Gender & The Philosophy Job Market
“The odds of women obtaining a permanent academic placement within two years is 65% greater than men when all else is held constant,” according to an analysis discussed by Carolyn Dicey Jennings, Patrice Cobb, and David Vinson (UC Merced) at the Blog of the APA.
Jennings and Vinson do not argue for any particular explanation of this finding, but note three possibilities, the first two of which they currently have plans to investigate:
- Women philosophy graduates are more likely to find permanent academic placements because those women who would have been less likely to find permanent academic placement are more likely to leave the discipline before graduating than men in the same position…
- Women philosophy graduates are more likely to find permanent academic placements because women are more likely to specialize in areas sought by hiring programs…
- Women philosophy graduates are more likely to find permanent academic placements because hiring programs have a preference for hiring women, all else being equal.
The analysis is based on data previously discussed here.
Read the full post here.
I would have thought that 2* and 3 both deserve some investigation. 2* is not that women are more likely to specialize in areas sought out by hiring programs, but that women are outperforming their male peers regardless of area of specialization. They are just better candidates to hire. This seems at least possible to me. All of the programs I have been in have disproportionately large numbers of men. They also tend to have women who vastly outperform men. (Then some further explanation for this outperformance would need to be provided. But women are outperforming men in many areas of education, including higher ed, so this seems at least worthy of investigation).
3 also seems possible. The explanation is as follows. Almost all departments have disproportionate number of male faculty. This is increasingly seen as a problem. There is therefore pressure to higher at least 50/50 men/women, if not more like 35/65 men/women. The pressure for these hiring numbers is drawing from a pool that is disproportionately male. So if you wanted gender parity in hiring (or better) and are drawing from a pool of candidates that has disproportionately large numbers of men, you would expect women to do better than men in the hiring process.
Just to be clear, I think it would be perfectly fine and good if both 2* and 3 were true.
As I argued on this blog recently (http://dailynous.com/2016/04/15/philosophy-placement-data-and-analysis-an-update/), the first hypothesis is unlikely to explain the data even if it’s true and, as far as I can tell, the only data we have for the moment that directly bear on it speak against it. If people are interested in the details, they can read the comments on Carolyn’s guest post whose link I give above, particularly my discussion with WP.
As for the second hypothesis, since the regression analysis whose results are reported on the APA blog controls for AOS and there is still a large effect of gender, it seems to me that it can already be ruled out. Carolyn and her co-authors on the APA blog seem to acknowledge that, but they make this point which I find rather mysterious: “Although our analyses accounted for first-reported area of specialization, they did not account for the area of specialization sought by the hiring program. To test whether hiring AOS helps to explain the gender effect, we intend to match our placement data to the job ads from the same time period.” I may be missing something, but since the regression already controls for AOS, the coefficient of the dummy-variable for gender should already factor in any effect that a difference in specialization between men and women might have.
Another thing, which has nothing to do with gender, that surprises me is that the effect of AOS found by the most recent analysis is completely different from that found by the previous one. I know that some of the coefficients were not significant back then and most still aren’t in the latest analysis, but given that I don’t imagine the data that have been added make up a large proportion of the total, I still find that a bit surprising. I only skimmed through the supplemental material on Patrice Cobb’s website though, so I may just have missed the explanation.
In my comment, “since the regression already controls for AOS, the coefficient of the dummy-variable for gender should already factor *in* any effect that a difference in specialization between men and women might have” should read “since the regression already controls for AOS, the coefficient of the dummy-variable for gender should already factor *out* any effect that a difference in specialization between men and women might have”, as you probably had guessed. The point is that, since the analysis controls for AOS, I don’t see how the effect of gender could be confounded by AOS, as the second hypothesis would have it.
I’m pretty far out of my element here, but let me take a stab at what they meant by the mysterious claim about AOS.
Suppose we control for AOS listed on the candidate’s CV and find what they found. But suppose we didn’t control for AOS listed on the job advertisement. Aren’t two possible explanations for the difference in likelihood (a) the (supposed) fact that women philosophy graduates are more likely to have AOCs in the AOS listed and (b) the (supposed) fact that women philosophy graduates are more likely to have research interests within their reported AOS that are more attractive to hiring programs? Scenario (a) doesn’t seem totally implausible — people regularly apply for jobs with an advertised AOS they don’t have. Scenario (b) seems less plausible but still worth exploring.
