On sample data on this year's TT hires

The past two years, I have presented data on hires from the Leiter job thread, complete with publication averages on new hires, etc. This year, because the process of collecting data is so labor-intensive and because I've had a very busy spring, I instead have taken a sizable sample of the tenure-track jobs reported (the first 100 jobs reported on the thread). Although I took pains to get all of the data right, I cannot promise there are no mistakes. Anyway, here is some of the data:

• TT jobs accepted by male candidates: 55 (55%)
• TT jobs accepted by female candidates: 45 (45%)
• TT jobs accepted by graduates of Leiter top-5 programs: 37 (37%)
• TT jobs accepted by graduates of Leiter 6-10 programs: 13 (13%)
• TT jobs accepted by graduates of Leiter 11-25 programs: 18 (18%)
• TT jobs accepted by graduates of Leiter 26-50 programs: 20 (20%)
• TT jobs accepted by Leiter-unranked programs: 12 (12%)
• Mean publications per TT job accepted: 3.46
• Median publications per TT job accepted: 1.5
• Avg. (mean) # of top-20  journal publications per TT job accepted: 0.77
• Avg. (mean) # of non-top-20 journal publications per TT job accepted: 2.40
• % of candidates accepting TT jobs w/no publications: 27%

Although I reported TT jobs and post-docs the past two years, the first significant thing I notice about this data is the mean # of publications (3.46) this year is 1.2 pubs higher than last year.

A second very notable thing is that although candidates from Leiter top-5 programs did significantly better than those lower in the Leiter rankings, the numbers of jobs accepted by candidates lower in the Leiter rankings were relatively evenly distributed (and relatively close to proportional to the # of programs in each grouping) -- so, aside from candidates from the top-5 programs, being higher in the Leiter rankings does not appear to be an incredible advantage.

Third, "top-20" publications were -- by my lights -- in surprisingly short supply. There were a total of 79 top-20 publications for the 100 TT jobs reported, and many of those were multiple top-20 publications (anywhere between 2-10) by single reportees. There were far, far more non-top-20 publications (271!) across the 100 jobs accepted.

Finally, there also appears to be a notable increase in the proportion of women accepting TT jobs (45% in this sample compared to 36% last year, though again this is just a 100 job sample).

Comments

[Michel X.]

Two small items:

1.) Your numbers for graduates by Leiter rank add up to 102%.

2.) When you detail the numbers for graduates by Leiter rank, are you using the overall international ranking, or just the US ranks? I ask because the first 100 posts in the Leiter thread have several entries by graduates of non-American PhD programs, and I was just wondering how they fit (or didn't!) into your picture.

[Tim]

These are interesting numbers to know. I wonder if anyone has tracked the number of TT hires of candidates who have previously held a TT and those who have not (hired straight from Ph.D. or from a post-doc, or a visiting job). My impression (perhaps mistaken) was that the proportion between the former and latter has been much higher this year than in past years. Of course, these stats might correlate with the publication stats to a significant degree.

[Gradjunct]

Deeply discouraging to see that 88-90% of TT hires were from Leiter-ranked schools. As someone outside those echelons it gives very little hope for success in finding TT work. I wonder, however, whether your sample might not be a bit skewed. After all, it stands to reason that the first 100 TT positions are going to be snatched up best the best candidates, those who optimally combine institutional prestige, research, and teaching excellence. As the job market season goes on, those people are taken off the market and things might open up for those of us at unranked schools with solid publications. I wonder if you'd get the same results w.r.t. institutional prestige if you had taken a random sampling of 100 TT hires without taking the first 100.

[Rachel]

I certainly had a hand in skewing that data. However, I received my TT hire last year, but deferred it for a year to take a postdoc, and I added a bunch of publications in the meantime, so the number of pubs when I was hired was 6-7, and that number now is 11-12 (depending on how one counts).

I suspect that most people who were in my situation will have been reported in the first 100 hires (since we knew about our jobs well before most people on the market *this year* will be able to announce theirs).

I don't think that ANY meaningful conclusions can be made from one year of data.

Also…I came from an unranked program.

[Marcus Arvan]

Hi Michel: Thanks for catching that. I accidentally transcribed one too many for Leiter top-5 and one too many for Leiter 6-10. It's fixed now! And yes, I used overall international rankings.

Gradjunct: You're very right. The rankings could be skewed in that way.

Rachel: Thanks for your comment. This is why I like to follow trends from year to year. I agree that no conclusions can be made from one year's data. But trends can be observed when looking at several years.

[Justin]

Is it surprising that top 20 publications were in short supply? Top 20 publications are hard to come by, first of all. Second, there are simply many more non top 20 venues in which to publish, and it's often a lot easier (which is not to say that it's easy) to get one's work into those places. So aren't those bits of data fairly predictable?

[Tom]

I'm sure this question has been discussed elsewhere on this blog, but I am very puzzled by the idea of "top 20 publications" or "top 20 journals". Is there an objective metric that determines which journals fall inside and outside that "top 20"? If so I'd like to learn what it is; if not then I do not see how helpful it is to keep throwing around the label "top 20 journals" (not just here but throughout the philosophical blogosphere).

