This Year in Philosophical Intellectual History

This Year in Philosophical Intellectual History


This fall, one of the most powerful institutions in the field of philosophy in this country began to collapse…

In “The Rise and Fall of the Philosophical Gourmet Report,” a brief post at the U.S. Intellectual History Blog, historian Ben Alpers takes a look at one of the major stories in the philosophy profession this year. Alpers is cautious about his account of the story—more marking it for future analysis than delving into it deeply. He is self-conscious about his status as an outsider to the profession, but he also takes a stab at understanding its significance:

I had wanted to write something about it… But I never felt that I really understood the details well enough to add anything interesting to the conversation you’ll find on philosophy blogs. But before the year is up, I wanted to post something on the controversy, because I think that the rise and fall of the PGR will be a wonderful future topic for U.S. intellectual historians. The appeal of such a ranked list tells us interesting things about the field of philosophy in the late 20th and early 21st centuries, as do the specific philosophical preferences of Leiter and the PGR.

I agree that this could be an interesting topic for future intellectual historians and sociologists studying the philosophy profession. Despite attempts to reduce it all to a personal smear campaign, what happened this year, in my view as an observer of the profession for over a decade, was the result of several important changes in attitude and power structure in the profession. In a previous and apparently divisive post, I gave the result of these changes a name: “The New Consensus“:*

Over the past couple of years, philosophers have witnessed the emergence of a new consensus—one that rejects acquiescence to abuses of power in philosophy, one that seeks to overturn rather than turn away from the profession’s problems, one that seeks to support rather than silence the vulnerable. At the same time, the consensus reinforces two very traditional pillars of philosophical practice: first, it recognizes that criticism is the currency with which philosophers pay respect to one another while insults are just cheap counterfeits, and second, that we should go wherever the arguments and inquiries take us, putting aside philosophical prejudices and erasing boundaries if need be.

What I had in mind with this admittedly somewhat purple prose was that there seemed, first, to be a greater recognition of the kinds of problems our profession faces and acceptance of and support for attempts to remedy them. Second, there seemed to be a wider range of voices confidently making themselves heard on a variety of issues, and, among other things, an increased willingness to call out more powerful people in the profession for making personal attacks on others. And third, there seemed to be a greater acceptance of a diversity of approaches to philosophy, and more skepticism about attempts to dismiss inquiries by saying “that’s not philosophy.”

I think these are improvements and I am happy for whatever small role Daily Nous played in helping them come about.

One thing I said in the New Consensus post that I think is often overlooked is: “There are still many with legitimate personal and structural complaints about our profession.” Things are far from perfect, clearly. One of the good things about the current state of affairs, though, is that there are so many outlets for bringing issues to light, discussing them, and working towards solutions. Daily Nous is one of those outlets, so feel free to bring your concerns here. I am not under the illusion that every or even any problem can get solved on a website, but at least information and perspectives and ideas can be shared.

Alpers notes, “It’s obviously much too soon to write this history.” That’s true. But what we can do now is draw attention to what we think would be important to include in that history.

(art: detail of “Nature” by Richard Tuttle)

*see first comment.

guest
46 Comments
Oldest
Newest Most Voted
Inline Feedbacks
View all comments
?
?
6 years ago

With all due respect, Justin, I’m not sure that what you think is being accomplished is being accomplished. Or if it is, whether it’s any better.

There are different voices, to be sure, but they’re trying to silence others just like the former voices were. Anyone who suggests that if they were a graduate student falsely accused of rape they would be very glad to have David Barnett on their side, or who suggests that the Site Visit Team did anything wrong, or who suggests that Carolyn Dicey Jennings’ data (that women who receive TT jobs have on average half the publications of men who receive TT jobs) indicates that women get preferential treatment, or who suggests that people in poverty face discrimination, gets shouted down, morally censured, and ridiculed. It’s what keeps me from posting this in my own name or with my own email, and what prompts me to use Tor to do so — because to question this new consensus is seen as morally wrong. (Usually one can see what group is in power by what group is willing to post under their own names.) But there are legitimate concerns with it, and they should be voiced and responded to. Responded to with actual reasons and arguments, which one doesn’t get from the new consensus.Report

JT
JT
6 years ago

?: “…women get preferential treatment…”

Really? That’s what you’re hanging your hat on? That philosophy is such a wonderful place for women must be a very, very well kept secret. It’d be hard to explain the discipline’s demographics otherwise. Unless you implicitly think that women are just bad at philosophy in virtue of their gender/sex. That’d be bad. That’s the kind of thing that gets you (rightly) “shouted down, morally censured, and ridiculed.”Report

Ben
Ben
6 years ago

The fact that Justin felt he had to add a “disclaimer” to his own post is pretty telling, I feel.Report

Roadster
Roadster
6 years ago

@JT – the CDJ hiring data are for recent years only. The overall demographics of the profession are caused by hires going back to the sixties. I agree that the climate for women isnt’t great in the profession, but you can see why some young males, especially those from less privileged backgrounds, are worried about the CDJ data. I have been on search committees. There is significant pressure to give women preferential treatment, especially at the shortlisting stage.

As for ?’s main point, I must say I’m in broad agreement. For many of us the lesson of the September Statement/New Consensus has been that there’s a group even more powerful than Leiter & co, a group less brazen but equally ruthless in disposing of its enemies. Just so we’re clear, I don’t think Justin belongs in the inner circle of this group.Report

Roadster
Roadster
6 years ago

Justin, “disposed” wasn’t meant literally! Anyway, exhibit A: even Leiter has (apparently) been putsched after messing with one particular elite philosopher. There are many others. Exhibit B: Faculty-grad student relationships are publicly condemned by prominent exponents of the New Consensus (again, you’re not one of them), but when their friends in elite circles engage in the same behavior they get a free pass.