So, if I understand you correctly, (a) is the hypothesis that although AOS doesn’t confound gender, AOC might and (b) is the hypothesis that, if we used a more finely-grained classification of AOS, the effect of gender would go away. I think they had something else in mind though, since in order to test (a) or (b), we don’t need to look at what AOS were listed in ads, yet on the APA blog they seem to be suggesting that we do something like that in order to test their second hypothesis.
Yeah, that’s what I had in mind!
I agree that what they said doesn’t seem to mesh well with what they said. However, on hypothesis (a) shouldn’t we expect AOS listed in ads to correlate with AOC (on CV)?
Yes, I don’t think (a) is plausible, at least not if (b) is false, precisely because AOC must be highly correlated with AOS.
As for what they had in mind, Carolyn explained it below. It turns out that applicants could report more than one AOS and, if I understand correctly, she and her team only used the first on the list to assign AOS-categories to the applicants. However, I don’t think it can explain the results of the regression, because I doubt that a lot of people have AOS that fall in more than one of the 4 broad categories used in the regression.
That leaves (b) as a possibility, though given the magnitude of the effect of gender, it would be remarkable if it could explain all of it. Of course, for all we know, it might not explain any of it and it could even be that, if we used a more finely-grained classification of AOS, the effect of gender would actually come out as even larger. Hopefully they will test that hypothesis soon or release the data so people can do it themselves.
I think you’re underestimating how many people have AOS/C that cross these categories. Some common combinations: epistemology and value theory, language and value theory, language and logic, history and whatever they’re interested in the history of, feminist philosophy (which I think falls under “traditions”) and value theory…
Using a more fine grained AOS classification seems important, given that there’s at least one wrench that we know is out there: women are far and away more likely to have feminist philosophy as an AOS/C, which is some years pretty popular in job ads (see http://www.newappsblog.com/2015/10/tracking-the-job-market-a-start.html).
It is interesting that the size of the gender effect seems to vary between AOS categories (see http://www.patricehazam.com/#!supplemental-material/c1mqf). It looks a lot smaller in value theory, especially compared to LEMM.
“I think you’re underestimating how many people have AOS/C that cross these categories. Some common combinations: epistemology and value theory, language and value theory, language and logic, history and whatever they’re interested in the history of, feminist philosophy (which I think falls under “traditions”) and value theory…”
It may be so, though I doubt it could significantly affect the result, because not only would there have to be a large proportion of people with AOS in more than one category, but moreover it would also have to be the case that the distribution of men and women across categories differ substantially depending on which AOS you use to assign categories and I would be amazed if that were the case. In any case, it should not be difficult to test that hypothesis, so again I think it would be great if the data were just made public as soon as possible.
“Using a more fine grained AOS classification seems important, given that there’s at least one wrench that we know is out there: women are far and away more likely to have feminist philosophy as an AOS/C, which is some years pretty popular in job ads (see http://www.newappsblog.com/2015/10/tracking-the-job-market-a-start.html).”
I agree that using a more fine-grained classification of AOS would be a welcome development. I wouldn’t be surprised if that reduced the effect of gender, though I also wouldn’t be surprised if that increased it. But I don’t think we should worry at all about what ads are saying. Even if there are a lot of ads that mention feminist philosophy and the departments posting those ads were really looking for someone who does feminist philosophy, which may or may not be the case, it could also be that specializing in feminist philosophy reduced one’s chances for jobs whose AOS are open, since even when a department post that kind of ad it still has AOS-based preferences. Of course, the same thing is true for every AOS, not just feminist philosophy, which is why we should focus on the AOS declared by candidates and not what ads say. However, I take it that you were just offering one reason to think that using a more fine-grained classification of AOS might reduce the effect of gender, not arguing that we should turn our attention to ads instead of candidates.
“It is interesting that the size of the gender effect seems to vary between AOS categories (see http://www.patricehazam.com/#!supplemental-material/c1mqf). It looks a lot smaller in value theory, especially compared to LEMM.”
I agree that it’s interesting and should be explored further. In fact, among candidates who specialize in value theory, the proportion of men who found a permanent position is slightly higher than the proportion of women who did, though the difference isn’t significant. (The proportion of women who found a permanent position is much higher in all the other categories, though the difference is only significant for LEMM. We also don’t know what things would look like if we controlled for year of graduation for each category separately, though I doubt it would change things much.) It’s still the case that, when we control for AOS-category and year of graduation, women have a substantial advantage, but the fact that it shows a lot of variation across categories is interesting. In fact, I think it’s the only thing anyone has pointed out so far that should reduce our confidence that employers have a preference for women, though I would advise caution until we understand why the effect of AOS-categories has changed so radically since the last revision of the data. As I noted in my first comment above, that is pretty surprising. Again it would be great to have the data so that everyone can perform their own analyses.
