Black and White in Black and White

May 2, 2007
Posted by Jay Livingston

The New York Times had a front page sports story today. Not ARod’s homers, not Daisuke’s K’s. It was a story based on an unpublished paper by two academic economists. Study of N.B.A. Sees Racial Bias in Calling Fouls,” said the headline.

“We find that black players receive around 0.12-0.20 more fouls per 48 minutes played (an increase of 2 ½-4 ½ percent) when the number of white referees officiating a game increases from zero to three.”

Bloggers everywhere are going to be all over this story, but here's my take.

Are the NBA refs racially biased? The Times couldn’t find anyone in the NBA who would say so. Doc Rivers and Mo Cheeks—both black, both coaches— declined to comment, and Rod Thorn, president of the Nets said he didn’t believe it. There may be a difference between what guys in the NBA can say publicly and what they really think. Still, “no comment” is hardly ringing endorsement of the economists’ thesis, and you’d think the Times might have been able to get at least one or two retired players to say that maybe the white refs might have made some questionable calls against black players.

Why haven’t any players made the racism call against the refs? Why did it take two professors? Mostly because the racism, if it exists, is invisible to the naked eye. First, any racism on the part of the refs has to be unconscious. I can’t imagine a real racist anywhere in the NBA, certainly not among the referees.

More important, the bias effects are so small, you have to collect a mountain of data in order to detect them. It’s like a coin that you have to flip 10,000 times to detect its slight bias. The economists used thirteen NBA seasons with 600,000 fouls. And what did they come up with? A difference of at most 0.20 fouls per player per game. Five players, one-fifth of a foul. Imagine an all-black team playing an all-white team; at the end of the game, the black team would have been called for one more foul than the whites. (In my mind’s eye, I picture the all-white team, their shooting guard hacked while attempting a two-hand set shot, then going to the line and shooting his free throws using the old underhand scoop technique.)

In the real NBA of course, there are no all-white teams; blacks account for 83% of all playing minutes. How often do you see a team with even three white guys on the floor, even when the coach has gone deep into the bench? So with 17% white players, it works out to less than one extra foul every five games. It may be “statistically significant,” but statistically significant is not always meaningful.

The Celtics finished the 2006-07 season 24-58; they lost more than 70% of their games. They are not the Celtics I remember, the Celtics of the 1980s with Bird and McHale and Parrish. The difference is painful. But it’s not about the refs calling one extra foul a week.

1 comment:

Brad Wright said...

I had similar thoughts when reading the article. A plausible hypothesis, but pretty tough to prove. Are there other variables that would make the relationship spurious? E.g., position played (do frontcourt players get more fouls, are black players more likely frontcourt), starters, fame (best known players get the calls, etc...