The figure above is from my paper with Adam Chilton, An Empirical Study of Political Bias in Legal Scholarship. I highlight it here because it received a great deal of attention in a recent workshop. It shows that Republican law professors (more precisely, law professors who make net donations to Republican candidates) tend to write a mix of conservative, liberal, and “neutral” papers, while most Democratic law professors write uniformly liberal papers or nearly so. (The numbers on the x-axis refer to the number of conservative papers minus number of liberal papers written by a professor out of a total of five. For example, -5 means that a professor writes five liberal articles; +2 means that a professor writes on net two conservative articles, which could mean three conservative, one liberal, and one neutral article, or two conservative and three neutral articles, and so on.)
What is the explanation for this pattern? I can think of five (which are not all mutually exclusive). (1) Democrats honestly write liberal papers that accurately reflect the world as it is, while Republicans do so only occasionally. (2) Republicans are open-minded and write papers contrary to their political leanings if truth leads them in that direction, while Democrats are ideologues. (3) Republicans who end up in academia are just not politically passionate, while Democrats are. (4) Republicans benefit intellectually from being in an environment where most people challenge their views, while Democrats suffer from herd behavior. (5) Republicans behave strategically, deliberately writing some liberal papers (or entering fields with weak ideological valence) in order to avoid being seen as excessively conservative by colleagues, deans, and students who mostly disagree with them.
In this recent workshop, a number of conservatives in the room argued that explanation (5) was the correct one. Apparently (and this is news to me), law professors sometimes (often?) advise politically conservative applicants for teaching positions to look for topics where conservative principles or methods would lead them to reach liberal conclusions. In contrast, the liberals at the workshop reported that they had not had not been told to mask their ideological leanings in whole or in part. This is anecdotal evidence from just a few people, but it does explain the pattern we observe in our data.