05.03.07
Racial bias in NBA refereeing?
In the spirit of the season, the Freakonomics
Pretty interesting stuff - certainly a lot more interesting than actual NBA Games are…
You’ll find the article here: Freakonomics Blog » Racial bias in NBA refereeing?
Diesel The said,
May 3, 2007 at 10:39 am
This guy pretty much blows the entire study out of the water: http://msn.foxsports.com/nba/story/6767806?MSNHPHMA
It just seems like a whole lot of much ado about nothin’ to me. The NBA as a whole has long been the most integrated of sports (at least racially - see: Amaechi, John).
Also, GO CAVS!
Big D said,
May 3, 2007 at 2:00 pm
Diesel -
The only thing Kevin Hench “blows out of the water” is the fact that he’s never taken a statistics course and that he never read the article. He *is* a good writer, and his points are good ones - the trouble is that the paper specifically addresses all of them.
The bottom line isn’t about individual situations (e.g., which specific ref made the call, etc.) - the whole point is that an analysis of the huge dataset (over 250k player-game observations) reveals a statistically significant increase in the mean number of fouls called as the disparity between the race of the player and the race makeup of the referee team increases. Specifically, they find that players earn up to 4% fewer fouls and score up to 2.5% more points on nights in which their race matches that of the refereeing crew.
While those differences in means are statistically significant (in some cases at a 99% confidence interval), this does not mean that you’ll see the effects in any one game or case: it’s all about the results *in general*. Also, note that “4% fewer fouls” is 0.21 fouls - it doesn’t even make sense to apply it in a specific case.
The data shows a difference in the mean results when you look at a white player with white refs v a white player and black refs. The point of the study was to determine A) If the difference is significant (e.g., what is the probability that the results could turn out this way purely by chance) and B) Evaluate the data amongst a variety of other control variables to see if it can be explained by some other means than by race.
The article does a very thorough job of this and takes care to consider a wide variety of possibilities. It’s well worth a read - shame on Kevin Hench for not spending enough time to understand what he was looking at before spouting off at the mouth. …Oh wait, I forgot that he’s a pundit - congrats on a job well done, Kev.
Diesel The said,
May 3, 2007 at 3:17 pm
Other than the general AP article about the study, the Hench article was all I had read. So shame on me.
By the way, here’s a direct link to the study so you can actually read it before commenting (unlike myself): http://bpp.wharton.upenn.edu/jwolfers/Papers/NBARace.pdf
That said, to be honest, I’m not sure I’m even going to read it. I don’t think I care that much.
Big D said,
May 3, 2007 at 3:51 pm
Having just finished a stat course this semester its fresh in my mind.