(By the way, I think there is a mistake in the numbers reported on Patrice Cobb’s website, because some of them seem inconsistent. For instance, on figure 2 the proportion of men who found a permanent position seems to be around 35%, while the proportion of women who found a permanent position seems to be around 48%. But the proportions that one can derive from the information in tables 2 and 3 are 38% and 36% respectively. I think it’s because the women who found a permanent position but whose AOS was N/A wasn’t added in the last column for that same row.)
I think this is pretty simple. Candidates have multiple areas of specialization. We looked at the first-listed AOS. The hiring program need not have been interested in the first-listed AOS, but some other AOS, and perhaps even an AOS not listed by the candidate but for which the candidate was deemed suitable. (A further point: we actually controlled for AOS category, which is more rough-grained than AOS. I am not sure if we will look at AOS or AOS category when we compare to job ads.)
Oh I see!
I didn’t know candidates could list more than one AOS. I just thought they were asked to pick one of the 4 categories. Do you know what proportion of candidates declared AOS in more than one category? The fact that you only looked at the first-listed AOS is unlikely to have affected the result if, as I imagine, the vast majority of people had all their AOS in one category.
I also didn’t know that you had data about AOS that were more finely-grained than what you call “AOS category”. As I said, I have only skimmed through the report, mea culpa. When indicating their AOS, were candidates able to write anything they wanted, or did they have to choose among a fixed list?
AOC also seems a very important thing to look at. In my recent couple years on the job market, I was somewhat surprised to see how in demand AOC like feminist philosophy, philosophy of race, non-Western philosophy, environmental philosophy, all sorts of history, and Continental philosophy were (surprised because at my extremely ahistorical, analytic PhD program, it was not possible to take a grad course in some of these topics, and most of us did not develop AOC in them).
It also seems plausible to me that women might have more of these in-demand AOC than men and that this difference might very well have an impact on the market. Think of the SC at a liberal arts college who is looking at 2 roughly equal candidates with an AOS in say, phil language, but that only one of them has an AOC in Chinese Philosophy, feminist philosophy, philosophy of race, or something else that might be in-demand among undergrads. So there is a 4th hypothesis about the gender difference: women are more likely than men to have a desirable combo of AOS & AOC.
From the armchair, this one kind of seems like it’s got to be a combination of the three.
I would also caution anyone against thinking that the third hypothesis points to some kind of structural injustice. Obviously we are trying to increase diversity, obviously an all-things-equal preference for women is a good means to that end. I think that more left-leaning folks often worry (for some reason) about admitting that things might be improving, but hey, we should just admit that they are improving when they are.
Perhaps also
(3*) Women philosophy graduates are more likely to find permanent academic placements because hiring programs have a preference for hiring women.
should be considered.
Is it supposed to be controversial that most hiring committees these days have a preference for hiring women? I would have thought that this is completely obvious to anyone minimally engaged in the profession. It is also certainly the simplest explanation of the hiring data.
In one sense, (3) is perfectly fine: if literally all else is equal, *something* must be the deciding factor, and if that something is the sex of the candidate, then so be it. On the other hand, affirmative action is (at best) a silly concept, so I’m a little hesitant about it. (Not to mention it would suck to be a male in a crappy job market and see that (3) is true…)
While it certainly would suck to be a man in a crappy job market and for 3 to be true, one may still recognise there are good reasons for adopting a policy like 3. I say this as a man who suspects that something like 3 is at least part (almost certainly not all) of the explanation of the data that the researchers have found. The truth is that our discipline is so messed up that no matter what we do, some people will be hurt. It seems to me that something like 3 is better than not attempting to counteract the massive structural injustices of our discipline, despite the fact that it effects people like me.
Well, I’m not inclined to think that our discipline is “so messed up” – there is certainly a disproportionate amount of men in the field, but I’m not inclined to view that as a great – or inherent – injustice. It is an injustice if women are being rejected on the basis of their sex, and perhaps that’s the case some of the time, I’m just not convinced it’s true in most cases. Anyway, I simply cannot support a policy of discrimination, even if its intentions are good.
All that said, I do partially sympathize with your reasoning, but cannot support it for the above reasons.
