Sunday, May 06, 2007

Michael Jordan and the NBA refs

Ian Ayres

The New York Times published an article a few days ago on the possibility of race-contingent foul calls in the NBA. The NY Times pieces was reporting on an empirical study of Joseph Price and Justin Wolfers. One of the coolest things about the article was the way that Alan Schwarz organized a set of econometricians to review some materials that NBA provided. It is important that empirical studies be subject to counter-crunching -- that they be susceptible to second opinions. So it is a great practice for journalists to do just what Schwarz did.

To be honest, though, it was a frustrating experience. Originally, I thought that I would have access to some of the NBA's data to see if Price and Wolfers results were robust to a more detailed set of data. But in the end, the NBA provided in drips and drabs just opaque descriptions of their results and methodology. They never really were willing to take the Price and Wolfers study head on by running a similar regression on the new data. I have great admiration for David Stern, but I think he is mistaken to conclude "“We think our cut at the data is more powerful, more robust, and demonstrates that there is no bias.”

If the NBA is serious about the possibility of race-contingent reffing, they should be more open to testing. It should be willing to give its referees the "Implicit Association Test". One of the coolest pictures in the Price and Wolfers study is Figure 2 a ref by ref rank of the estimated racial disparity in calling fouls:

Notice how the empty circles representing white referees disproportionately call fouls on blacks and the black dots representing black referees disproportionately call fouls on nonblack players. Imagine that the NBA refs took the IAT and we found the same pattern -- that refs with higher implicit association bias against blacks were most likely to disproportionately call fouls on blacks. This shouldn't be the end of the story, but it should cause additional concern. I applaud the NBA for refusing to circle the wagons in the Crawford/Duncan incident. But at least privately they should be willing to test whether bias is a problem. By the way, I'd offer to offer my services in this regard. I'd even be willing to pay $100 per ref to take a 5 minute IAT. One of the reasons I'm willing to make this offer (which i'm sure I could find funding for) is that it is important to find out whether the IAT predicts behavior in real world decisionmaking. Researchers have shown that it can predict behavior in the lab (for example, people with IAT aversion to blacks are less likely to make eye-contact in interviews). But an important next step is to test for validation in market contexts.

If there is a problem of unconscious racial bias in reffing, there is a separate very difficult question of what should be done about it. One possibility that I raised in my book "Pervasive Prejudice" is to make hiring decision turn in part on the results of the test. This might especially be appropriate for employees like refs and police who have to make important Blink decisions where like the IAT it's hard to keep unconscious bias from affect your behavior. It might be easier for us to consciously overcome unconscious bias when we have time to make more deliberate decision. But it might be harder to disable bias when you have just a second to make a decision.

Alternatively, some research suggests that unconscious bias might be ameliorated by reframing the context. And here is one of the great ironies. Simply showing people a picture of Michael Jordan (or Bill Cosby) reduces implcit racial bias in the IAT. This suggests that people so much want to "Be Like Mike" that he transcends racial categories, making us think less about race. Priming the experiment with his picture reduces the bias of the general public. What's unclear is whether that picture or the picture of other NBA stars would reduce the bias of refs themselves.

Finally, I was quite disturbed by the wording the New York Times "correction." I agree that the Times should have reported that Larry Katz was Justin Wolfer's advisor. But the correction said "After the article was published, The Times learned that one of the three experts, Larry Katz of Harvard University, was the chairman of Mr. Wolfers’s doctoral thesis committee." Larry (who was my roomate at MIT) has said that he told Alan Schwarz before the article was published that he supervized Wolfers' dissertation. I know Larry to be a person of the highest integrity and I believe that this was disclosed. In saying this, I don't want to impune the integrity of Schwarz. He was under incredible time pressure and pressure from the NBA that claimed that their study refuted Price and Wolfers. But this is a case where the correction itself needs correcting.


I'll pitch in for the "Referee's take the IAT" collection. We have been getting some follow-up calls of journalists making the links that you did in that story. I am interested to see how this develops.

Brian Nosek

When I look at Table 3, it seems that the bias is when the ref crew is all the same race. So, couldn't the bias only come out when one is in a "own kind" environment? And the presence of outsiders keeps everyone in check?

I don't remember the priest telling me when I went to Confession when I was a kid, "Well, Lance, it was wrong of you to disobey your mom and talk back to her like that, but since you set the table every night and do your homework and sent your aunt a birthday card, what the heck! You're a good kid. Your sins are forgiven automatically. No need for you to do any penance." 糖尿病 文秘 心脑血管 糖尿病 高血压 高血脂 冠心病 心律失常 心肌病 心肌炎 中风 低血糖 胰岛素 血糖仪 胰岛素泵 And maybe it's happened a few times and I haven't heard about it but I can't recall a judge ever letting somebody walk on the grounds the crook was a good guy and his friends really like him.

Post a Comment