“?” Is right to say that who chooses to post under their real name tells us who is to ne feared.Report

JT
JT
6 years ago

@ Roadster

Whether that’s a reasonable worry depends on what we take the climate problem to be. If part of the climate problem is that women are systemically disadvantaged in other respects, then we can’t reasonably infer from there being preferential hiring practices to the conclusion that this that women are given preferential treatment on the whole or that these practices are unfair to men. It may just be that these practices confer advantages that merely offset disadvantages elsewhere. Moreover, I take it that you’re that not claiming that these practices amount to an explicit affirmative action scheme across the profession, so I expect that the advantage conferred will likely be marginal.

Also, CDJ has already addressed the problems with inferring from her data to the conclusion that there is significant preferential treatment in hiring here: http://www.newappsblog.com/2014/12/gender-and-publications.html#more. The upshot is that the disparity in publications appears only when looking at the mean, but not the median and mode. She suspects that the difference is generated by a “difference in opportunities prior to the tenure-track.” This post has been up for a while now, so it’s somewhat disingenuous to make the bald faced claim that women get preferential treatment just on the difference in mean number of publications alone–at best the data is inconclusive. This is especially so in light of the climate problems you yourself point out.

I agree that a selective reading of the CDJ hiring data can be construed as a reason for worry on the part of young underprivileged males, but it also seems to me to be a pretty weak grounds for attacking efforts to improve the climate for women. These efforts don’t preclude a movement towards improving conditions for males who are in poverty or come from an impoverished background. Redressing systemic injustice is not zero-sum. This point is so obvious that I can’t help wonder about the sincerity of advocates of this argument.Report

Roadster
Roadster
6 years ago

@ Justin – can’t Leiter be both an executioner and a victim? And what do his philosophical views have to do with this? We’re talking about a political struggle for power within the profession.

@JT – CDJ’s attempt to provide an alternative explanation for her data seems rather tortured, and has widely been recognised as such. I agree that we don’t know whether AA in hiring overcompensates for other, previous discrimination. I don’t agree that academia isn’t a zero-sum gain. The most privileged are least likely to be harmed by AA towards an underrepresented group. People who are, say, privileged by race and gender but disadvantaged by class and so educational opportunities will likely suffer more from AA than white upper class males.Report

Captain Sensible
Captain Sensible
6 years ago

With respect Justin, it is perfectly plausible to see the campaign against Leiter as not a response to some recent testy comments but a focus for a bunch of people who disliked him and/or the PGR for a whole host of reasons going back for a long time and is plausibly read as a smear campaign. And as for your remark that you have been ‘an observer for more than a decade or so’, and so you don’t see it that way, there are others having been observers for at least twice the length of time and don’t share your opinion. To put it down this sea change as the emergence of a ‘New Consensus’ is divisive and not merely apparently so. You apparently have decided which side is right and expressed in such a way that the September campaign is something other than a smear campaign.Report

Lion Rampant
Lion Rampant
6 years ago

A small group of people demonstrated that they had the power to oust Leiter from the editorship of the PGR when one of them was attacked by him. If that’s not a philosocabal (thanks, Justin!) I don’t know what is.Report

Roadster
Roadster
6 years ago

I think Lion Rampant’s view is more plausibile than Capt Sensible’s. The longstanding PGR critics got some breathing space and serious airtime only after a small elite group drafted and signed the September Statement, and that in turn only happened after one specific person was attacked by Leiter. The wording of the SS is clear about this. The many signatures on the SS don’t seam to have made much difference, but the clout of the original drafters and signatories with the PGR board did.Report

Lion Rampant
Lion Rampant
6 years ago

Fundamental power asymmetries in the profession haven’t changed: the New Consensus still means that we have to keep on the right side of well-connected people at elite institutions. The people and their priorities have changed a bit.Report

Captain Sensible
Captain Sensible
6 years ago

Roadster, I don’t consider my view as inconsistent with Lion’s views, with which I agree,Report

JT
JT
6 years ago

@ Roadster

Did you even read CDJ’s post? We don’t need to accept CDJ’s explanation to see that the median and mode number of publications are pretty even. That clearly does not support the conclusion that women get preferential treatment in hiring.

Also, the point wasn’t that academic is zero-sum. Rather it is that efforts to increase the number of women in the profession does not preclude increasing the number of members from other underprivileged groups. You might be right that AA-type efforts for women may in effect take jobs away from other underprivileged groups, if no efforts are made to improve the situation for those other groups. But shouldn’t the response be to make efforts to improve the situation for those groups, including implementing AA-type practices, rather than give up the game all together? Alternatively, one might think that AA-type practices are not the way to go here, but that’s not the claim on the table.Report

Roadster
Roadster
6 years ago

@JT – did you read the comments under CDJ’s blog post? What do you think is the inference to the best explanation, given the data?

On AA pratices, a lot seems to depend on whether we think some forms of disadvantage are more fundamental than others. I happen to think that economic disadvantage is more fundamental. Very roughly, if that’s the case then AA for, say, women will turn out to be largely a power grab by an underrepresented subset of the most advantaged group. The ‘Market Boost’ committee are a prime example: they take themselves to be a disadvantaged minority when arguably their elite undergrad education (all but one of them) places them at a significant overall advantage relative to most aspiring or junior philosophers.Report

JT
JT
6 years ago

@ Roadster

I did. Which comment are you referring to? In any case, I’m not sure what the best explanation is, but I’m pretty sure it’s not that women are given preferential treatment in hiring across the profession. The mode is 0 and the median is 1. You tell me what the best explanation of that is.

Maybe economic disadvantage is more fundamental, though I’m highly sceptical of these kinds of claims. I’m not sure why this means that AA for women constitutes a power grab. Again, shouldn’t the solution be to extend AA-type practices to the economically disadvantaged than to deny them to women, who are also disadvantaged? We might even think that if economic disadvantage is really more fundamental, we should design AA schemes so that they benefit from them more than women, in lieu of denying them to women. I’m not sure what the best answer here is, but the point is that there are plenty of better options on the table.Report

Roadster
Roadster
6 years ago

@JT – comment n4, and and CDJ’s convoluted response to it.