Surely representation in the field isn’t the only way in which things can be messed up, though. It’s messed up that women in philosophy experience such a high level of sexual harassment. It’s messed up that women in philosophy end up being asked to do all kinds of extra service because women need to be represented on various committees and there are so few of us to go around. It’s messed up that there is so much additional pressure on women not to screw up in the minor ways that we will do, because of the justified fear that when we mess up it will be seen as evidence that women aren’t good at philosophy, not that an individual made a mistake. Etc etc etc. And those are all things that there are some reasons to think would get better if there were more women in the field, and in particular departments.
Those all strike me as good reasons to make gender a deciding factor when something has to be. *Maybe* it would be objectionable if gender were doing more work than breaking a tie – but why think that is true? We *know* that there are far, far more highly qualified people on the market than there are jobs. We also know that past a threshold it’s incredibly hard to make fine merit distinctions. I would imagine that hiring committees are very often in the position of having to decide between several extremely qualified candidates, any of whom they would be delighted to have. And if that’s the case, then hiring a woman might help in all of the ways mentioned above in which things are indeed messed up. Sure, it would suck to be a man who doesn’t get a job in that case. But it also sucks to be a woman in a department that is 80% male.
I’m not really sure how that is relevant to what I said. I mean, I conceded that if X and Y are equal, and we need a deciding factor, then sex would probably be an acceptable candidate for it. This should not be a policy, however, because we should not implement discriminatory policies. That is not right.
“Sure, it would suck to be a man who doesn’t get a job in that case. But it also sucks to be a woman in a department that is 80% male.”
This is just silly. I’m sure that all the men who are unemployed will be very comforted to hear this. It also depends what department a woman is in…if a woman is in a good (non-creepy) philosophy department, I don’t see why her being part of the 20% would necessarily suck. I mean, perhaps it sucks to be a continental philosopher in an analytic department, but I don’t think such ‘sucking’ is significant, and I certainly don’t think that the employed philosopher should complain to her unemployed counterparts. I also don’t see why we should just assume that it sucks to be part of a minority. Did it suck to be Richard Swinburne when he was at Oxford and was one of the like 5 substance dualists in the world? Perhaps, but not in a significant way. Again, this is not to say that some women are harassed or mistreated by their philosophy department, it’s just to say that we shouldn’t’ assume it to be the case if they are a minority.
What is the relevance of whether unemployed men (or unemployed women) will be comforted by a particular rationale. The present job market is guaranteed to screw over large numbers of applicants. Thus the fact that some candidates will reasonably feel screwed over is just not eligible as a reason to object to any given policy. Now you can certainly adopt a deontic principle which requires gender-blindness in all decisions, even when that will do nothing to effectively promote fairness or minimize overall harm (indeed, even when they predictably lead to even greater systemic unfairness); but the reasons for adopting a principle simply cannot be based upon its predictable effects.
I’m not advocating that we adopt any principle. I’m saying that it isn’t right to discriminate, and that we shouldn’t construct discriminatory policies.
That sounds like a principle.
I am the only woman in my department. It does ‘suck’. It’s a very strange feeling sitting around a table discussing a paper and being the only woman there. I also notice subtle ‘boys club’ behaviours where the men know more about what each other is working on, making plans outside the office, etc. I came from a place where women were represented much more equally so it has been an adjustment.
Why does that suck and wjhy is it strange? I don’t see why one would even care about the sex of those àt the table so long as you’re talking about philosophy – unless your department is full of creepy people. However, I suppose I can understand it in one sense: I can sort of grasp how it would have sucked to be Swinburne when discussing Phil of mind with his peers since 99% of them probably didn’t take his ideas seriously. Is that sort of what you’re getting at in your situation?
Dude. She just told you why it sucks.
No she didn’t. She said it sucks and then said “also…”and listed some stuff. …”dude.”
“Those all strike me as good reasons to make gender a deciding factor when something has to be. *Maybe* it would be objectionable if gender were doing more work than breaking a tie – but why think that is true?”
Do you realize how many job competitions there are in which a tie-breaker is needed? There is a huge pool of immensely qualified applicants for every job. It’s very hard for the best people to weigh them up fairly on the basis of the best available evidence. Now, if you want to flip a coin or roll a die each time so that the winner is chosen from the best without bias, then great. But even to make it a policy (de facto or de jure) to give the position to a woman as a tie-breaking heuristic would make something like 80-99% of the hires female.
But I suspect that wouldn’t bother you, because you think that the male candidates applying for the same jobs had their chance by proxy when the male parents of the now-privileged female candidates had it easy, even though they were non-identical with those males. But still, genitals, so those good, highly qualified, unemployed men who won’t have a chance if you implement a tie-breaker-goes-to-women policy can go screw themselves! Right?
