There's a thread over on Brian Leiter's page with quite an animated discussion over which sorts of publications do and do not look good to hiring committees. Setting aside the usual squabbling over which journals are "good" and which are "crap" (not my words) -- by the way, this is an issue that I will discuss at length soon in a post -- most of the discussion consists in different people sharing their (incredibly disparate) opinions about what does and does not look good.
I figured I would try to bring some empirical facts to bear on the discussion. Because I finished two papers yesterday, the computer that I do research on is broken (no data lost, thank goodness), and my wife is out at work, I've had so much time on my hands that I decided to go through the entire first page of hires on BL's tenure-track hiring thread for this year, and put together a database of everyone's publications (using BL's surveys on journal rankings).
The data I collected are interesting, to be sure. I'll begin with the data and then give some take-home messages. Here's the data:
Total # of Research-1 Jobs: 94
Hires with at least one top 20 publication: 30
Hires with more than one top 20 publication: 15
Hires with no top-20 publications:52
Hires with at least one non-top-20 publication: 47
Hires with more than one non-top-20 publication: 27
Hires with no publications (from Leiter top-25 department): 31
Hires with no publications (Leiter non-top-25-department):3
Hires with at least one publication in a top-5 specialty journal: 24
Hires with more than one publication in a top-5 specialty journal: 16
Total # of Teaching jobs: 31
Hires with at least one top 20 publication:6
Hires with more than one top 20 publication:1
Hires with no top-20 publications:19
Hires with at least one non-top-20 publication:11
Hires with more than one non-top-20 publication:8
Hires with no publications (from Leiter top-25 department):12
Hires with no publications (Leiter non-top-25-department):6
Hires with at least one publication in a top-5 specialty journal: 6
Hires with more than one publication in a top-5 specialty journal: 4
Takeaway lessons:
Lesson#1: If you want an R1 job, you must either have at least one top-20 journal publication or have no publications at all but come from a Leiter-awesome department.
Lesson#2: Non-top-20 journal publications DO NOT harm you, either in the case of R1 jobs or teaching jobs.
Lesson#3: If you want a job at a teaching university, you do not need any top-20 journal publications.
Lesson#4: If you want a job at a teaching university, non-top-20 journal publications appear to help you.
Lesson#5: If you come from a Leiter-ific department but have no publications, you are a very strong candidate both types of job (research and teaching).
Fwiw, all of these lessons cohere more or less with my own personal experience on the market, and with things I've heard second-hand. Sooo...who out there is surprised by the data? Reactions?
Wouldn't we also need to,know the number of people in each class to know the relevant percentage of each class hired?
Posted by: Kevin | 05/24/2012 at 11:28 PM
Very useful. There seems to be a missing data point: how many of the 52 no top 20 publications had publications in lower ranked journals?
Posted by: Neil Levy | 05/24/2012 at 11:28 PM
Neil: I didn't keep precise track of that, but from what I recall (I just did the whole thing tonight), most of the people who published in lower journals had also published in top journals. Very few people who got R1 jobs had *only* published in lower journals (though there were some).
The lesson for me, then, is this: people who get R1 jobs who have published tend to be very successful publishing in *both* domains (top journals as well as lower ones). Some people are just incredibly productive, and publish all over the place. They tend to get the R1 jobs...them and the Leiterific people who've never published at all.
Posted by: Marcus Arvan | 05/24/2012 at 11:41 PM
Sorry, I should probably clarify one thing. Although most of the R1 hires who published in top-20 journals *also* published in lower journals, the same was not true for the teaching hires.
Teaching hires mostly split into two categories: those who *only* published in top-20 journals and those who *only* published in lower journals (so, yes, several people got teaching jobs with only "lower" ranked publications).
Posted by: Marcus Arvan | 05/24/2012 at 11:52 PM
This is really interesting and conforms with my impressions of how the market works. One thing I've been curious about, and I know folks who worry about this sort of thing, is how publications on CVs figure when they are NOT in the dissertation area/core area of specialization. I definitely know folks who have top-20 and lower journal articles not in their AOS. Sometimes these are their only publications. I wonder your analysis gave you any initial impressions about the following questions: Do publications outside of one's AOS not count for much on the job market? Do they count differently depending on R1 or teaching jobs? Do they only count for something if the candidate already demonstrates a coherent research profile, so an extra, non-AOS publications are just more proof of professionalism? Perhaps these questions also apply to articles one co-authored with one's adviser or a senior faculty member and that were more in the latter folks' research areas.
Posted by: Kyle Whyte | 05/25/2012 at 03:02 PM