AA is a liberal “solution”, and as such it is inadequate to tackling structural inequalities, because it is agency-focused. I realize this isn’t a popular view in the profession, where bland liberalism still rules the day.Report

Ligurio
Ligurio
6 years ago

I think that there are overlapping sets of different parties and groups that have come into increasing conflict over the past few years for a multitude of reasons. Philosophy is pretty far behind the curve among the humanities and many social science disciplines when it comes to sexual and gender equity. If you’ve been in academia for any length of time, it seems to me you simply must acknowledge this. It is pretty obvious. The more difficult question is to what extent this is a result of the combative and aggressive intellectual culture of philosophy, to what extent it’s a result of a (until recently) more or less accepted culture of males behaving badly, and to what extent the intellectual culture feeds or encourages the sexual one.

If you believe there is a close connection between these two cultures, then you might choose to fight sexual and gender injustice both directly and indirectly, by challenging the norms of the intellectual culture that (perhaps) contributes to such injustice.

Now consider that there is also a pretty obvious generational which is separate from but directly related to the issues above. People under the age of forty (it seems to me) tend to be much more concerned about civility and tone and in general avoiding offense in written and spoken discourse than is the older generation of philosophers (and academics from other disciplines as well). There is a new kind of moralism among the younger set that is not *simply* (I think) a function of their giving a damn about the sexual and gender inequities of the discipline.Report

JT
JT
6 years ago

@ Roadster

I read CDJ’s response as tentative and speculative in light of her recognition of the complexities involved. But sure, w/e.

In case, NS’s comment doesn’t cite any reason whatsoever to think that there is preferential hiring, i.e., ‘positive discrimination’ _on the basis of CDJ’s data_. And neither have you. If there was preferential hiring, we should see differences between the median and mode numbers of publications for men and women. The data, not CDJ’s explanation, shows that these numbers are actually fairly close. I’ve brought this up a number of times now and you’ve side stepped it every single time. Needless to say, it’s pretty frustrating. If this argument is wrong, tell me why.

Also frustrating is your change in tune. I’m certainly no bland liberal. And I’m not sure that AA is the way to go either. (To be perfectly clear, this is not me ceding the argument on the question of whether women get AA-type preferential treatment in hiring. I’ve yet to see any good reason to think that there is widespread adoption of AA or AA-type practices in the profession.) But what was at issue in our exchange wasn’t the merits of AA or AA-type practices in general. What was at issue was whether the fact that other groups are underprivileged too gives us a reason to abandon AA/AA-type practices for women, assuming that such practices are in place for the sake of argument.Report

Anon23
Anon23
6 years ago

They are happy to call others out in the most strident of terms, but none so happy when they themselves are called out. So, when one of their commentators posted:

‘I’d find nothing whatsoever problematic about a black professor who expressed indifference about violence about the cops’

I decided to call them out:

‘On your way out tonight I’d like you to stop by a cop and explain to them why it is ever legitimate for someone to be indifferent as to whether they get to go home tonight to see their family or fight for their life on the floor of A&E. Seriously, you should be ashamed. It’s not ‘right on’ to be indifferent to violence towards other human beings. It’s disgraceful…’

Now, on the question of silencing, this comment was promptly deleted and when I questioned why I was told that what I’d done was ‘pretty awful’. Readers can decide for themselves, in light of recent events, just how ‘awful’ what I had to say was…Report

Philippe Lemoine
Philippe Lemoine
6 years ago

I don’t understand CDJ’s argument relative to her data on hiring, but perhaps I’m just missing something. According to her data, the mean number of publications for women who were hired during the period she considered is 1.13, whereas it’s 2.17 for men. I haven’t run a statistical test, but I’m guessing that such a difference is extremely unlikely to result from chance alone, so we need another explanation.

The fact that, if you look at median number of publications, this difference evaporates, doesn’t mean that we don’t need to explain why the mean number of publications is significantly different for men and women. The hypothesis that women get preferential treatment at the hiring stage, at least for tenure-track jobs, is one explanation.

I take it that CDJ’s, and perhaps JT’s, reasoning is that, since the median number of publications is the same for men and women, it means that it’s not the case that most men who were hired during the period she considered had less publications than most women who were hired during that period. But this doesn’t show that women didn’t get preferential treatment since, as I already noted, it doesn’t make the difference in mean number of publications disappear.

Even if you only care about median number of publications, since the mean number of publications for men is almost twice as much as that for women, it must be that, for some n, the median number of publications for men with at least n publications is greater than the median number of publications for women who had at least n publications. We still need to explain that fact, even though when you look at the whole sample of people who were hired, there is no difference between the median number of publications for men and that for women.

The way I see it, looking at median instead of mean numbers of publications just obfuscates a difference between men and women in the sample CDJ examined, but it doesn’t make that difference go away. Speaking more generally, mean and median numbers give different kinds of information about a distribution, but the information given by one doesn’t annihilate the information given by the other. This seems right to me, but perhaps I missed the point of the argument.Report

Philippe Lemoine
Philippe Lemoine
6 years ago

In the above, instead of “it’s not the case that most men who were hired during the period she considered had *less* publications than most women who were hired during that period”, you should read “it’s not the case that most men who were hired during the period she considered had *more* publications than most women who were hired during that period”. Of course, both statements are true, but I take it that CDJ had the second one in mind. Sorry for the mistake, I should have checked my comment before posting it, but I rarely do.Report

?
?
6 years ago

Hi Justin-

In response to the last two, I think you might find yourself in an uncomfortable position in the New Consensus. Perhaps I’m wrong, but I don’t think they’ll like that you’re willing to grant that women receive preferential treatment, even if you think that’s okay. (In fact, I believe the same.) And I’m sure you could tell that some prominent members of the New Consensus didn’t appreciate the poverty thread, which drew attention away from the issue of women in philosophy.

It will be interesting to see how much agreement there is between members of the New Consensus (for example, does Feminist Philosophers speak for the new consensus), and whether others are willing to go as far as the more vocal members (e.g., that a critic is a rape apologist).Report

Roadster
Roadster
6 years ago

One example: Gabriele Contessa, fiercely attacked on very thin civility grounds on FB by a prominent FP, her prominent husband, and a bunch of bandwagon-jumpers. His crime?Report

Philippe Lemoine
Philippe Lemoine
6 years ago

I’ve been thinking some more about CDJ’s argument from the data she compiled on hiring and here is what I think she might have been thinking when she wrote that post.