Look, the point was that fairness to job candidates is not the only thing that matters – needs of departments matter too. And one way in which many departments with few women (although the same will hold for people of color and other groups as well) would be well-served would be to make those departments less shitty for their current female members. This is so for all of the reasons mentioned above that I won’t go over again. And it’s certainly open to men to say that they don’t know why it should suck to be the only women in a philosophy department. Cool, bro – it’s the kind of thing that is probably indeed hard to understand if you have never been in a position in which you were a) part of a very small minority in your profession and b) in which there was a significant amount of both implicit and explicit bias directed at members of your group qua members of your larger profession. You probably have not been in this position, and epistemic humility should suggest that you might want to listen in this case to those who have.
As for whether using gender as a tie breaker would lead to 80%-99% of hires being female? Obviously it wouldn’t, a) because I was claiming that gender would be a permissible thing to use as a tie-breaker, not an obligatory thing, and b) because even if 80%-99% of first offers were to women, 80%-99% of hires would not be. I don’t work in the most desirable department in the world, but I work in one that is still I think more attractive than the majority – and we consistently end up hiring our second choice candidate instead. I’m sure we’re not the only ones.
In the end, am I worried about the hiring process being unfair to some candidates? Yes, of course I am. Do I think that that is the only thing to worry about here? No, I don’t.
As someone in the same position, I completely agree (though I guess I remain agnostic on which factors explain the current data). Many departments still have very good reason to want to diversify their department in terms of gender (and other factors!). The idea that such preference is unjustified because it is unfair is undermined by the fact that *any* way of selecting candidates will lead a significant percentage of candidates to be harmed through no fault (or professional deficit) of their own. I believe the reasons Anca Gheaus provides for not considering gender-based reasons for conference invitations unfair can be extended to the job market, with suitable adjustment for the different context: roughly, we routinely accept the inevitability of any number of nonmerit factors influencing the outcome, so it’s not clear why this one should be different. http://www.academia.edu/2436536/Three_Cheers_for_the_Token_Woman_
Derek Bowman,
At this point, I’m having a hard time discerning what’s even motivating your side of the discussion. At first I thought that what you were after was a way to ensure that the hiring process is fair and that a candidate’s gender is left out of consideration. But at 5:35, you rejected “a deontic principle which requires gender-blindness in all decisions” on the grounds that, you feel, it would “lead to even greater systemic unfairness”. That seems strange to me, but I thought, OK, you just have different nonnormative beliefs and are just trying to find some way to ensure that, whether or not search committees follow the deontic principle of gender-blindness, they in fact hire women at rates equal or greater to their proportions in the applicant pool. But that can’t be it either, since we already have not only achieved that but have gone way beyond, and you’re still here arguing for your conclusion. Then I thought that perhaps your aim was to reduce the number of nonmerit factors influencing the outcome of hiring decisions, but you’ve now just openly said that you accept the inevitability of that sort of unfairness and don’t really care about it.
Could you please walk us through what you’re after? You want departments to be more ‘diverse’ on the basis of sex or gender (you don’t seem to be as interested in promoting diversity based on other things like political viewpoint, or at least you never seem to mention it), but why? Not, it seems, for the familiar reasons. You want more women, but not in order to correct a hiring bias (which goes the other way now and has for years), not because you’re principally concerned with merit, not because you want gender-blindness in hiring, not to ensure that women are hired at their ratios in the applicant pool, and not in order to reduce the number of nonmerit factors in hiring. So why? What, on your view, is the actual argument that supports what you want? And how is your argument any more compelling, objectively, than the argument of a sexist, old-fashioned male who wants to keep hiring men for some antiquated reasons? Most of us hate traditional male-oriented sexism because it’s unfair and because it unjustly harms people of one sex, but you seem not to care about those things and hence the traditional anti-sexism reasons are not open to you. Please explain.
Surprised Not Surprised,
Sorry if I was unclear. I’m starting from the assumption that there are good reasons for the discipline and for departments to increase the representation of women among tenured/tenure-track faculty. In light of that assumption, I’m offering an argument against the understandable, but mistaken claim that acting on those reasons by having a preference for women candidates is wrong because it is unfair.
I think your difficulty is that you assume the only issues of justice and fairness at play are ones that work at level of discrete (or aggregated) hiring decisions. You’re right that I don’t think that’s the right level of analysis. I don’t think there is any way of running a search in the current job market that will not predictably result in undeserved harm to many individual candidates. There are nonetheless morally better and worse ways to make those decisions – such as not causing gratuitous harm, or not causing harm that also contributes to larger structural forms of injustice. But I think many of our usual reasons for avoiding bias in discrete hiring choices are based on the contribution of those decisions to larger forms of systematic injustice, and I see no evidence that a small bias against men in hiring does this.
