Problems with ‘Problems with the Discrimination Hypothesis’

Via Daily Nous, philosophers Neven Sesardic and Rafael de Clercq have written a paper called ‘Women in Philosophy: Problems with the Discrimination Hypothesis’. The paper is broad in scope – covering, among other things, problems with the empirical science behind the ideas of implicit bias, IAT tests, and stereotype threat and the possibility both of pro-feminist bias in which papers on these topics get accepted for publication and which papers in philosophy get accepted for publication (discrimination hypotheses to which the authors seem to apply rather more lax standards for evidence). It’s quite a list of topics for a relatively short paper.

I don’t have the time or the knowledge base to adequately discuss all the issues brought up in the paper. So in this post I’m just going to focus on two bits of it. (Hopefully our readers can weigh in in the comments.) First, this passage:

In her 2008 article “Changing the Ideology and Culture of Philosophy: Not by Reason (Alone),” MIT professor of linguistics and philosophy Sally Haslanger provides a table with percentages of women among the faculty of the top twenty graduate programs in philosophy in the U.S., ranging from 4 percent to 36 percent, and concludes that “the data mostly speak for themselves.”[5] But again, the data don’t speak for themselves at all. Without additional information, it is impossible to draw any conclusion.

To establish even a prima facie case for anti-woman hiring bias, it would be necessary to first compare the percentage of women among the job applicants with the percentage of women among job recipients. Only if the latter percentage is lower than the former is there prima facie evidence for hiring discrimination against women.

University of British Columbia professor of philosophy Andrew Irvine compared exactly these percentages in the Canadian academic job market over a twenty-year period ending in 1996.[6] According to his estimations, the percentage of female job recipients was on the whole higher than the percentage of female job applicants, which led him to conclude that “if systemic discrimination is occurring within contemporary university hiring, it is more likely to be occurring in favor of, rather than against, women.”[7]

To begin with, when Haslanger says that ‘the data mostly speak for themselves’, she isn’t referring merely to the employment data. I encourage everyone to refer back to her paper to see just how much data she compiled. Secondly, she’s focusing on the number of women employed in ‘top’ PhD granting philosophy departments in the US. In response, Sesardic and de Clercq refer to this paper which looks at employment information across all academic disciplines and all universities in Canada. How the information in the latter speaks to the former is unclear – especially since no philosophy-specific information whatever is given in the latter. Moreover, the paper in question doesn’t actually do what the authors suggest, which is to “compare the percentage of women among the job applicants with the percentage of women among job recipients”. The paper has no data – from philosophy or anywhere else – about the actual numbers of women and men applying for jobs. It estimates those numbers, based on the assumption that these percentages will (again, across all disciplines) exactly mirror the number of PhDs awarded. That’s a pretty big assumption. The authors go on to discuss a few more studies in a similar vein, but again none of these give any philosophy-specific, or indeed any non-Canadian, information (and some are specifically focused on the sciences).

What’s striking is that we do have quite a lot of – much more recent – philosophy-specific data courtesy of the hard work of Carolyn Dicey Jennings that’s directly relevant to the issue the authors are discussing. (Jennings data suggests that women get jobs in philosophy at roughly the same rate they get PhDs in philosophy.) The authors, however, do not discuss this data.

Now I’m going to turn my attention to this passage:

“Isms” Found in Article Titles

Stanford Encyclopedia of Philosophy


Number of Times Used

Feminism, feminist


Empiricism, empiricist


Materialism, materialist


Idealism, idealist


Rationalism, rationalist


Naturalism, naturalist


The number of SEP articles that contain “feminism” (or “feminist”) in the title is more than four times higher than the number of all articles combined whose titles mention empiricism, materialism, idealism, rationalism, or naturalism.

These numbers suggest that, far from sidelining feminism, philosophers make extra efforts to dedicate an inordinate amount of space to feminism, and precisely in those all-important publications that are focused on presenting the state of the art in philosophy to its practitioners as well as to the wider public.

Furthermore, even top philosophy journals have occasionally relaxed their standards to make place for feminist articles.

