Sports: Study of N.B.A. Sees Racial Bias in Calling Fouls

[Guy Berger]

Hey Guy - do you know that paper that came out sometime in the last couple years that found that the results of academic studies are biased towards finding significance levels of just under 5%? They had histograms of the p values that showed a huge spike right at 4.9. It was pretty funny, and somewhat appropriate to the issue at hand.

I remember reading about that. (Though I thought it was 10%?) Definitely appropriate -- obviously this study could be debunked by further research. I assume the first step will be to see if another academic can replicate the results.

Guy

[Guy Berger]

Mr. Wolfers said that he and Mr. Price classified each N.B.A. player and referee as either black or not black by assessing photographs and speaking with an anonymous former referee, and then using that information to predict how an official would view the player

Ok... Jason Kidd. Black? or White? So we better ask the 'anonymous former referee'?

Seriously...

In the article it says that 12 players could fit into either racial group, and that their classification did not substantially change the conclusion.

Guy

[Guy Berger]

There's a post and thread on this article at the Marginal Revolution economics blog.

One of the commenters says:

The one statement in the NYT that I object to is that the paper was evaluated by three "independent" experts: Larry Katz was one of Justin Wolfers advisors and is a close friend to Justin. Ian Ayres is a frequent co-author of John Donohue, who in turn is a co-authors of Justin Wolfers. I'd be very surprised if Wolfers and Ayres didn't know each other pretty well. None of that changes the merits of the paper, but I would have picked differently if I had chosen the independent experts.

Another says (re: Noj's comments on all-stars and Laker scrubs):

The coefficients on the instrumental variables in table 4 are also interesting. For example, all-stars receive 2 fewer fouls per 48 minutes than do reserves. Also, fewer fouls are called on black players as a group (.763 fouls per 48 minutes), but 90% of this difference is explained by “observable characteristics” including position, starter-status, and physical stature.

Another good one:

The effect is big enough that an all-white team would, all other things equal, win two extra games over the course of an 82-game season

This is actually a nice lesson in explicating the practical size of statistically significant effects. When was the last time an all-white team played 82 NBA games in a row?

Guy

Edited May 3, 2007 by Guy

[alocispepraluger102]

fantasy football in the nba in a hardcore music forum?

why arent we discussing 'who taught art tatum'?

Edited May 3, 2007 by alocispepraluger102

[J Larsen]

fantasy football in the nba in a hardcore music forum?

why arent we discussing 'who taught art tatum'?

That was me. Took the better part of a nice weekend. Any other questions?

[Guy Berger]

N.B.A. Commissioner David Stern said in a telephone interview that the league saw a draft copy of the paper last year, and was moved to do its own study this March using its own database of foul calls, which specifies which official called which foul.

“We think our cut at the data is more powerful, more robust, and demonstrates that there is no bias,” Mr. Stern said.

Given a choice between an inexact measurement that shows "bias" and an exact measurement that shows no "bias", which do you think is a better designed study?

This is a non-story. A flawed measure found bias. An unflawed measure found no bias. In short, there is no rational basis for the claim of bias.

BTW -- ESPN quotes economist Larry Katz on the NBA "study":

In fact, after studying the NBA. data, Katz, one of the nation's most respected economists {Guy: and Wolfers's advisor}, told us: "It was so poorly presented that it was hard to figure out what they were doing. And to the extent you could figure out what they were doing, there was such incoherence you couldn't draw any conclusions from it."

also:

"As long as the assignments of crews to games is random, Justin doesn't have to know the official's [race] in order to know an all-white crew calls less fouls on white players than an all-black crew," said Derek Neal, an economics professor at the University of Chicago who was in the audience for Wolfers' presentation.

"It doesn't matter who officiates. You get the same number of fouls called on black players. But if the officials are all black, there are more fouls called on white players. And if the officials are all white, there are less fouls called on white players."

Edited May 6, 2007 by Guy

[J Larsen]

As soon as I read the NBA's bullshit reason for not releasing their data, I was skeptical of their study.

[Patrick]

Haven't looked at the study. Even if the results are statistically significant, they strike me as not too significant economically (or should I say in terms of wins/losses). Given the small effect, it also would not surprise me if someone found some other confounding factor that also explains the difference. Sure hope junior faculty member's tenure decision won't rely too heavily on this--or is everyone, faculty included, getting super-freakonomic now? Some sizzle, and some discussion outside econ lecture halls, but anything useful/insightful?

How do the authors translate the very slight increase in fouls called into 2 wins over the course of a season?