In an earlier post, I analyzed the association between a philosophy department’s ranking and the placement of its graduates in permanent academic positions. The motivation there was to examine the assumption that going to a “higher ranked” program will significantly improve one’s chances of getting a permanent academic job. That analysis showed that, indeed, a department’s ranking is positively correlated with its placement rate, but that association only explained 22% of the variation. Thus (as one would expect) there are likely a number of other factors at play that determine how well departments are able to place their graduates.
One other obvious factor to consider is a graduate’s area of specialization. A job candidate is only going to be hired if there are actually departments looking to hire candidates with that area of expertise. Some areas of philosophy may be more in demand than others, and this level of demand for areas may change from year to year. So for new philosophy graduate students interested in maximizing their future job opportunities, it may be valuable to know: Are there trends in hiring for different areas of specialization?
The graph below sheds light on this question, using data from all the job adds ever posted to PhilJobs.org, showing how the number of tenure-track job opportunities each year (blue bubbles) compares to fixed-term opportunities (yellow bubbles) or senior tenured opportunities (red bubbles), stratified by area of specialization. (Note that the bubbles are all semi-transparent and can overlap, which is why some of the bubbles may look sort of orange or purple.) You can mouse over bubbles to see the exact job counts. You can filter by whether the job type by clicking on the legend at the top.
What does this show?
Before offering an interpretation, it is first important to explain that the “area of specialization” (AOS) field in PhilJobs’ data does not use a structured vocabulary. This results in considerable heterogeneity in how AOS’s are described, and consequently, it is not straightforward to process this data and simply “count up” the jobs per AOS per year.
To get around this problem, I trained a “Bag of Words” model to identify a list of AOS terms, based on the most frequently occurring words in the AOS field of Philjobs’ dataset (filtering out conjunctions, articles, and other words that don’t relate to a specific field of specialization). Then to generate the data on jobs per AOS per year, I wrote a Python script that searches the text of each job’s AOS field for the term (or “string,” in Python parlance), and if it finds the string, it adds 1 to the total for the corresponding year. For example, the code uses the string “feminis” to capture jobs that describe an AOS including the words “feminist” and “feminism". Or the string “medic” will capture “medical ethics” and “philosophy of medicine” (which are not exactly the same area, but this stratum can still be usefully thought of as capturing opportunities at the intersection of philosophy and medicine broadly construed).
Further, since jobs can include more than one AOS, these bubbles really represent the number of opportunities to apply (i.e., jobs for which an AOS would be consistent with the ad), not the number of actual jobs themselves.
With those caveats in mind, what do we see? The most striking results here are probably the large number of fixed term jobs with an “open” AOS (around 105-110 each year), and the “ethic” category, which consistently sees 50+ opportunities both for tenure track and fixed term positions. “Politic” and “science” are also looking relatively strong, with 20-30 opportunities each year. Science also shows a noticeable uptick with 40 jobs in 2019, compared with 21 the year before. So for new graduate students in search of a project that is likely to increase their job market desirability, working on the ethical, social, or political implications of science would seem like a reasonably safe bet.
In terms of trends over time for tenure track, most of the AOS’s fluctuate around 2-10 jobs/year. There may be a slight downward trend for some fields, like “mind” and “modern,” and a recent uptick of interest in “gender” and “race”, for example. The pattern for philosophy of biology suggests that the area may have fallen on hard times, with 6 jobs in both 2015 and 2016, but then none in 2017 and only 1 2018 (although biology is back up to 3 in 2019). It is also interesting to note what may be emerging areas, like “information” and “data,” which had almost no offerings back in 2013-2014, but more recently there have been at least a few opportunities each year. In fact, “data” is a keyword in 6 jobs in 2019 (twice as many as “biology”).
However, the lack of big changes over time within most strata is also worth thinking about for new graduate students as they develop and shape their projects. For example, in the past 7 years, there have only ever been a handful of opportunities in math, medieval, and aesthetics—even for fixed term positions. So if your interest is in one of these more “niche” areas, it may be prudent to think about building a connection to a higher demand AOS in order to increase your future academic job opportunities.
Notes and Updates
December 9, 2019: Updated the figure and discussion for new PhilJobs data through December 2019.
March 17, 2019: Many thanks to all the people who reached out to me over e-mail with feedback and suggestions for ways to expand and improve on the analysis. In particular, I want to thank Eni Mustafaraj, a computer scientist at Wellesley College, who pointed me in the direction of using the Bag of Words approach to generate a more robust set of AOS terms.
I also added the data on the “Tenured, Continuing, or Permanent” job types. This is perhaps less useful data for graduate students or new PhD’s to consider, but it does still provide an illuminating picture of where departments are looking to hire at the more senior level.
March 11, 2019: After sharing this, a few people commented on social media about other areas that they were curious to see—specifically: “art,” “asian,” “continental,” and “non-western” (and I added “analytic” and “tech“ as well). So I’ve updated the figure to include these. (If you’d like to see other terms added to the list, please just reach out and let me know.)