For authors very concerned that feminists are being careless in what the data suggests, that’s a relatively strong claim for what these numbers suggest. Suppose, for example, that we turned our attention from how often words appear in the titles of papers to how often they appear in the SEP in general. We get a rather different picture. A quick googling suggests the following. ‘Feminism’ appears 385 times in the SEP. In contrast, ’empiricism’ appears 573 times, ‘naturalism’ 569 times, ‘idealism’ 505 times, ‘materialism’ 431 times. (Some other ‘ism’s, if you’re curious: ‘realism’ 1050 times, ‘consequentialism’ 303 times, ‘pluralism’ 492 times, ‘dualism’ 435 times. For whatever that’s worth.)

Whether the occasional admission of feminist philosophy to top philosophy journals is evidence of those journals ‘relaxing their standards’ is, it’s fair to say, not a a hypothesis fully supported by the evidence. The authors – despite their strong objection, within the paper, to claims of bias based on anecdote alone – give only one purported example (of a paper published in AJP, which they don’t like.)

I want to leave aside the question of whether Sesardic and de Clercq are correct in their assessment of what the standards of evidence should be for establishing gender bias in philosophy. The simple point I am making here is that their paper doesn’t meet their own standards of evidence.

35 thoughts on “Problems with ‘Problems with the Discrimination Hypothesis’

  1. One point 1., discrimination can affect the pool of applicants as well as who gets selected from the pool. A work environment that is unwelcoming to women (in small ways and sometimes severe ways) can reduce the number of female applicants and do so in way that count as discriminatory. And of course Haslanger makes the case that the environment in philosophy is often just like this.

  2. Some of Sesardic and de Clercq’s arguments seem very weak indeed. For instance, the relatively larger number of encyclopedia articles with “feminism/t” in their title could very well be seen as a sign that philosophers make extra efforts *not* to include feminist perspectives among their mainstream articles, and that such perspectives then have to get their own entries. The data they give can be used to support the exact opposite of what they are claiming.

    Yet, I have to agree that much better data are needed to make the kind of assertions Haslanger made in her paper. In particular, data from top institutions and journals cannot be assumed to be representative of the whole field. It is true that we have many more data now than when Haslanger wrote this paper. This being said, it is not clear that we have enough data to sway partisans of the status quo or to convince people who are ready to attribute the underrepresentation of women in the field to differences in interests or abilities. Moreover, women in other fields are discriminated against even though they are not underrepresentated.

    A paper I wrote with Guillaume Beaulac is forthcoming in Ergo shortly and addresses many of these topics in more detail. In it, we suggest that the lack of data is not a reason for not acting. In fact, we argue that we need to address structural problems as soon as possible in order to make the profession a better place, for everyone. A preprint version is accessible at

  3. That’s a really important point, Adam! For some reason, the authors seem to construe Haslanger as making specific claims about hiring in philosophy, and targeting her discussion of bias at hiring practices. Whereas, of course, her discussion is much broader.

    David, thanks for the head’s up. (Also: weird!) I’ll see if I can fix the link.

    Yann, I look forward to reading your paper.

  4. Wait, are we not addressing the proverbial elephant in the room? I mean Sesardic and de Clerq’s claim that “even top philosophy journals have occasionally relaxed their standards to make place for feminist articles.”

  5. Also, the list of issues the National Association of Scholars (which publishes the journal the paper appeared in) is concerned with is quite interesting ( It includes “Overemphasis on issues of race, gender, class, sexual orientation”, ” Exclusion of conservative and traditional viewpoints” and ” “Multiculturalism,” “diversity,” “sustainability””…

  6. I am not familiar with this journal. However, I am wondering how much attention to feminist philosophy is considered “ordinate”. Is there a consensus on this matter?

    I am unsure what it means for a top philosophy journal to relax its usual publication standards. Does this mean that individual peer reviewers, upon confronting a paper with a feminist theme, approved it for publication more readily than they normally would have or remained oblivious to its flaws? Or did they find the feminist content inordinately interesting? I hope the authors did not mean it was an intentional decision on the part of the editors to lower standards simply to obtain a feminist philosophy paper; would that mean they abandoned the blind peer review practice, knowing a paper they desired to publish wouldn’t succeed? Why did they want one of these feminist papers so very badly, anyway? It sounds like something about that topic must be unusually desirable to editors, which is odd since I thought it was relatively rare to see papers on that topic in the most selective journals in the field.

