my asking why she’d decided on a career in industry, replied “I’d never go into academics after I’ve seen how they treat you.”

    I rest my case.

  18. colnago80 says

    Re Al Dente @ #1

    Maybe they reached their conclusions before they did the research, like creationists do.

  19. Esteleth is Groot says

    It doesn’t even have to be that blatant. It can be questions (addressed only to women) about how you juggle kids/marriage and an academic life in STEM. It can be astonishment when, in reply to those questions, you indicate that you are single and have no children, coupled with comments about how they hope you don’t regret “throwing away” your mid-twenties “when the time comes.”

    It can be a lack of help and mentoring that your male counterparts are given as a matter of course. The lack occurring due to both a dearth of qualified women to mentor, and the men in the upper ranks not stepping in – both because it does’t occur to them, and because some hesitate because they are concerned over impressions of impropriety and/or snarky commentary when they take a young woman under their wing.

    It can be because you meet your soulmate when you’re both grad students or post-docs, and when the time comes to move on, you don’t get placements in close enough proximity for it to work, and it just makes more sense for her to drop out.

    I cannot say that any of these things happened to me. My problems, I think, had nothing to do with the fact that I’m a woman.

    It’s just that I find myself as another one of those data sets: I’m a woman. I have a Ph.D. in a STEM field. And I just sent out a batch of resumes to nursing jobs.

  20. sff9 says

    The cause of this is not that women applicants are not being hired, but rather that they are choosing to opt out of academic science.

    I find it telling that using this line of argument, the conclusion would be the same even if there were only one woman in academic science (it probably could even be argued that it would be the same if there were no woman at all, by vacuous truth). How it is possible to not realize this and write

    In sum, with a few exceptions, the world of academic science in math-based fields today reflects gender fairness, rather than gender bias.

    is beyond me.

  21. says

    Ignoring the barriers to entry could eliminate problems in other areas, too, like income inequality. Poor people could be buying yachts if they chose to. We’ll just ignore why they aren’t choosing to do so.

  22. Ariaflame, BSc, BF, PhD says

    Good scientists are the ones who when they create the models, know that the model is not the reality.

  23. anteprepro says

    How the fuck do you get published by saying the opposite of what your data is showing? That’s the most obvious fail state I can think of for a science paper, aside from completely asinine methodologies…

  24. twas brillig (stevem) says

    Just to pile on:

    But for the women, the number had plummeted from 72% in the first year to 37% as they finish their studies.

    The unaddressed question: Why did the number of women plummet? Me thinks stuff happened in college to shoo them away from their targeted discipline. And WHY do you blame their kindergarten experience as the culprit? And the overarching question: WHY are these women “opting out”? Okay, so they’re not in their targeted position yet, so the position itself isn’t the primary cause. But, think again, those positions aren’t isolated in bubbles; maybe there is something about the whole environment that makes them hesitate to join. Address THAT, paper writers; everything ain’t OK, get it?

  25. numerobis says

    The number of women plummeted because females biologically are predisposed to like feathers and feathery things, natch.

  26. numerobis says

    Oh wait; “plummet” is from “plomb” not “plume” — my folk etymology was off.

  27. ragdish says

    I only skimmed over the abstract. Haven’t had time to read the article in depth. Are the authors shifting the gender bias to earlier development a la as the twig is bent, so grows the branch? Or are they making the fatuous claim that there is no gender bias?

  28. chrislawson says

    Premise 1: Black US Presidents re-elected: 100%
    Premise 2: Non-black US Presidents re-elected: 42.9%
    Deduction: Black men can and do prosper if they choose to become President in the first place.