Thread by 🐈⬛👻 Dan but with zombies 🎃🐈
- Tweet
- Oct 1, 2022
- #Socialphilosophy #Academy
Thread
To celebrate 🎃October🐈⬛ I'm thrilled to share with you all this paper with @diceyjennings and @pcontrerask. onlinelibrary.wiley.com/doi/10.1111/meta.12591 It's a long-ass paper, so you get a long-ass summary 🧶 1/30
All the data we used comes from @APDAtweets ' surveys of philosophy PhDs and job market placement records, 2012-2019 (when APDA has the best coverage) 2/30
Study 1: Cluster analysis. APDA surveys students, asking them (a) their own AOS and (b) what keywords characterize the research in their PhD program 3/30
We used these to cluster programs together. A 2-cluster solution seems to correspond to the Analytic/Continental divide. But! 4/30
A 3-cluster solution fits the data slightly better. The third cluster? Philosophy of science onlinelibrary.wiley.com/doi/10.1111/meta.12591 5/30
We also examined a 6-cluster solution for comparison. Here we can distinguish a certain kind of program in phil religion + medieval (Notre Dame, Baylor), and feminist+applied phil. But! 6/30
The 6-cluster solution also includes a "random cluster," some programs with nothing in common. Clusters can be overinterpreted, and we should keep this in mind w/ 2- and 3-cluster sol'ns. 7/30
Btw average gender distribution is about the same, 28%, for recent PhDs from all three of Analytic, Continental, Phil Sci onlinelibrary.wiley.com/doi/10.1111/meta.12591 8/30
Study 2: Prestige. @Helenreflects argues philosophy suffers from prestige bias. Can we identify a prestige hierarchy using APDA data? 9/30
We used the hiring network data: PhD program --> hiring department. Similar to methods used by @aaronclauset et al in several cool papers: (1) advances.sciencemag.org/content/1/1/e1400005 (2) www.pnas.org/content/early/2019/04/24/1817431116 (3) www.nature.com/articles/s41562-022-01425-4
10/30
10/30
"Prestige" was identified using 2 methods. 1 is Fancy Math (eigenvector centrality); the other starts with NYU, asks who they hired, who those programs hired, etc., until we reached a closed subnet 11/30
Both methods agree, picking out the same set of 83 "high-prestige" programs onlinelibrary.wiley.com/doi/10.1111/meta.12591 12/30
Does prestige reflect genuine differences in quality? APDA data can't tell us this, but we review the literature arguing that prestige ≠ quality 13/30
We don't include a list of the high-prestige programs, btw. There's a figure that shows some, but not all, of their names. We're not interested in creating a new ranking scheme for academic philosophy. There be dragons. 14/30
Study 3: Hiring effects. Take the clusters (study 1), prestige categories (study 2), country of PhD program, size and diversity of PhD program, plus individual gender, AOS, and grad year 15/30
Which factors seem to contribute to a job applicant's success in securing a permanent position, and how much? This is a good question for logistic regression 16/30
Most notable effects appear w/ gender, country of PhD, and prestige category onlinelibrary.wiley.com/doi/10.1111/meta.12591 17/30
Gender: Women are 10-17 percentage points more likely to get a permanent position than men. A modest-small advantage, but not what we would expect from the literature on underrepresentation in phil 18/30
What might explain this? We consider 2 potential mechanisms: Maybe bias against women drives out all but the most talented, who go on to succeed on the job market. Maybe DEI approaches are working. 19/30
Neither seems totally satisfactory, but both are compatible with empirical evidence of bias against women and a relatively recent change in job market success 20/30
Prestige: Grads of high-prestige programs are 7-14 percentage points more likely to get a permanent position than grads of low-prestige programs 21/30
The direction is what you'd think, but probably not the magnitude. Going to a high-prestige program gives an advantage, but only a modest-small advantage 22/30
Also worth noting that there's lots of variation in placement success within the prestige categories. Lots of high-prestige programs have placement rates < 50%, lots of low-prestige have placement rates > 67% onlinelibrary.wiley.com/doi/10.1111/meta.12591 23/30
Message for potential grads is: **pay more attention to a program's placement track record than its prestige** 24/30
Country: Compared to US, PhDs from other Anglophone countries have lower placement success. Canada is 9-22 percentage points lower than US. 25/30
It's possible this finding is due to data gaps, because APDA is Anglophone and US-centric. But the coverage of Canadian and UK PhD programs is pretty good. 26/30
This finding made me really sad. I have a bunch of friends who graduated from great Canadian PhD programs and left academia because they simply couldn't get a job. 27/30
Last year I mentored a student (US citizen) through the grad application process. They were considering a couple Canadian programs, and I had to discourage them because the placement rates were ~30% 28/30
Message for ACP/CPA: **You need to figure out what's going on here, and do more to support your PhDs** 29/30
Okay! You've got the major findings! Go read the details, fight about it on the blogs (I don't read any of them), don't @ me in your discussions just because you feel like it 30/30