  7. It’s important to note that especially at top tier programs, the majority of faculty were not hired as the result of the traditional sort of search from which junior folk are typically hired (i.e., candidate selected from a pool of applications received in response to a widely advertised national search). Most top departments are senior heavy, and most senior faculty at these top departments presumably got there as the result of senior hires, which often differ quite radically from their traditional junior counterparts. As such, the profession-wide recipient/applicant percentage for junior-hire women being roughly the same as that for junior-hire men says very little about the presence or absence of hiring discrimination in top tier phd programs, when those programs are staffed mostly by recipients of a procedurally far less formal (and well-documented) search much more narrow in its focus with a much, much smaller candidate…and so perhaps one traditionally more prone to the effects of discriminatory biases (implicit or otherwise).

  8. The above quotations from Sesardic and De Clercq’s article are misleading. The author (“Magical Ersatz”) cites S&DeC’s table (Table 2) showing that the number of SEP entries containing the word “‘feminism’ (or ‘feminist’) in the title is more than four times higher than the number of articles combined whose titles mention empiricism, materialism, idealism, rationalism, or naturalism.” Magical criticizes S&DeC saying that the word “feminism” appears in the SEP (i.e., not just in the titles of entries) 385 times, compared with 573 times for “empiricism,” 303 times for “consequentialism,” etc. In other words, feminism does not seem to be getting so much more attention than the traditional fields if we use a different measurement. The problem is that, in Table 1, S&DeC do report the “frequency of appearance” of the word feminism/feminist compared with empiricism/empiricist etc. in the Routledge Encyclopedia of Philosophy, and the relative frequencies they find are similar to those that Magical finds in the SEP. So S&DeC effectively do what Magical criticizes them for not doing (though they do it for the Routledge Encyclopedia of Philosophy, not the SEP).

    Magical quotes S&DeC’s statement, “Furthermore, even top philosophy journals have occasionally relaxed their standards to make place for feminist articles.” She makes it seem like S&DeC’s evidence for this assertion is simply the fact that sometimes top philosophy journals publish feminist articles. She omits the following *example* that they give of “a feminist approach to the logic of negation” from the Australasian Journal of Philosophy: “Such an account of ~p specifies ~p in relation to p conceived as the controlling center, and so is p-centered…In the phallic drama of this p-centered account, there is really only one actor, p, and ~p is merely its receptacle. In the representation of the Venn diagram, p penetrates a passive, undifferentiated universal other which is specified as a lack, which offers no resistance, and whose behavior it controls completely.”

  9. As the former co-editor of Feminism in the Stanford Encyclopedia, I’m in a position to easily explain their observation. 1. Feminism is an area with its own co-editors, who put the word ‘feminism’ or ‘feminist’ in every title commissioned. Empiricism is not such an area.
    2. The SEP is incredibly unusual in how receptive it is to feminism.
    3. Feminist philosophy, as readers here know, is very broad. There are feminist approaches to (I would suspect) the vast majority of traditional areas of philosophy, and to quit a few topics. We have tried to commission entries on each of these areas or topics, which does indeed lead to a huge multiplication of topics.
    4. These entries are not redundant, despite being on topics covered elsewhere, precisely because feminist approaches tend to be ignored.
    5. I kind of suspect (though I’ve only had hints that this is the case) that other area editors are not choosing to commission nearly as many entries as we did (and probably are still doing).

  10. Also, if the area you’re editing is e.g. Philosophy of Language, you don’t have anything in common that you put in every article title. Feminism is unusual in that way.

  11. Nossonbenreuven, if you read my post you’ll see I did mention that they give an example of a journal ‘laxing their standards’ to admit feminist philosophy – though just one, in a paper where they have previously railed against arguments by (lots and lots of) anecdote when used by feminists.

    And whatever they say about other sources, the point about SEP is just that the frequency of discussion of feminism in SEP looks very different depending on what we’re measuring – titles or general usage. That’s the kind of sloppiness they’d jump all over if it came from feminists. (I’m pretty sure they’d also jump all over the idea that we can read off much at all about the status of feminism in the discipline from what happens at SEP, for exactly the reasons Jenny gives above.)

