This year’s law prof twitter census results were published back in September. After some delay—due in part to my international move during a global pandemic, in other part to my having actual work to do, and in remainder to 2020 being such a great year—I’ve gotten around to mapping the law prof network and subsequently putting together some rankings and discussion.
The network is a little smaller this year than last. There were significantly more private/deleted accounts than in previous years, so perhaps more law prof twitter users are taking a break this year. That said, the total number of unique twitter users who follow at least one law prof has still increased by over a million to 4,074,832. The network density (0.088) has also increased steadily each year, as law profs increasingly follow one another on twitter. The image below shows the network, clicking it should take you to an interactive version. Colors correspond to detected communities where ties are somewhat more likely amongst same-colored nodes than to users outside the community.
Top-20 law profs by total followers:
|JonathanTurley||Jonathan Turley||George Washington||230591|
|StevePeers||Steve Peers||Essex (UK)||133656|
|jentaub||Jennifer Taub||Western New England||121805|
|orinkerr||Orin Kerr||UC Berkeley||96071|
|rickhasen||Rick Hasen||UC Irvine||82662|
|McCannSportsLaw||Michael McCann||New Hampshire||70137|
Top-20 law profs by law prof followers:
|orinkerr||Orin Kerr||UC Berkeley||715|
|AnthonyMKreis||Anthony Kreis||Georgia State||529|
|DBRodriguez5||Daniel B. Rodriguez||Northwestern||524|
|HoffProf||David A. Hoffman||Penn||483|
|leahlitman||Leah Litman||UC Irvine||455|
|daniellecitron||Danielle Citron||Boston University||453|
|OrlyLobel||Orly Lobel||San Diego||407|
Top-20 schools based on the number of law prof follower sums:
The previous few law prof twittersphere mappings generated some discussion around the gender dynamics of following relationships and community structure on twitter. Two years ago, profs with male names had significantly more law prof followers than profs with female names. Last year that gap had pretty much closed. This year, the gap goes the other way, with profs that have female-identified names having more law prof followers than their male counterparts. Male-named law profs still have significantly more non-prof followers.
In digging into this a little it more, I subsequently looked into gender dynamics at the ego network level by examining the female ratio within each law prof’s immediate network. This shows the gender mix of the other law prof accounts that the prof in question chooses to follow. I subsequently average this across the ego genders, as you can see below.
The takeaway here is that there is a clear gender homophily tendency within the law prof twittersphere. The overall female ratio of the subset of accounts for which I was able to infer gender (based on a fairly low tech gender name dictionary lookup approach) is about 0.45. So, one would expect that if profs followed accounts at random, they’d have that same female ratio in their own ego network. What we see instead is that female accounts have a female ratio of around 0.52, while male-named accounts are at about 0.39—demonstrating that women are more likely to follow other women and men are more likely to follow other men.
Caveats and data
As always, this process is imperfect. The census relies on self-reporting and is thus incomplete. If a user isn’t listed in the census they’re not in the network. There were also quite a few handles in the census that threw errors of various types when trying to get their data via the Twitter API. Errors included: no such user, user suspended, or private profile. I removed isolates from the network, so if you don’t have any follower/following relationships with other law profs you won’t show up in the network data. I did very little data cleaning from the original census data, so if schools are represented with multiple names or there were typos in the input data those issues remain. I was unable to infer gender for all accounts, so the above gender analysis excludes some nodes from the network. Raw data available for download here, network data available on request.