I know that Leiter has done some poll on this issue, but the results were so heavily skewed towards analytic M/E interests that I found it hard to take them seriously at all.

At the very least, it seems to me that one would have to take into account the AOS of the relevant candidates when counting which candidates have "top 20" publications. It seems to me crazy to think that for someone doing History of Phil., having a publication in a distinguished, selective History Journal (such as JHP, BJHP, Archiv, Oxford Studies in Ancient Phil., etc.) should necessarily mean less than having a publication in journals such as PPR or Nous or Phil Review. Same holds for people doing ethics (surely a publication in Ethics or Phil./Public Affairs is the gold standard for people doing value theory, not to mention the fact that their pieces will actually be read by readers of those journals).

[Marcus Arvan]

Rachel: quick follow-up. Given that you accepted your job in last year's job season, shouldn't it have been reported in last year's Leiter thread? I always thought the Leiter thread was supposed to be an annual list of the present year's job season (and indeed, Leiter's instructions refer to hires made "this year"). I don't mean to nit-pick, but consistency of reporting is important for data-collection purposes.

Justin: No doubt top-20 journal publications are always in short supply. But sometimes it is implied -- both on job threads and by graduate school programs -- that one needs publications in top-ranked journals to get a good TT job. The data over the past several years suggest that this common belief is false. This year, almost 70% of the first 100 jobs reported were by people with no top-20 journal pubs. And, as one commenter noted above, the first 100 people reporting are arguably "the cream of the crop."

Tom: That's a good point. But I counted publications in Ethics/PhilPublic Affairs, JHP, BJHP, etc. as publications in "top-20 journals." After all, they are, as you note, clearly top specialty places.

[Tom]

@Marcus: I see, this is helpful. So, to clarify: when you use the label "top 20 journals", you actually mean a wide range of journals (more than 20) that are of roughly the same quality across different areas? that would make sense to me.

[Marcus Arvan]

Tom: Yes!

[Rachel]

Marcus: No, only the postdoc was reported in both Leiter and Caroline's data. I've been talking about double counting issues a few times. And like I said, a few of us got jobs *and* postdocs at the same time, only the postdoc was announced…and now the job is.

So, e.g., people think that it's more common for people to get a postdoc before a TT job, but with us we got both at the same time, and merely deferred the TT job for a year to take up the postdoc.

[jmugg]

Marcus:

When you count the number of publications, I assume you are not including conference proceedings or book reviews? I am under the impression that these don't really help you on the job market (though they can't hurt, right?). Also, how are you counting coauthored publications?

[Marcus Arvan]

Jmugg: I did my best to screen out conference proceedings and book reviews. However, I did count co-authored publications. I didn't know exactly what to do with those. Fortunately, co-authorship is still relatively rare in philosophy, and the vast majority of articles were not co-authored.

[Scott Clifton]

Marcus: It would be nice to see what teaching experience the hires have, too--especially since we often include the ability to teach well as a criterion for candidacy. As I've mentioned in earlier posts, I think that it counts very little, but here is where the data should confirm or disconfirm the claim.

[Marcus Arvan]

Hi Scott: I agree, it would be nice to know that. Unfortunately, I don't have the time to go hunting through everyone's CV for that much information! Just gathering data on publications takes a ton of time. Collecting teaching data would take a ridiculous amount of time...far more than I have.

[Justin]

Marcus,

I suspect it depends on what people mean by 'good' TT jobs. If we're talking about a job at a place with a top grad program, you probably do need top 20 publications or to come from a top grad program yourself or both. If you want a job at a SLAC, then while top 20 pubs might help, they're not required, as you're data indicate.

[Marcus Arvan]

Justin: I refute your argument thusly, "Any TT job is a good job." :) On a more serious note, I get your point. Top-20 pubs surely are important for getting an R1 job. No doubt. But getting *a* TT job? It would appear not.

[Rachel]

Justin: The median number, year over year, of "Top 20" publications at R1 universities is still around 1.

How many times can we say this: GETTING A TT JOB ISN'T A CV COMPETITION. It's not about how many publications one has. There isn't a magic number that will make it >50% likely one will get a TT job any given year. I know plenty of people who landed R1 jobs without a single publication.

And Marcus is right, any TT job is a good job. Whether it's a *good fit* for the candidate is a totally different question. There are some people who love 4-4 or 5-5; there are some who don't like 2-2 research-heavy jobs. Others love 2-2 jobs.

[Justin]

Rachel: It's not clear whether your remarks in the second and third paragraph of your latest response were directed at me. But in case they were, I'll note that I only said that a person probably needs a top 20 pub to get an R1 job. I didn't say it was strictly necessary. Nor did anything I say imply that getting a TT is a CV competition. I'm all too aware it's not. Nor did anything I say imply that a TT job with a 4-4 load is not a good job.