  12. I’m refraining (with some difficulty) from posting terrible excerpts from papers having nothing to do with philosophy, in various top journals. It would help to show that publishing the occasional paper with a terrible bit is not at all correlated with feminist philosophy. Why? Well, one reason is that without full knowledge surrounding context I don’t think any of us should be judging the paper cited from the AJP to be terrible. Another reason is that it would be wrong to hold such papers up in that way, contributing to a truly ugly dynamic in our field that we should all be trying to stamp out.

  13. Sesardic’s publication list also includes articles criticizing philosophers of biology who reject the biological concept of race and articles criticizing philosophers who are critical of the hereditariness of IQ. Those sentences are practically unbearably twisted because (as might be obvious) the strategy of the articles appears very similar to the Women in Philosophy article: not put forward any positive view; rather, argue that a current explanation is unsuccessful. By only criticizing views, Sesardic can leave to the readers imagination alternative explanations for a given phenomena. For example, in criticizing Ned Block’s criticisms of The Bell Curve (!!!!), the nurture explanation for intelligence is made to look less plausible thereby making the nature explanation look more plausible without having to SAY that IQ is hereditary. In other words, the strategy allows for the spreading of fear, uncertainty, and doubt.

  14. Christy, is that true?
    My current department (Brown) and my alma mater’s (Princeton) both have *mostly* philosophers who were hired without tenure. Maybe this is unusual.

  15. Wow. Yes they do. The authors attribute this quote:

    “individual panel members’ ratings could be periodically reviewed for patterns that might indicate bias. If such patterns are found, these could be—carefully and privately—raised with panel members, as reflecting on one’s past biased judgments is an extremely effective way of reducing bias”

    to Margaret Crouch, citing it as quoted in Jenny Saul’s paper on ranking exercises and implicit bias. In fact, Jenny Saul is – pretty clearly, if you look at the text – not quoting anyone in this passage of her ranking exercise paper. She’s talking about some implications one might draw from from Crouch’s work, but she (very clearly) isn’t quoting Crouch.

    Jenny’s paper is here:

  16. Jamie, I certainly take it to be true of top programs only recently established (e.g., NYU, Rutgers) as well as for those long-established programs not traditionally known for tenuring their junior faculty. This together with the fact that movement in and out of the top-tier is largely a matter of the gains and losses of senior faculty, I had just assumed most senior faculty at top programs arrived via senior hire. That said, I’ve nothing more than mere conjecture to offer here and so am happily open to correction.

  17. Christy’s point strikes me as important even if we replace the claim that *most* people at top programs ended up there via senior searches with the claim that *many* people people at these places ended up there via senior searches. (And this latter claim definitely seem true.) That’s enough to make it the case that hiring at top PhD granting programs is often quite a bit different than it is across the board, where senior hires – though they are of course happen – are somewhat rarer.

  18. To put into context the article’s scepticism about stereotype threat, here’s a quotation from an excellent 2010 survey article concerning female under-representation in STEM subjects (

    “In one of the earliest experiments looking specifically at women, Spencer et al. (1999) recruited 30 female and 24 male first-year University of Michigan psychology students with strong math backgrounds and similar math abilities as measured by grades and test scores…The students were divided into two groups…One group was told that men performed better than women on the test (the threat condition), and the other group was told that there were no gender differences in test performance (the nonthreat condition)…They found that women performed significantly worse than men in the threat situation and that the gender difference almost disappeared in the nonthreat condition…

    In the ensuing decade more than 300 studies have been published that support this finding.The results of these experiments show that stereotype threat is often the default situation in testing environments.” (pp. 39 – 40)

  19. I honestly don’t think this article warrants attention. On the *most* charitable interpretation, it is a thinly-cloaked polemic against feminism (and women) in philosophy.

  20. Tom Dougherty, where does the article suggest that “it’s impossible to do meta-analyses concerning stereotype threat”? Sesardic and De Clercq cite 7 studies where the stereotype threat effect could not be replicated (see reference #73 in their paper). John A. List (U. of C.) said that he and his collaborators couldn’t induce stereotype threat even by telling women, “Women do not perform as well as men on this test and we want you now to put your gender on the top of the test.” Citing List, Sesardic and De Clercq suggest that there might be a file drawer problem, i.e., researchers who fail to replicate stereotype threat don’t submit their papers for publication. It’s a well-known fact that, in areas of research where there is a file drawer problem, meta-analyses will yield misleading results.