If it’s true that there is a pro-woman bias in hiring that is due to explicit pro-woman preference, I don’t see how that undermines the general reasons for greater representation unless that has already done more to equalize the representation of women in philosophy faculty overall. Tenure-track careers are long, so how many years exactly do you think this effective pro-woman bias has been in effect?
But even if we restrict ourselves to the question of parity in hiring, I find the argument strange. Imagine a community that has demanded an increase in police patrols for years because they have a higher crime rate than neighboring communities. After these calls are finally heeded, new statistics show that community now has a slightly below average crime rate. It would be absurd to respond to this by returning police patrol to their former level. At best, then, these statistics would be a reason for slightly less use of (or less weight to) a preference for hiring women by hiring departments.
Thanks for the reply, Derek.
I do understand that you’re “starting from the assumption that there are good reasons for the discipline and for departments to increase the representation of women among tenured/tenure track faculty.” However, what I’m questioning is what your basis is for that assumption. To be clear, I can understand that someone could legitimately have that assumption because he or she thought that the current hiring trends are unfair. But that doesn’t seem possible in your case, since you seem to admit that you’re not concerned about fairness.
So could you please set out why *you* feel the assumption is warranted? What, in *your* mind, are the good reasons to increase the numbers of women? They don’t seem to be the familiar moral reasons.
Your final metaphor about police patrols is helpful on the more limited issue you seemed to think I was asking about, but even there I’d like to look more deeply into it. If Community C has until recently had a high crime rate and now, thanks to extra police patrols, has a remarkably low crime rate, that isn’t yet a sign that the extra police patrols should be completely removed from Community C. But there are clearly limits. For one thing, imagine that nine other communities have to share the same police resources. Wouldn’t it be right for people in the other communities, or really any fair-minded person, to suggest that the extra patrols be scaled back a little for a few months just to see whether the problem disappears? Also, if Community C now has by far the lowest crime rates, but the advocates of Community C were still out there loudly and aggressively advocating at all times for even *more* of the police budget to go toward Community C patrols, regardless of what the evidence comes to show, then that would seem to be a good sign that something unhealthy is going on with their thinking.
Another way of putting this is: what are your falsification conditions? What empirical data would lead you to say “OK, we’ve gone far enough with this thing, now let’s ease off a little”? I think many of us are worried that, for many of the more vocal advocates, perhaps including you, it’s a moral imperative to tip the scales further and further in favor of more and more women in philosophy, regardless of what the evidence shows and regardless of what else needs to be jettisoned to achieve this end. And it would be good to hear from you what would be sufficient to tip you off that things have gone far enough and that it’s time to dial it back. Because for many of us, we’re past that point. What’s the point where that would kick in for you? Thanks.
I don’t know why you continue to insist that the familiar moral reasons are only those that apply at the site of hiring decisions. Existing gender imbalances contribute to an environment that is conducive to overt sexism and sexual harassment. They impede our ability to notice and respond to burdens of care that fall disproportionately on women. They contibute to forms of stereotyping and implicit bias that are as detrimental to our students as they are to our colleagues. They compound and are compunded by larger social disparities in pay and social status for women.
If these don’t sound like familiar moral reasons, you would be better served reading some real feminist scholarship than focusing on the brief points I can cram into a comment box.
Hi again, Derek.
You didn’t answer my question: what sort of evidence would, on your view, be sufficient to make the case that it’s time to dial things back on this issue? Comparatively, under which circumstances would you admit in your own hypothetical case, that the time had come to stop giving preferential treatment to the community that now has a lower crime rate than the rest? Thank you.
I have never insisted, even once, that “the familiar moral reasons are only those that apply at the site of hiring decisions”, so I don’t “continue to insist” that. But if I might outline your reasoning as you suggestively present it, you’re reasoning
1. The mere fact of having more men than women working in philosophy, whatever the reason for it, leads to overt sexism and sexual harassment.
2. The mere fact of having more men than women working in philosophy, whatever the reason for it, makes it difficult for the rest of us (including you) to notice that there are burdens of care that fall disproportionately on women, and makes it difficult for us to respond to those burdens.
3. The mere fact of having more men than women working in philosophy, whatever the reason for it, contributes to forms of stereotyping and implicit bias and thereby harms our students and our colleagues.