(1) There is a difference between the mean number of publications for men and that of women that is unlikely to result from chance alone.
(2) One natural explanation is that, at least for tenure-track jobs, women get preferential treatment at the hiring stage.
(3) But if women got preferential treatment at the hiring stage, the median number of publication for men would be greater than that for women, which it’s not.
(4) Therefore, whatever explains the difference between the mean number of publications for men and that of women, it can’t be that women get preferential treatment.

On that reading, she is not denying that we need to explain why the mean number of publications for men is greater than that for women and therefore she is not suggesting that we ignore that difference, but she is arguing that data about the median number of publications rule out the possibility that women get preferential treatment. The obvious problem with that argument, if indeed that’s what CDJ had in mind, is that (3) is clearly false. At least, if it’s not, I really can’t see why.Report

?
?
6 years ago

Hi Justin-

If I must restrict myself to merely political positions, and definitively show that actual prospects in the profession have been damaged or destroyed, I won’t do it. I can only show that people who have disagreed with what I’m thinking of as the New Consensus have been the targets of attempts to do so. And I’d do so by pointing at the people who have commented with their own names on http://feministphilosophers.wordpress.com and exhibited less than full-throated agreement. It’s a mistake they won’t soon make again, I think. I don’t want to reproduce their names because I don’t want to remind the people who run that blog, lest the attempt begin anew. (I see you’ve done the same by changing the name of the person I wrote to “a critic”, which I support.)

And I am decidedly not the same person that came up with Brian Leiter as a target of abuse. I do still find it interesting that it took him attacking a popular attractive tenured white person before the profession stood up to him. I think it’s good that we finally did, and I think his latest three attacks were horrendous, but there were earlier attacks on less privileged people that should have prompted the same reaction and didn’t. (I suppose this speaks even more to the power people enjoy by being vocal supporters of the new consensus.)Report

Roadster
Roadster
6 years ago

@? – I didn’t mean to morally condemn the attacks on Leiter, and I certainly wouldn’t call them “abuse”. I used that example simply to illustrate the power of the New Consensus philosocabal who drafted the September Statement.Report

JT
JT
6 years ago

@ Philippe Lemoine

I agree that the difference in mean numbers needs to be explained. And of course the median numbers are compatible with the explanation that the mean is skewed because there is preferential treatment in some cases. But taking the median into consideration as well undermines the inference to the claim that there is widespread preferential treatment, no? (I realise now that I’ve put the claim somewhat infelicitously in some of my comments. I was taking aim at ‘?’s and Roadster’s claim that the CDJ data is a reasonable cause for worry for men in the profession. My objection was that this worry is not reasonably supported by the CDJ data, given the median numbers. Sometimes I put it in terms of ‘widespread preferential treatment’, and sometimes it was in terms of just ‘preferential treatment’. I’m sorry for any confusion this caused.)

In her post, CDJ notes that the mean appears to be skewed by a difference between the percentages of men who have 5+ publications (15%) and the percentage of women who have 5+ publications (5%). This difference disappears, she tells us, “If we look only at those placed candidates without (reported) prior positions, the percentage of men with 5+ publications drops dramatically to 4.67%, whereas the percentage of women with 5+ publications stays about the same at 4.08%. Thus, what appears to be pushing the mean away from the median are male candidates outside the top-20 who have prior positions.” CDJ suggests that “the difference in mean publications may be partly caused by the difference in opportunities [between men and women] prior to the tenure-track.” Another possible explanation is that it is easier for men to move from a non-TT post to a TT job. As I mentioned in an early comment, I’m not sure which explanation (including that there is preferential treatment for women in some cases) is best. Neither is CDJ, at least in the Newapps post. (I think it’s telling that those who steadfastly insist that the data shows worrisome preferential treatment for women take her uncertainty on the matter, and gender issues in the profession more broadly, as her being ‘convoluted’ and ‘desperate’.) But to my mind, these numbers go to show that even if there is preferential treatment for women, it’s only in a limited number of cases.Report

Philippe Lemoine
Philippe Lemoine
6 years ago

I don’t really see why the fact that there is no difference between the median number of publications for men and that of women undermines the hypothesis that women get preferential treatment. Since the mode is zero for both men and women, I find it entirely unsurprising that, even if women get preferential treatment, one should not be able to see that if one is only looking at the median numbers of publications.

But after reading your comment, I read CDJ’s post more carefully and noticed another point she makes, which I think might have more merit. She notes that, although women get tenure-track jobs in the same proportion they are awarded PhDs in philosophy, they get post-doctoral/VAP/instructor positions in a much smaller proportion. Therefore, if the difference between the mean number of publications for men and that for women disappears once people who already had a position when they got a tenure-track job are ignored, it suggests that although women get preferential treatment at the hiring stage for tenure-track positions, this advantage may at least in part compensate a disadvantage they suffer from at the hiring stage for post-doctoral/VAP/instructor positions.

Based on what you say in your last comment, I suspect this might also be what you were getting at. However, although for all I know this argument might well be correct, I think a few remarks are in order.

First, for the argument to work, it has to be the case that if people who already had a position when they got a tenure-track job are ignored, there is no significant difference between the mean number of publications for men and that of women. CDJ says a few things which suggest this might be the case, but as far as I can tell she didn’t check.

Moreover, this argument will only be interesting if women are unfairly disadvantaged at the hiring stage for post-doctoral/VAP/instructor positions, but the fact that they get this sort of position in a much smaller proportion than they are awarded PhDs is not enough to show that. Among other things, we should also make sure that women who get a post-doctoral/VAP/instructor position don’t on average have significantly less publications than men who get that kind of position. If they do, then it’s not clear that women are unfairly disadvantaged at the hiring stage for post-doctoral/VAP/instructor positions, in which case CDJ’s argument wouldn’t be particularly convincing.

Finally, even if CDJ’s argument turns out to go through once we have checked all that, it will not really show that women don’t get preferential treatment at the hiring stage for tenure-track jobs, but only that at least part of that preferential treatment merely compensates a disadvantage they suffer from at the hiring stage for post-doctoral/VAP/instructor positions.

