Marcus Arvan at the Philosophers' Cocoon posted sample data from the new appointments site at PhilJobs, which is discussed in a great post by Helen de Cruz here at New APPS. In comments at de Cruz's post and in a new post Arvan discusses the impact of Gourmet ranking on women and men seeking tenure-track jobs. I wanted to follow up on Arvan's post by looking at the full set of data currently available at PhilJobs. I did this in part because I knew that the sample Arvan collected was skewed on gender, due to an earlier analysis on gender I performed for a comment on a post at the Philosophy Smoker. With that convoluted introduction aside, here is a summary of the findings, in keeping with the findings by Arvan: the gourmet rank of one's PhD granting institution appears to have a greater impact on men seeking tenure-track jobs than on women seeking tenure-track jobs. Although I cannot yet speak to the source of this discrepancy, I (like Arvan) find the difference troubling. I welcome comments on the source of the difference below, although any comments will be subject to moderation. Let's look more closely at the data (Note: the linked spreadsheet was updated on May 14th):
By my count, there are 150 tenure-track or equivalent jobs posted to the PhilJobs site. 86 of the hirees (57%) are men, and 64 (43%) are women. If you look at the ranking of these candidates' PhD granting institutions from the English-Speaking World version of the 2011 Philosophical Gourmet Report, 38 of these candidates' departments (25%) are unranked. If you look at the mean score, instead, which allows you to use any of the rankings, 22 (15%) are unranked. So 15% of those who have gotten tenure-track or equivalent jobs so far this year come from PhD granting institutions that are not scored or ranked at all on the 2011 Philosophical Gourmet Report.
I split up my analysis into two parts, one looking at ranking (in the English-Speaking World version of the 2011 Philosophical Gourmet Report) and one looking at mean score (from the 2011 Philosophical Gourmet Report).
Starting with the ranking analysis, I split up the rankings in groups of 10, using 1-10, 11-20, 21-30, 31-40, 41-50, and unranked. Here is a table of the findings:
|Total Hirees per Rank||Men per Rank||Women per Rank||Percentage Total per Rank||Percentage Men per Rank||Percentage Women per Rank||Rank Categories|
And here is a chart showing the percentages of tenure-track hirees per rank (click for full size):
Since the slope for women is shallower than that for men (for the ranked departments), the ranking of one's Phd-granting institution appears to have a greater impact on men than on women.
One worry with this data is that there are so many unranked candidates. I decided to look at mean score to solve this problem, leading to a new analysis.
For the mean score analysis, I split up the scores into scale of 4-5, 3-3.9, 2-2.9, 1-1.9, and unscored. Here is a table of the findings:
|Total Hirees per Scale||Men per Rank||Women per Rank||Percentage Total per Rank||Percentage Men per Rank||Percentage Women per Rank||Scale Categories|
And here is a chart showing the percentages of tenure-track hirees per score category (click for full size):
In this analysis, the highly ranked programs seem to dominate a smaller percentage of the pool, but that is likely only because the rankings do not track the mean scores in a linear fashion. This is because only 9 institutions have a mean score of between 4 and 5, while 28 have a mean score between 3 and 3.9, and 46 have a mean score of between 2 and 2.9, etc. Thus, the rank steps for institutions with mean scores between 2 and 2.9 correlate with a much smaller difference in mean score than the rank steps for institutions with mean scores between 3 and 3.9, and the rank steps for these institutions correlate with a much smaller difference again than the rank steps for institutions with mean scores between 4 and 5. If we thought that mean scores represented some sort of objective value, this would make the rankings a misleading representation of quality, especially given that they cut off at 50, which for the English Speaking World Ranking is just between a mean score of 2.7 and 2.6, a difference that means very little in terms of the supposed objective value but very much in terms of perceived value (assuming that absence from the ranking makes a great deal of difference to perceived value).
In any case, the greatest number of tenure-track and equivalent hirees come from institutions with a mean score between 3 and 3.9 for both men and women, and for women more hirees come from institutions with a mean score of 2-2.9 than from institutions with a mean score of 4-5. Of course, given the very small number of the latter and the high number of the former, this is not reason to suppose that women graduating from these institutions have the same chance of success. The trend for men, as in the previous analysis, seems to represent a greater impact of score on one's chances of landing a tenure-track job.
If you would like to play around with the data yourself, here is the spreadsheet.
Update: I changed the charts to focus on percentages, rather than numbers, since I think this is a somewhat more useful comparison. You can view charts created with numbers at the linked spreadsheet.