4. The mere fact of having more men than women working in philosophy, whatever the reason for it, leads to larger social disparities in pay and social status, to the disadvantage of women.
5. These harms to women are not only clear and significant; they are in fact so significant that other values and interests and goods can and should be jettisoned in order to secure even a speculative chance of reducing the male to female ratio in philosophy, regardless of the costs to anyone else and regardless of any apparent evidence that the system already privileges women over men by a significant extent.
I find more or less all these premises dubious and in serious need of support. I find 3 somewhat plausible, but there are real questions even there about how much harm is involved. The rest seem flat-out dubious.
Also, it’s interesting that you assume that I haven’t read some “real feminist scholarship.” I have indeed read a considerable amount of it. I find that a few feminist thinkers have made important points that others have neglected. However, feminist scholarship, even at its best, is a partial view of a very complex set of interactions and factors. At the risk of sounding like someone trying to rescue others from a cult, I recommend that you have a look at some non-feminist scholarship on the same issues. The world is much messier, and has many more things whose value we ought to preserve, than are dreamt of in feminist philosophy.
1. On your interpretation of me: “The existing gender imbalances” =/= “the mere fact of having more men than women”
2. I don’t know how else to interpret your repeated claim that I don’t seem to endorse any of the “familiar moral reasons” for opposing gender disparity just because I’m not terribly concerned with isolated bias in particular hiring transactions.
3. I also think I’ve made it plain what would be relevant evidence. Either evidence that the recent hiring trends have in fact eliminated serious gender disparities in tenured/tenure-track faculty positions (*not* in last-year-hiring for those positions) or evidence that gender disparities as they exist (*not* as imaginary, abstract, isolated variables) don’t have the sorts of deleterious effects I’ve referred to.
4. But seriously, while I appreciate the self-examination provided by your attempt at playing the Socratic gadfly, you’re wasting your time on me. I’m not on any hiring committees, nor am I likely to be on any anytime soon. Nor am I an expert on feminist scholarship. I think I’ve more than explained why I hold the position I do. If you’re not convinced that’s fine – the evidence for this is complex; I’ve mostly just reported to you my reasons, not argued for them. Even if I were an expert – which I’m not – I couldn’t offer a full defense in this forum. If you’ve read the feminist scholarship on these issues then engage with the serious, professional, peer reviewed works by experts. Not the current best judgments of a blog commentator who is happy to have a distraction from end of semester grading. So by all means, feel satisfied with your own judgment that you’ve defeated me. That’s not a substantial accomplishment either intellectually or practically.
I think we need to pay close attention to the fact that “all other things being equal” is quite loaded. We have seen that what it takes for a woman to be seen as sufficiently educated, prepared, and productive is quite different from what it takes for her male competitor to be seen as such. Surely implicit biases play a role…and they play a role all the way back…
H, it’s true that we’ve seen *some* evidence that what it takes for a woman to be seen as sufficiently educated, etc. is quite different from what it takes for a similarly qualified man. However, we’ve also seen evidence that things go in the opposite direction. It’s a complex issue, and if anything, the final results from the CDJ data seem clear where the most pressure lies at present.
I’m all in favor of identifying likely implicit biases. But we should look at them on both sides, and not get carried away, as I think many of us do, in buying into a grand narrative that compels us to believe in a network of dark and overwhelmingly powerful forces behind everything even when there is strong empirical evidence to the contrary. Your comment “Surely implicit biases play a role… and they play a role all the way back…” sounds like another way of saying that whatever evidence is presented against the ‘women need way more help in the job market’ hypothesis, attention will be shifted to these implicit biases even though the evidence is against their playing a decisive role against equally qualified women getting the jobs; and if that point must be conceded, you will retreat to ever deeper and deeper suspicions of implicit bias, “all the way back.” At what point, if any, is it time to say that things are pretty well as they look, and it’s time to dial things back a little instead of constantly ramping them up?
Where do we see evidence that things go in the opposite direction, Surprised Not Surprised?
The final results from the CDJ data might not take into account that “all things being equal” wasn’t the case from the start. That is my suggestion.
What would be really interesting is to get data on how many SCs blind CVs and the like before reviewing them.
Maybe we could just borrow from other “disciplines” and apply post hoc fantasies that align with our biased experiential knowledge and posit an intrinsic, motivated discriminatory force that pervades every facet of society and culture (at every level all the time) to simplistically explain disparate representation in something as heavily nuanced as profession and career choice in the modern world…
But that would be irresponsible wouldn’t it?