The good news is that CDJ also posted the data she compiled, which I think is really cool, so it should be relatively easy to check all the things I argued that we should check to determine whether her explanation is convincing, at least if it contains all the information we need. I will try and do it eventually, but the problem is that I have no idea how to use Excel, so I will have to learn first.Report

Philippe Lemoine
Philippe Lemoine
6 years ago

As I already explained, since the mean number of publications for men is greater than the mean number of publications for women in a way that is unlikely to result of chance alone, it seems reasonable to assume that women get preferential treatment at the hiring stage for tenure-track jobs. But I guess it also seems right that, had the same thing been true of the median numbers of publications for men and women, the case for that conclusion would have been stronger. It just occurred to me that it might have been what you meant in your last comment, in which case I don’t disagree with you on that point.

I still think that CDJ also made another argument, which should be distinguished from the one I just summarized, based on the suggestion that perhaps women are unfairly disadvantaged at the hiring stage for post-doctoral/VAP/instructor positions and that it’s what explains the difference between the mean number of publications for men and that for women. As I argued in my previous comment, although this argument may have some merit, I think it’s not clear at this point and, in any case, it wouldn’t show that women don’t get preferential treatment at the hiring stage for tenure-track jobs.Report

Philippe Lemoine
Philippe Lemoine
6 years ago

I have used CDJ’s data spreadsheet to check the hypothesis that, even if women get preferential treatment at the hiring stage for tenure-track jobs, it only compensates the unfair disadvantage they face at the hiring stage for post-doctoral/VAP/instructor positions. As I noted above, in order to determine whether this hypothesis is not disconfirmed by the data CDJ compiled, we must check that (1) the mean number of publications for women who got a post-doctoral/VAP/instructor position is not significantly less than the mean number of publications for men who got that kind of position and (2) there is no significant difference between the mean number of publications for men and that for women once people who already had a position when they got a tenure-track job are ignored.

Relative to (1), I found a mean number of publications for men who got a post-doctoral/VAP/instructor position of 1.65, whereas it’s 0.87 for women who got a post-doctoral/VAP/instructor position. I didn’t run a statistical test, but it seems clear that such a difference is very unlikely to result from chance alone. Thus, although women get post-doctoral/VAP/instructor positions in a smaller proportion than the proportion in which they are awarded PhDs, CDJ’s data suggest that it’s not because they are unfairly disadvantaged, at least not at the hiring stage for that kind of position. (Of course, even if there were no difference between the mean number of publications for women who got a post-doctoral/VAP/instructor position and the mean number of publications for men who got that kind of position, it would not show that women are unfairly disadvantaged since, among other things, it could be that female applicants have on average less publications than male applicants.) If that’s right, then giving women preferential treatment at the hiring stage for tenure-track jobs because people assume they are unfairly disadvantaged when they apply to post-doctoral/VAP/instructor positions doesn’t seem justified. Of course, women may be unfairly disadvantaged at some earlier stage, but CDJ’s data are silent on that.

As for (2), I found a mean number of publications for men who already had a position when they got a tenure-track job of 1.37, while it’s 0.86 for women in the same situation. Thus, even if we ignore people who already had a position when they got a tenure-track job, there is still a difference between the mean number of publications for men and that for women, although it’s smaller than before. Again, I haven’t run a statistical test to make sure of that, but I’m guessing it’s significant. So it seems that CDJ’s hypothesis, if indeed that’s what she had in mind, is disconfirmed by the data she compiled.

Although I’m confident they undermine CDJ’s hypothesis, I’m not sure how to interpret those results. It could be that women don’t fare as well as men when they apply to post-doctoral/VAP/instructor positions because, for that kind of positions, affirmative action policies are not as strong as for tenure-track positions. In that case, the fact that when people who already had a position when they got a tenure-track jobs are taken into account, the difference between the mean number of publications for men who got a tenure-track job and that for women increases, might just be the indirect consequence of the fact that affirmative action policies are less strong at the hiring stage for post-doctoral/VAP/instructor positions. But I’m just a grad student, so I have no experience with the way in which hiring for post-doctoral/VAP/instructor positions differ from hiring for tenure-track jobs and, therefore, I have no idea whether this hypothesis is plausible.

If you want to check my analysis, you can find the spreadsheet I created here: https://www.dropbox.com/s/jyd2od6vdafl98z/PlacementData.xls?dl=0. I just added a sheet called “Quick & Dirty” to CDJ’s data spreadsheet, who I think should be thanked for her work, which must have taken her a lot of time and effort. I don’t think I made any mistakes, but since I had never used Excel before (actually, I used Calc from OpenOffice, because I don’t have Microsoft Office), I strongly encourage people who are more familiar with this software to check and make sure of that.Report

Philippe Lemoine
Philippe Lemoine
6 years ago

In the above, you should read “I found a mean number of publications for men who did *not* already have a position when they got a tenure-track job of 1.37, while it’s 0.86 for women in the same situation” instead of “I found a mean number of publications for men who already had a position when they got a tenure-track job of 1.37, while it’s 0.86 for women in the same situation”, as you probably have guessed since otherwise what I say doesn’t make any sense.Report

Bijan Parsia
6 years ago

All other things being equal, I would think it’s pretty obvious how the equal medians would undermine a bias hypothesis given unequal means. This undermining isn’t necessarily fatal, of course, but enough that saying ” who suggests that Carolyn Dicey Jennings’ data (that women who receive TT jobs have on average half the publications of men who receive TT jobs) indicates that women get preferential treatment” is not really sensible.

Let there be 20 candidates, ten of each comparison group (e.g., women/men). Now imagine that for each group and for each number from 1 to 10, there is exactly 1 member with that number of publications. So, at the moment, median and mean are equal and equal between groups. Now take the 10 from one group and change it to 20 publications. Now the mean is 6.5 for that group and remains 5.5 for the other (the medians are still ≈5.5).

