All Things Political (Science), vol. 2

Someone just forwarded me an article about an academic paper that's going to be published citing racial bias in NBA foul calls. To (over-)summarize, in any one game, an all-white officiating crew will call between .12 and .2 more fouls on black players than an all-black crew would, or 4.5% more fouls per year, which apparently results in a real loss of about 2 games per year for the teams with the most black starters.

The most interesting part of all this (as far as I'm concerned) is that someone actually conducted a fairly rigorous academic/statistical analysis of something as non-academic as NBA foul calls. Being that I'm fairly well stuck in statistics/"What can I publish?" mode, my mind immediately set off in search of how I might re-model such an experiment and arrive at different results.

The study controlled for things you might expect to affect foul calls like veteran/All-Star status, position, player "assertiveness", coaches' behavior, etc. If one were going to do a counter-study, they might look at things like flagrancy vs. non-flagrancy of fouls, since the former are much less a judgment call for officials; actual physical style of play (not measured by stats), since some teams/players may just be more physical than others; at what points in the game fouls are most likely to be called, and which players are most likely to have the ball at those times (or to be defending someone who has the ball); etc.

I'm tempted to drop a stat model right here, but I won't. At least not until I figure out how to turn it into a publishable paper, which seems to be 25% of what I care about these days.