On DeCruz's post on academic pedigree & hiring -- some more data

Helen De Cruz has an interesting and important post over at NewAPPS questioning whether the sample hiring data I provided the other day display a kind of racist and classist obsession with pedigree. In particular, De Cruz notes that in my sample, 37% of all TT hires came from Leiter top-5 programs, and only 12% came from Leiter unranked programs.

I agree that these figures are disturbing, and share De Cruz's worries about pedigree. Just last week, for instance, I learned that one of my students at the University of Tampa -- a minority male from a disadvantaged background -- got into Harvard but was unable to attend for financial reasons (viz. a single mother with a terrible credit history, and no one else willing to cosign his students loans). This student is utterly brilliant but, because of his pedigree (or lack thereof), is extremely unlikely to get into a Leiter-ranked PhD program in philosophy. 

However, although I share De Cruz's worries, I think the actual data I gathered provide some reasons for optimism. For here is the hiring data broken down by gender:

	Male	Female
Leiter Top 5	25	13
Leiter 6-10	4	10
Leiter 11-25	13	5
Leiter 26-50	8	12
Unranked	5	7

While a full 50% of male TT hires went to people from Leiter top-5 schools, only 29% of female hires did -- and a full 42.2% of female hires came from either Leiter 26-50 or Leiter-unranked schools. This suggests (but, I should be clear, only suggests) that pedigree may play significantly less of a role in TT hiring for traditional disadvantaged populations (at the very least, for female philosophers) -- though, again, I still do share De Cruz's worries about pedigree.

Comments

[B.M.]

It's not clear that the numbers suggest that pedigree plays significantly less of a role for female applicants. If you look at the top-10 schools as a group, they provide 53% of male hires and 49% of female hires.

The fact that the top 10 programs dominant TT hiring to such an extent is deeply disturbing. Do you have these same statistics from the past two years? What I'd really be curious to know is if there are any other academic disciplines where 10 graduate programs fill half of the TT jobs.

[Marcus Arvan]

B.M.: You make a good point re: the top 10 programs. However, I still think the sample suggests that prestige played a lesser role for female hires than for males.

49% of female hires came from the Leiter top 10. However, *58%* of male hires did -- which is significantly higher. Similarly, only 26% of male hires came from Leiter 26 and down, whereas 42.2% of female hires came from 26 and down -- again a significant difference. Thus, although the clustering in the top-10 is disturbing across both genders, the overall trend appears to be that pedigree played less of a role for female hires (with the caveat, again, being that this is just a sample).

[jeb]

Considering the number of unranked programs out there and similar results for previous years, it looks like maybe you should only enter one of them for a PhD if you can viscerally accept that it's much more likely you won't get a tenure track job than that you will. Almost lottery paradox unlikely.

[jeb]

I’m not sure why some seem surprised by the effects of pedigree on hiring. It’s not new, it’s not a secret, and it’s not confined to philosophy (not sure why anyone thinks it's unfair, either, but here I focus only on its ubiquity):

Daniel B. Klein and Charlotta Stern

Independent Review, Spring, 2009

 http://www.independent.org/publications/tir/article.asp?a=731

http://www.insidehighered.com/news/2014/04/08/study-traces-relative-success-those-who-earn-phds-more-and-less-prestigious#sthash.XuHDiXre.dpbs

polisci.fsu.edu/news/Placement_Rankings.pdf (look at table 1, for example)

A half hour on google will produce many similar results. Or google various departments and see where their faculty received degrees. You might argue about any one study, but the trend is robust and unmistakable. As the market gets worse, and hiring institutions have more choices, one would think this trend might accelerate.

[Carolyn Dicey Jennings]

Great post, Marcus! I put up a further analysis on NewAPPS, which had similar results: http://www.newappsblog.com/2014/04/the-impact-of-gourmet-rank-on-women-and-men-seeking-tenure-track-jobs.html .

[Rachel]

I'm sorry Marcus, but what's the p value you're using for the significance difference between 58% for the men from Top 10 departments, and 49% for the women? Those numbers may not be statistically significantly different.