So, knowing nothing else, this sort of separation of median and mean have a lot of plausible explanations other than preferential treatment including ones people might find objectionable (e.g., that the best men are the market are “much better” than the best women but then there’s a convergence in quality as you go down from the top tier) as well as other ones (there might be a skew in area or teaching vs. research focus that let other properties dominate).

I’ll also note that we have a potentially tricky sample and that hiring isn’t made on the basis of publication alone and certainly publication *numbers* alone. Or so we should all hope. And, these are stats for successful candidates, yes? Unless we know the shape of the unsuccessful candidate pool, it seems really very hard to conclude a bias. Do we really thing that the pool of unsuccessful candidates is going to have the same distribution of publication numbers, esp at the high end, as the pool of successful candidates? If departments are selecting for publication number this would be a surprise unless there were a megaton of 5+ publishers and then how did all the male 1+ publishers beat them out?

I think CDJ’s causal hypothesis is interesting but I’ve not fully absorbed Philippe’s critique (@35) yet. Prima facie, I still think there’s potentially an error in focusing on the whole cohort rather than stratifying given the weight of the high end.Report

Philippe Lemoine
Philippe Lemoine
6 years ago

I agree that, if there is no difference between the median number of publications for men and that for women, depending on why this is the case, it might undermine the hypothesis that women get preferential treatment. It’s also true that CDJ’s data do not contain a lot of information that we’d like to have, especially the distribution of number of publications for applicants whether or not they were successful. And, of course, the decision to hire someone is not based only on what the applicant published, let alone how many papers he published.

But, at the moment, we only have CDJ’s data. According to her data, the mean number of publications for men is significantly greater than the mean number of publications for women, so we can still ask what could explain that difference. It should go without saying that, whatever conclusion we reach on the basis of CDJ’s data, we should be careful before we accept it as true and try to obtain more data to adjudicate the facts.

My point is that, although there is no difference between the median number of publications for men and that for women, it doesn’t undermine the hypothesis that women get preferential treatment. This should hardly come as a surprise because, even if women get preferential treatment, there are many ways in which the median number of publications for men could still be the same as the median number of publications for women. Moreover, CDJ’s data also don’t support the hypothesis that, even if women get preferential treatment at the hiring stage for tenure-track jobs, it at least in part compensates the way in which they are unfairly disadvantaged when they apply to post-doctoral/VAP/instructor positions. For all we know, this hypothesis might be true, but that’s not what the data CDJ compiled indicates.

You’re saying that, since there is no difference between the median number of publications for men and that for women, the difference between the mean number of publications for men and that of women might just result from the fact that men at the right tail of the distribution on average have significantly more publications than women in that situation. If I understand you correctly, your point is that, if this is why the mean number of publications deviates more from the median for men than for women and not in the same direction, then it undermines the hypothesis that women get preferential treatment.

But this hardly seems obvious to me. Perhaps it makes sense if you assume that how many publications applicants have is the only factor in hiring decisions, for in that case people whose number of publications is relatively large are de facto the strongest applicants. But, as you noted yourself, that’s clearly not the case. It might be that people with a lot of publications are in fact compensating for other characteristics which on the job market are weaknesses, one of which might be the fact that they are men.

(This also means that, even if we consider a subset of the sample for which there is absolutely no difference between the mean number of publications for men and that for women, it doesn’t follow that at least women in that sample did not also benefit from preferential treatment. The fact that, when we consider the whole sample, there is a significant difference between the mean number of publications for men and that for women is, I claim, prima facie evidence that they did, although of course we lack so many information that it could easily be defeated.)

Of course, if the reason why the mean number of publications for men deviates so much from the median was that the 5 persons with the largest number of publications – who represent 1% of the individuals in the sample CDJ examined – were all men and all had a ridiculously high number of publications, so that the situation was analogous to what’s going on in your toy example, I agree that it would completely undermine the hypothesis that women get preferential treatment. But, if you look at the data, you will see that it’s not at all what’s going on in the sample CDJ examined.

CDJ says that, if the mean number of publications for men is twice that of women, it’s because 15% of men but only 5% of women have 5 publications or more. But I have checked and, even if you ignore those people, the mean number of publications for men is still 36% higher than the mean number of publications for women. And even if ignoring those people had been enough to eliminate the difference, they represent more than 12% of the sample, which is hardly negligible. In fact, even when you ignore the people who had 2 publications or more (that’s almost 40% of the sample), the mean number of publications for men is still 16% higher than the mean number of publications for women, which I’m guessing is unlikely to result from chance.

So, although there is no difference between the median number of publication for men and that of women, it seems to me that since there is a significant difference between the means and this is not merely some artifact due to the fact that a tiny number of men in the sample had a ridiculously high number of publications when they were hired, I think the hypothesis that women get preferential treatment is not unreasonable. Of course, there are other possible explanations, so we should be careful and try to get more data.

CDJ also made another suggestion as to why the mean number of publications for men is greater than the mean number of publications for women, which I think is interesting but ultimately not convincing, given what her data say. The suggestion is that although women may get preferential treatment at the hiring stage for tenure-track jobs, this at least in part compensates for the fact that they are unfairly disadvantaged for post-doctoral/VAP/instructor positions.

Against this hypothesis, I first noted that it’s not clear at all that women are unfairly disadvantaged for post-doctoral/VAP/instructor positions. But, if women are not in fact unfairly disadvantaged for that kind of positions, a compensation at the hiring stage for tenure-track jobs could obviously not be justified by the fact that women are unfairly disadvantaged for post-doctoral/VAP/instructor positions.

Now, although women are awarded that kind of positions in a significantly smaller proportion than the proportion at which they receive PhDs, the mean number of publications for men who got a post-doctoral/VAP/instructor position is significantly higher than the mean number of publications for women who got that kind of positions. So, at best, CDJ’s data are highly ambiguous relative to the hypothesis that women are unfairly disadvantaged for post-doctoral/VAP/instructor positions.

Moreover, if the preferential treatment that women get for tenure-track jobs, assuming of course they do get preferential treatment, only compensates at least in part the unfair disadvantage they face when they apply for post-doctoral/VAP/instructor positions, then the difference between the mean number of publications for men and that for women should disappear once we ignore people who had a position of that sort before landing a tenure-track job. But, as I explained in comment #35, there remains a large difference even when we do that.

