New APPS

When I first looked at placement statistics at the Philosophy Smoker I performed some analyses that I shouldn't have. First, I performed too many analyses. Second, I used the wrong kinds of analyses for some of the data. I did not imagine that these statistics would take off as they did and I was overworked*, which contributed to some mistakes on my part. One of these mistakes was running correlation analyses over gender:

I also found a negative correlation between PhD granting institution and number of publications (-.17: the lower your PhD granting institution is ranked the more peer-reviewed publications you have) and between gender and number of publications (-.21: if you are a man you likely have more publications than if you are a woman). 

While at the time I suspected that this negative correlation had to do with the increased difficulty women have in publishing their work, others worried that women had an upper hand on the job market. I brushed off this latter worry because the proportion of women who found tenure-track jobs was about the same as the proportion of women who obtain PhDs in philosophy. In fact, in the 2011-2014 data set I found that there is not a significant difference between the proportion of women who graduate from each department and the proportion that find tenure-track jobs from each department (but there is a significant difference for postdoctoral/VAP/instructor positions, which are awarded to a smaller proportion of women relative to women graduates). But this worry regularly comes up in comments and I feel a responsibility for having possibly led people astray with analyses I shouldn't have used in the first place. For that reason, I want to provide some more appropriate analyses here, as clarification on the relationship between gender and publications in the placement data from 2011-2012 and 2012-2013. Those who want to check this work can use the spreadsheet at the bottom of the post here, which is the one I used. (I do not use the more recent data because I decided not to collect publication data in this last round, due to time constraints.)

Note first that I am here simply reporting on an old set of data to clarify the relationship between gender and publications within that data set. This data set could be improved, so I do not want to stand by these numbers as being representative of the discipline or placement as a whole. 

Of 498 placed candidates in 2011-2012 and 2012-2013, 149 were women (30%). 110 women (32%) and 229 men (68%) were placed in tenure-track jobs. For comparison, the mean percentage of women graduate students per department, as reported in the 2013 APA Graduate Guide, is 32.58%, and in 2009 one source put the percentage of women graduates from doctoral programs at just over 30%. 

What is the mean number of publications for women and men in this data set? For all of the jobs (tenure-track, postdoctral, and VAP) and for all peer-reviewed publications, placed women have an average of 1.13 publications, whereas placed men have an average of 2.17 publications. Thus it looks as though placed men have one more publication, on average, than placed women. Yet, if we look at median number of publications, this difference evaporates: the midpoint of publications by both women and  men is 1 publication. (The mode is 0 for each.) Why this difference between mean and median? The difference comes down to those at the extremes: 15% of men and 5% of women have 5+ publications. Interestingly, the extremes change for men, but not women, from the top 20 departments (using the 2011 "English-Speaking World" PGR, since I take this to be the disciplinary standard at the time of placement). Placed women from these departments have a mean of .89 peer-reviewed publications, compared to a mean of 1.86 for placed men, whereas both have a median of 1 publication, which is similar to the data above. Yet, whereas 5% of the placed women from these departments still have 5+ publications, only 10% of men from these departments have 5+ publications. If we look only at those placed candidates without (reported) prior positions, the percentage of men with 5+ publications drops dramatically to 4.67%, whereas the percentage of women with 5+ publications stays about the same at 4.08%. Thus, what appears to be pushing the mean away from the median are male candidates outside the top-20 who have prior positions. Since the proportion of women placed in postdoctoral, VAP, and instructor positions (24.5% for 2011-2013) is significantly smaller than the proportion of women graduate students (average of 32.6% for departments in the 2013 APA Graduate Guide), the difference in mean publications may be partly caused by the difference in opportunities prior to the tenure-track. At least, I think this is something that readers should consider when looking at these numbers. 

*In 2011-2012 I was a full-time graduate student in the Psychology Department at Boston University with a 4/4 course load, a half-time research assistant in Takeo Watanabe's Vision Lab, and the organizer of a popular monthly talk series (Neuphi), while also defending my dissertation, looking for an academic job, and getting married. I took on this project partly because it was a diversion from weightier pursuits, but mostly because I was on the job market, I was interested, and the numbers were not available.