[Marcus Arvan]

Rachel: the p value was .38. As you probably know, that's actually very high. In the social sciences, .1 is considered a "small" relationship, .3 "moderate", and .5 "strong."

[Abstruse]

I haven't read through the other threads, but has there been discussion of the fact that the category of people who post their jobs on Leiter is probably skewed pretty heavily toward Leiter-ranked, especially top-ranked, programs? This includes both candidates themselves but especially placement directors. Year after year, it seems, the placement directors from the top 10 or so schools post all their students' hires as the news become available and periodic summaries, with the practice seeming to drop off with ranking/prestige in certain circles. Likewise, it seems far more likely that students socialized in the top-5 environment are unlikely to be fully satisfied with the kinds of jobs most even of them are likely to get the first couple times they go on the market, so there's bound to more lateral TT movement in that demographic. I worry these and other confounding factors make the appropriate conclusion of these analyses quite weak: top-5 folks are just more likely to turn up early in the season on Leiter's thread and other public venues, without that providing much illumination into the TT market as a whole. That's consistent with the thought that they do far better on the job market, but also with the thought that the gap is narrower than one might expect, especially since some of those programs produce a lot of graduates.

[Rachel]

Waaaaiiiit, Marcus, WHAT?!

p value of 0.38 is very, very low. That shows me that the numbers are NOT statistically significantly different.

.1 is "good"
.05 is "excellent"
.01 is "gold standard"

Are you thinking r value?

[Marcus Arvan]

Rachel: yeah, sorry, my bad. That was the r-value. I haven't run the p-value (yet--I only have SPSS on my office computer).

[Fritz Allhoff]

Sorry, maybe I'm not following. If this means that it's easier for non-Leiterrific women to get jobs than non-Leiterrific men, where's the sexism? It seems women have the advantage.

This wouldn't surprise me at all if it were true. Many departments are under marching orders to hire minorities, whether gender or race-based. (Fortunately, these sort of marching orders are illegal in Michigan, where I work.) So hiring committees have to go further down the lists to find these sorts of people because they're more highly in demand. If a school wants a male, they have their choice of top-10 candidates. If a school wants a woman, they might not.

[Someone else]

"not sure why anyone thinks it's unfair, either"

Really? I thought it was pretty obvious why people _believe_ that it's unfair. Some people say that admissions to these programs often depends too much on factors that don't matter to ability. They'll also say that pedigree doesn't correlate with the factors that determine which is the best candidate for the job. You might disagree with these opinions, but surely you can't be unaware of them, can you?

[Carolyn Dicey Jennings]

Fritz,

I posted the below to your comment on NewAPPS, but I wanted to post it here, too, since your comment shows up in both places.

I think you might not be taking enough possibilities into account with these descriptions. Take, for instance, your description of what is happening for your MA program. At my low-ranked PhD institution, admissions didn't have to use affirmative action policies since women apply in such high numbers. This seems somewhat inconsistent with what you say until you realize that women may not be as inclined to apply to a program with an all-male faculty, such as yours. Given the problems of climate in the discipline, an all-male faculty might be a sign of a hostile climate. Women may be worried about just this kind of thing when they decide not to apply to your program. At least, it sounds like you have a relatively low number of applications from women.

Further, your use of the word "fortunately" and "marching orders" when referring to affirmative action policies makes it sound as though you have not considered the potential advantage of such a policy. One such advantage of diversity is that it could change the climate of a department, making it more comfortable to people of different backgrounds and social groups. Another is that it could improve the intellectual quality of a department, since diversity seems to improve problem solving and decision making in general (see, e.g., http://www.sciencedaily.com/releases/2010/09/100930143339.htm). Another, more obvious advantage is that it might re-balance otherwise unjust practices. That is, affirmative action may simply push back on biases against minority groups, such as women. If you believe that such biases exist then it seems reasonable to introduce pressure to overcome these biases (in the form of something like affirmative action), especially in cases where anonymization is difficult. How much pressure is another question, but given the other advantages I mentioned, I think the balance favors a more inclusive perspective than what I see in your comment above.