I hope this is a little bit clearer than my previous comments and that it will allow you to determine whether we disagree and, if so, on what point exactly. In any case, thanks for the reply, I think it was thoughtful and it forced me to clarify the points I was trying to make, which hopefully I did.Report

Bijan Parsia
6 years ago

Hi Phillipe,

I’m going to try a quick hit as it’s late and I have social duties 🙂 I appreciate your reply. I won’t be able to dig into the numbers more deeply until later (if at all). (i

“But, at the moment, we only have CDJ’s data. According to her data, the mean number of publications for men is significantly greater than the mean number of publications for women, so we can still ask what could explain that difference. It should go without saying that, whatever conclusion we reach on the basis of CDJ’s data, we should be careful before we accept it as true and try to obtain more data to adjudicate the facts.”

I agree with both of these.

“My point is that, although there is no difference between the median number of publications for men and that for women, it doesn’t undermine the hypothesis that women get preferential treatment. ”

I think we might be talking past each other. I generally agree that difference in mean and no difference in median is compatible with a preferential treatment hypothesis. So is no difference in either, as you pointed out. However, the dialectical context here is something like, “There’s a difference in the mean, thus preferential treatment for women is supported.” (See comment 2.) I think that such a prima facie case is weakened by there being no difference in the median. I think it cases where you have a lower bound and no upper bound, plus the probability of a very high high end, the mean is pretty unreliable for these sorts of thing.

“If I understand you correctly, your point is that, if this is why the mean number of publications deviates more from the median for men than for women and not in the same direction, then it undermines the hypothesis that women get preferential treatment.”

If you were inferring that the fact that the mean for men was higher than the mean for women was a signal for preferential treatment for men, then yes, I think the fact that you have an unbounded high end with high valued data points in a pretty standard shape for a system with “stars”, then yes, I think your inference is undermined.

“But this hardly seems obvious to me. Perhaps it makes sense if you assume that how many publications applicants have is the only factor in hiring decisions, for in that case people whose number of publications is relatively large are de facto the strongest applicants.”

This is what I take the people making the prima facie case for preferential treatment are roughly arguing. (Note it can be more sophisticated, e.g., that publications are correlated with other factors, but basically.)

“But, as you noted yourself, that’s clearly not the case. It might be that people with a lot of publications are in fact compensating for other characteristics which on the job market are weaknesses, one of which might be the fact that they are men.”

Reconsider my sample distributions. Ex hypotheses, nothing about the *decisions* was changes by the change in the upper value. The person who was changed from 10 to 20 pubs didn’t lose by gaining 10 pubs, nor did they gain. The 10 extra pubs were wasted as contributors to the decision.

“Of course, if the reason why the mean number of publications for men deviates so much from the median was that the 5 persons with the largest number of publications – who represent 1% of the individuals in the sample CDJ examined – were all men and all had a ridiculously high number of publications, so that the situation was analogous to what’s going on in your toy example, I agree that it would completely undermine the hypothesis that women get preferential treatment. ”

Ok, great.

“But, if you look at the data, you will see that it’s not at all what’s going on in the sample CDJ examined.”

Ok! This is what I was gleaning from CDJ’s post. But apparently I was wrong and I need to look more carefully. So I’m going to stop here on this until I’ve had a chance to get a grip on her spreadsheets.

By the by, I still think there’s a confusion in looking at the successful candidate cohort alone and as a whole for this sort of inference. Let’s say we had a single competition and ranked all candidates by (say) publication count and all jobs by desirability. Furthermore suppose that a fair allocation would be that that no candidate with a lower publication count gets a more desirable job than a candidate with a higher pub count. Our procedure is to assign in order with random allocation to break ties. (I.e., if A has 5 pubs and B has 5 pubs and we have job 1 more desirable than job 2, then we just randomly allocate to A and B.) It’s pretty easy to construct a population of jobs and candidates that would generate the means and medians we see in the data even if the allocation was unbiased in this way. Basically, even though the peaks of the curves are at different places, there are enough jobs that you get pretty far down the way both curves when allocating jobs.

I suspect it gets more complicated when you don’t have the single list, but people actively matching themselves with jobs they think they have a shot at and not applying for every possible job (if only for variance in AOS).Report

Bijan Parsia
6 years ago

Oops! I forgot to add before hitting submit that I very much appreciated your reply and data analysis.Report

Carolyn Dicey Jennings
Carolyn Dicey Jennings
6 years ago

Philippe, this is a nice bit of thinking and work. Perhaps we can write a blog post together at some point on these issues, if you would like. I hope at some point to look at a random sample of those looking for employment. For a start, I looked at the c.v.’s of my cohort at BU, since I know the names of everyone in that class and since we were four women and four men, which is a nice balance. One man left the program. Two women and one man found tenure-track positions. Two women and two men found fixed term positions, lectureships, VAPs, or postdoctoral positions. At the time of employment in these positions, the average publications for the 4 women was 1.5, whereas it was 2.7 for the 3 men. First employments, largely in 2012, were almost all VAPs/lectureships/postdocs, with one woman in a tenure-track job. At the time of first employment, the average publications for the 4 women was 1.25, whereas it was 1.33 for the 3 men. However, the two women with tenure-track employment also have 2 publications each in top 15 publications (according to Brooks Blog), whereas none of the men do. This is just a tiny snapshot, but I hope that it gives a sense of what sorts of things we might look for in a sample of job seekers.Report

Carolyn Dicey Jennings
Carolyn Dicey Jennings
6 years ago

Actually, one of the women has 3 top-15 publications.Report

pen wing
pen wing
6 years ago

Although I thank CDJ, Philippe and Bijan for their considered analysis of the data, I am sceptical of its value. Suppose that we establish that women philosophers’ lower hiring rate for VAPs/PDs does not explain their lower mean publication rates in TT hiring. Then what? Perhaps these questions are of personal interest to some. But it seems to me that the motivation for their discussion is generally political (NB., I am not speculating about the specific motivations of CDJ, PL or BP). Supposing that the answers to such questions are interesting for largely political issues in the profession, it would be useful to discuss two related questions:

Suppose we establish that men are advantaged in TT hiring. Then what?
Suppose we establish that women are advantaged in TT hiring. Then what?