I won’t how the “confidence” hypothesis is supposed to work. I can think of two models for it.
Model 1: Confidence is dynamic. Everyone starts graduate school with and initial amount of confidence (Ci). Everyone has to have a Ci > x, where x is some minimum amount require for one to pursue a Phd. Let x = 100 confidentiles for simplicity. As graduate school progresses some students gain confidence and others lose it. In general women are more likely to lose confidence due to worse treatment. Many women (moreso than men) find their confidence dropping beneath 100 and hence leave the program. Those women with high Ci’s remain and those women are more likely to succeed on the market compared to men with lower Ci’s, who were able to remain because grad school did not reduce their confidence as much.
The problem with Model 1 is that it’s only focused on initial confidence. However, it seems obvious that it is the amount of confidence at the time of performance that improves/harms performance. We have no reason to think that the women on the market have more confidence (at the time they go on the market) on average than men on the market. They have higher initial confidence, but they also have lost more confidence in grad school by hypothesis.
Model 2: Confidence is static. Instead of confidence fluctuating throughout grad school, what fluctuates is x– the amount of confidence needed to remain in the program. Due to disparate treatment in grad school x is gender dependent, so there is a xm for the confidence a man needs to stay in the program and a xf for the confidence a woman needs to stay in the program. xf > xm, again due to disparate treatment. Thus the women who make it to the market have on average more confidence than the men who remain.
Unlike Model 1, this model yields the result that women on average have more confidence than men at the time they are on the market. However, it seems unlikely to me that confidence is static in the way the model assumes it is.
I’m sure something this complex can be explained in more than the two ways I suggested, but those are the only ones I could think of. I’m hoping someone could explain what is the theory behind the confidence hypothesis, because I suspect I’m missing something.
It would be nice to have more details on the model that was used for the analysis. The equations on Patrice Cobb’s website (http://www.patricehazam.com/#!supplemental-material/c1mqf) don’t make a lot of sense to me. If I understand correctly, they allowed the intercept to vary for each PhD-granting institution, but not the coefficients of the predictors that appear in the regression model used at the individual level. As far as I can tell, this makes sense for gender and year of graduation, but probably not for AOS. Looking at the previous report, I’m also under the impression that it wasn’t the case before, which might explain why the results were so different for AOS in the latest analysis. But I can’t really tell because the material on Patrice Cobb’s website also doesn’t explain what the independent variables in the equation for the regression model at the group level are. The previous report is no more helpful on that issue and, as a result, I have no idea what the model exactly is. I suspect the equations on Patrice Cobb’s website contain typos because different coefficients are used for the same independent variables both at the individual and group levels. I think it’s because some of the subscripts for the independent variables are messed up. Also, while in that case a multilevel regression model definitely strikes me as the best choice if the data satisfy the conditions for that kind of model to be reliable, it would be nice to have a brief discussion of whether that’s indeed the case. I’m not a statistician, which I’m sure is part of why I’m a bit lost here, but usually I can figure things out if I have enough information about the model used. I would be very grateful if one of the authors or someone who understands what’s going on better than me could explain to me what the model they used is exactly.
Gender Bias is prevalent and is here to stay unless there is transparency in job environment, which again is highly unlikely!
If I understand correctly, hypothesis 3 actually does make some testable predictions. Specifically, it seems to predict that the advantage of being a woman should be greater for those AOSes in which there are relatively few women, while being lesser for those AOSes in which there are a large number of women.
After all, suppose that the search committee is hiring in one specific AOS and has a preference to hire a woman. If you are one of the few women in that AOS, this preference should confer quite a substantial advantage on you, whereas if there are many other women in this AOS, it should confer a smaller advantage.
(In any case, I should say once again that I am very grateful for and impressed with the work that Carolyn and her colleagues have done on this project.)
Late to the conversation here, but it seems to me that there might be an additional hypothesis that might (partly) explain the gender hypothesis: typical woman philosophers might have better social skills than the typical male philosopher, and this might give women an edge in the “collegiality” category.
I’m not certain whether this hypothesis will hold weight, but I certainly think that *on average*, even very nerdy women receive much more training in basic social skills than their very nerdy male counterparts. And at on-campuses, you are being evaluated as a colleague they might potentially be stuck with for decades, so these things matter a lot, things like: asking people about others’ work and not just talking about your own, making small talk effectively and in a way that shows you have basic empathy, *not* reacting in an overly hostile or defensive way when your work is challenged in Q & A but rather incorporating feedback graciously, etc.