a comment about agent based models in sociology in response to freese
About a week ago, Jeremy stopped by ye olde alma mater to give a talk on some new work. I was at SocInfo 2014, but my spies told me he made a quip about me. He mentioned that I thought that computer simulations were on the decline, even though his talk was about simulations. Of course, haters being haters,* the whole thing got blown out of proportion. Maybe, but it almost came to fisticuffs.**
Still, there remained a basic point – was I wrong? First, it helps to clarify. I never said that simulations were declining overall. In fact, simulations are a core technique in engineering, biology, physics, and computer science. Simulations also have a long history in *some* social science areas. Demographers, for example, have used them for population projections for decades. So, I fully admit (and have always admitted) that outside sociology, simulations are alive and well.
My point is specific to sociology. Simulations are honestly quite rare. Sure, a few folks do them. James Kitts, Kathleen Carley, and Peter Bearman are card carrying sociologists who have routinely used simulations. But how frequent is this? Not very, I’d hazard that less than 5% of papers in our main soc journals (top 4 + regionals). And yes, a few famous papers have been simulations (the 72 Cohen, March, and Olson comes to mind), but that generally doesn’t trigger a wave of simulations *IN SOCIOLOGY.* For example, how many authors in the Journal of Artificial Societies and Social Simulation have been tenure track faculty in sociology programs? Some, but not that many. How many readers of this blog have ever read JASSS?
I’d love to be wrong. I would love for their to be a large and growing contingent of social simulation in sociology programs.But right now, it’s niche area. Why? My guess is that there is a lot of inertia and there is a selection effect. I hope that changes.
* Hater = Fabio looking for twitter action on a Sunday night.
** How old people fight.