I suspect that nothing of political interest follows from answering these two questions. As in national politics, no answer, however sound, will convince a partisan that women or men are not advantaged in hiring if they are not already convinced. I counsel a healthy cynicism and disinterest here. Nothing with any political impact follows from these statistics unless it is both clear enough to convince a grade-schooler and totally unimpeachable. The data do not meet this lofty standard.Report

Bijan Parsia
6 years ago

Hi pen wing,

You raise a very important point. To be clear on my motivations, I picked it up because 1) I disagreed with some of the analysis and 2) I’m generally interested in looking at data like this. I have a bias (or priors if you are generous :)) toward thinking that women face challenges in hiring in philosophy, though I’d be delighted to discover otherwise.

In both cases, my first answer to “Then what” is to figure out the causal story. Depending on which question has an answer “yes” and the causal story, I would make concrete recommendations. Whether anyone would pick up on these recommendations is a bit more open than I think you allow. E.g., suppose we found that implicit bias was a source of disadvantaging women. Anonymization at key stages might help combat that. This is something familiar and increasingly popular.

Now, how strong conclusions can we draw from this data alone? As we all acknowledge, not hugely. But it’s a first step. and can at least point us in some useful directions. (I’m looking a stratifying by AOS to see if different areas have different publication rates and to set up some simulations.)

If we can get some well worked out analyses, that might help us gather data that would be more determinative (e.g., we could perhaps convince at least select departments to give data on all candidates (anonymised, perhaps)).

It’s definitely the case that confirmation bias is a real danger. But it seems like we have enough people with enough different inclinations to usefully combat that.Report

Philippe Lemoine
Philippe Lemoine
6 years ago

Bijan, thanks for your reply, I think we mostly agree. The only point I wanted to make is that, since we know that a large proportion of people are hired with little to no publication, even if women get preferential treatment, it’s hardly surprising that the median number of publications for men is the same as that for women. (I should also note that, if you look at Carolyn’s data, you will see that although the median number of publications is 1 for both men and women, it came really close to being 2 for men.)

Of course, when the mean number of publications is greater for men than for women even though the median is the same, it’s also conceivable that it’s because a handful of men have a very large number of publications. But, for this to explain a difference between the mean numbers of publications as significant as that which Carolyn found, the number of publications of those men would really have to be ridiculous. So ridiculous that we can pretty much rule out this possibility at the outset, because we know that nobody goes on the market with that many publications.

In any case, even if you disagree with that, I don’t think it really matters since, as I have explained in my previous comment, the analysis of Carolyn’s data shows that it’s not what’s going on. (If you want to check my analysis, I have uploaded the updated version of the spreadsheet here: https://www.dropbox.com/l/kMwfPSq7d7zMhT96fOfdtt.) That’s why I say that, since the mean number of publications for men is significantly higher than the mean number of publications for women, we have prima facie evidence for the hypothesis that women get preferential treatment.

However, I completely agree that it could easily be defeated, not only because there is a lot of information that is not contained in Carolyn’s data, but also because our analysis of her data so far has been very crude. I think that, even in the absence of data about unsuccessful applicants, we could still get a lot more from Carolyn’s data about people who were successful by doing things more rigorously. Since I think it would be very difficult to obtain data about a large enough, truly random sample of people looking for employment in philosophy, I’m inclined to think that we should focus our efforts in getting as much as we can from Carolyn’s data.

The first thing we should do, I think, is construct a model of the way in which the academic job market works, to help us figure out what could explain the data we have beside the hypothesis that women get preferential treatment and how plausible alternative explanations are. I think that, if we spent enough time thinking about it, this could be really helpful to exploit Carolyn’s data to the fullest. If the hypothesis that women get preferential treatment still looks plausible after we’ve built such a model and used it to analyze her data, then perhaps it will make sense to try and collect more data.

Carolyn, thanks for joining the conversation, as well as for your efforts in collecting all the data without which there would have been no conversation. If you would also like to explore more rigorously what the data you have already published can tell us about gender and hiring on the academic job market in philosophy, I think it would be really cool to write a post together about that issue. But, to be honest, I’m very busy at the moment and will remain so for the foreseeable future. The problem is that I don’t even have a PhD yet, let alone a job, so I need to focus on finishing my dissertation first 🙂 However, thanks for the offer, I will definitely keep it in mind for when I have more time.

Pen Wing, you are certainly right that the question of whether women get preferential treatment at the hiring stage is mainly interesting for political reasons, but although I agree with you that so far our analysis has been too crude to have any clear political consequences, I wouldn’t be so sure that a more careful analysis of the data couldn’t have some. One of the reasons I’m interested in that issue, beside the fact that I’m interested in pretty much everything, is that I think a lot of people just assume that women are systematically disadvantaged in philosophy as if it were obvious, even though I usually find the evidence they use to substantiate that claim utterly unconvincing.

So, if you want to know where I’m coming from, my interest is not so much in showing that women get preferential treatment at any particular stage of the pipeline (I don’t really know whether that’s the case), as much as in arguing that we are not justified in believing they suffer from the kind of unfair disadvantage that I think many people take for granted. It seems to me that, no matter where the facts take us, it will have at least some political consequences. Perhaps it will not be those I imagine right now, but that’s not really a problem, as I like to think that I and a lot of other people in the profession are intellectually honest. (To be clear, I don’t mean to imply that you said I, Bijan or Carolyn were not, I don’t think you did.)Report

Bijan Parsia
6 years ago

Hi Philippe,

I’ve put a first cut of some analysis on my blog (it’s too cumbersome to post here as a comment…it almost was too cumbersome for my blog :)):
https://bparsia.wordpress.com/2014/12/29/some-philosophy-hiring-data-analysis/

I welcome comment here or over there or by email.

I agree with you on first steps. I’ll poke a little further.Report