  21. “Ganley at al. point out that while published articles had a tendency to confirm the existence of stereotype threat, none of the three unpublished dissertations showed that effect. They also complain that they were unable to perform a meta-analysis because the number of available empirical investigations was too small and because many of the studies did not provide information necessary for calculating effect sizes”

  22. OK, so stereotype threat is more controversial than implicit bias. Any psych grad student could have told us that.

  23. Tom Dougherty,

    For what it’s worth, the quoted complaint made about an inability to do a meta-analysis was attributed to the paper by Ganley, and did not come directly out of the mouths of the authors of this paper.

    And what is obvious is that the real question here is: what is a proper meta-analysis? If one is sloppy in the criteria as to what counts as a relevant component study to include, then of course it can become quite easy to conduct a meta-analysis.

    Certainly the meta-analysis conducted by Geary and Stoet was very careful to include only those studies that included certain crucial elements, and it too was rather hamstrung by the relative paucity of studies that were on point.

    One might regard the problem with conducting such meta-analyses as of a piece with the larger crisis in social psychology, in which far too little effort is made to replicate previous studies.

  24. Your heavy-handed moderation is making it difficult to have an open conversation here. As a result the paper gets discussed elsewhere–in high-traffic venues such as LR and PMMB–in ways that are far more detrimental to your cause than any dissenting comments on here would have been.

  25. Concerned, since you aren’t one of my fellow bloggers I suppose I’m surprised that you think you’re in a position to know what comments we are and aren’t posting here!

    Fwiw, there have only been two comments I haven’t posted, and they were both clear violations of our comments policy. If that’s ‘heavy handed moderation’ by your standards, then so be it, but I’m not going to post comments that are just flagrantly rude. If anyone would like to post comments on this post or this paper – taking any position whatsoever – and can manage to do so without being a dick, I’ll happily post those comments.

  26. Magicalersatz, people talk. Maybe the colleague who told me of a post you didn’t approve misrepresented his tone or the content of his post. But even accounting for selection bias the conversation here seems extremely one-sided relative to what is going on at Leiter’s blog (let’s leave the PMMB aside, shall we).

  27. Yes, I’m sure people do talk. They can also read, if they want to. I’m not in the habit of not approving comments just because I disagree with them, or just because they’re critical of me, as this thread and plenty of others show. (Just scroll up to comment 9.)

    I don’t know what’s happening on Leiter’s blog, but I can’t say I’m inclined to think that a difference in tone or content between our blog and his is a bad thing. Again, I reiterate, that if anyone wants to post politely critical comments here, they are welcome to.

  28. Unfortunately, I haven’t the time to follow the details of the discussion on this post. However, I’ve been concerned about the complaints about censorship. i don’t think we’re a censorous bunch, so I’ve worried about what’s going on.

    One thing is that what our spam filter selects out has grown immensely. We can’t sort through the 500 or so emails a day that it dislikes. The spam watcher hates links and so some comments may get lost if they have links.

    Another thing is that what seems harmless to one person may seem much more problematic to another. I’m thinking here of attacks on people who are not regular posters on this blog. So if someone claims that Jackie of the UVA story has been finally discredited by the Washington Post, I’m going to pause before approving it. That’s an extremely serious charge with really no evidence. I tried to track down which WAPO piece it could be, and as far as I could see, the article is taking almost any one seriously who just denies what was in the Rolling Stone article.
    Tocut to the chase, like magical ersatz, I try to let through all reasonably civil comments. I hunt through the results of the spam filtering, but the numbers are often very large. The ones I fret over have real problems with matching the seriousness of the comment with the weight of the evidence offered.

  29. Hi. Can we please have a thread on Emily Yoffe’s ‘The College Rape Overcorrection’, published in Slate on Dec. 7th?

    It was featured on NPR a few days and I’ve seen discussions in many places, but never on the philosophy blogs. Lots of comments there need addressing.

    Thank you.

Comments are closed.