On Wednesday the New York TImes covered OB contributor Justin Wolfers’s paper showing: White referees called fouls at a greater rate against black players than against white players. … [There is] a corresponding bias in which black officials called fouls more frequently against white players, though that tendency was not as strong.
Without seeing the NBA's data sets (which includes data on specific refs and specific players--which Wolfers's study explicitly notes it's lacking) this whole issue just screams out for a Simpson's Paradox or, at the least, Absent or Hidden Data construal. Without knowing which ref called fouls on which player, there are other hypotheses consistent with the data. Add to that that part of the authors' methodology relied on the memories (sic!) of former refs as to who was working what game and this thing starts to look a lot weaker than it's playing.
Without seeing the NBA's data sets (which includes data on specific refs and specific players--which Wolfers's study explicitly notes it's lacking) this whole issue just screams out for a Simpson's Paradox or, at the least, Absent or Hidden Data construal. Without knowing which ref called fouls on which player, there are other hypotheses consistent with the data. Add to that that part of the authors' methodology relied on the memories (sic!) of former refs as to who was working what game and this thing starts to look a lot weaker than it's playing.
The study looked good to me; see here:http://www.stat.columbia.ed...
The study is criticized here:
http://mahalanobis.twoday.n...