The following is a guest post by Jeff Colgan, Richard Holbrooke Assistant Professor at Brown University, and is @JeffDColgan on Twitter.

It’s that time of year again, when professors are designing syllabi as fast as they can with deliberation and care. Recently I analyzed IR syllabi for PhD students. The data suggest a gender bias that instructors could easily correct.

The case that gender diversity is good for IR and political science has been made elsewhere, repeatedly and persuasively. According to APSA, women are 42 percent of graduate students in political science (in the US), but only 24 percent of full-time professors. If we assume that part of what it means to encourage female students to pursue academia in IR involves showing them examples of great research by women, early and often, then we ought to pay attention to our syllabi.

My findings come out of a broader analysis of graduate training: what it can teach us about the decline of IR theory, the value of Google Scholar (a subject of much recent debate), and much else.  That analysis is forthcoming at International Studies Quarterly.  It uses a new dataset of 3343 required readings assigned in the core IR course (“proseminar”) from 42 US universities.[1] Unfortunately, the data do not speak to diversity issues besides gender, like race or ethnicity.[2]

The data suggest that 82 percent of assigned readings in IR proseminars are written by all-male authors (i.e., women or co-ed teams account for the other 18 percent).  That percentage is high, but it is also roughly consistent with the gender pattern of articles published in top IR journals (81 percent male-authored).[3] At first glance, there does not seem to be any additional gender bias at the syllabus-design stage.

The gender balance does change over time.  Men wrote 90 percent of assigned readings that were published prior to 2000, compared to 73.5 percent since 2000 (i.e., 26.5 percent were published by women or co-ed teams).

I couldn’t help wondering, though, whether the instructor’s gender mattered.  It looked that way in the original data, but the sample of courses taught by female instructors was too small to be sure. I needed more data.

A fabulous undergrad, Miriam Hinthorn, helped me collect data from 73 additional syllabi: female instructors taught 35 of them, and male instructors taught the rest.  Collectively they contained 4148 required readings (each reading could be an article, a book, or a book segment). These courses were slightly different from the original dataset, because (a) they were not IR proseminars and (b) we occasionally used more than one syllabus per university.  They were still always IR courses designed for PhD students.

We found that female instructors tend to assign more readings by female authors than male instructors.  Men or all-male teams authored “only” 71.5 percent of readings in courses taught by female instructors.  By contrast, male authors wrote 79.1 percent of readings in courses taught by male instructors.  The difference is statistically significant (p=0.01). Put differently, female instructors assign 36 percent more readings by women (including co-ed teams) than male instructors do, or about 5 readings per course.

Of course this is correlation, not necessarily causation. The findings might be at least partly due to differences in course composition, or instructor age, rather than bias per se – though those other variables would raise additional gender questions (e.g., is it all ok if women are teaching fewer security courses than men?). Either way, the findings are hard to ignore.

Female instructors are also considerably more averse to assigning their own research as required readings.  They assigned an average of 1.68 readings that they themselves had written (as solo or co-author).  Male instructors assigned about twice as much of their own work: an average of 3.18 readings.  Again, the difference is statistically significant (p=0.01).

The gap between male- and female-taught courses would grow even larger if female instructors assigned their own work at the same rate as men. If they did (adding more of their own without subtracting anything else), research by male authors would then account for only 69.3 percent of their readings.  That rate would be 10 percentage points lower than the gender composition in male instructors’ classes.  That is, female instructors would be teaching 47 percent more readings by women than male instructors.

Is this type of analysis a case of political correctness gone mad? Some would argue that instructors should just assign the best readings. I agree with that. But “best” is partly subjective, and gender affects such judgments. My own experience is that revising my syllabus with gender in mind was not only feasible, it made it better. (Dan Nexon had a similar experience revising an article of his to address the citation gender gap.)

So if you’re teaching a PhD course in IR this fall – or any social science course, really – take a moment to consider how your syllabus will appear to your students, especially women and minorities.

[1] Yes, I will make the data public, anonymized to hide instructors’ identity. (No, not until it’s published.)

[2] Diversity in IR, beyond gender diversity, merits attention and study. The data are a lot harder to collect. I know, that’s a crappy excuse.

[3] Data from a recent article in International Organization by Maliniak, Powers, and Walter