The NY Times piece on racial discrimination by NBA referees has generated quite a buzz from sports websites ranging from SI.com to the Dallas Morning News in addition to an extensive discussion on the sports economists listserv. As some have expressed on the sports econ emails, the paper is quite meticulous and interesting. Nonetheless, I’m skeptical of the result. Yes, discriminatory biases of various kinds exist, but a systematic bias among NBA refs, even a small one, is another thing.
Among the various criticisms of the study’s details, three have stuck out to me :
1) What about the race of the crime’s victim?
2) Are there key influences left out of the analysis that may offer an alternative explanation for the observed differences — in particular what about player “styles”?
3) Is an aggregated analysis of “fouls” really meaningful — for example, would anyone think an NFL study of “penalties” that didn’t distinguish between holding for OL and pass interference for DBs meaningful?
The study does address question #2 but not sufficiently. At a general level, charges of discrimination in sports often don’t fully take account of more fundamental differences that seem to be present between players of different races. For example, in the last 15 years, no white player has won the NBA Defensive Player of the Year Award (the last was Mark Eaton in 88-89). Of the 75 first team all-defensive players over this time frame, only 1 was white (5 for second teamers). Among the 150 “All-League” first and second teamers in the last 15 years, only 14 were white (see Basketball-Reference). That’s a much more obvious race-based difference than in the foul study, but one that no one is attributing to discrimination. Yet, the underlying reasons for the differences may be quite similar.
In the data in the study, their are obvious differences in the averge white and black player. Whites were 2 inches taller on average and played center at a 22% higher rate than blacks while shooting about the same number of 3-pointers, rebounding nearly the same, blocking only slightly more shots, and doling out only slightly fewer assists. The average difference might crudely defined by the “soft” tall white player who likes to shoot from the perimeter. Also, 27 percent of white players were foreign born as opposed to only 3 percent for blacks — a difference in “style” quite obvious in international competitions. These “stylistic” differences are only accounted for by height, weight, position, or other characteristics but don’t really get at the more complex interaction reflected by “style.”
Question 3 implies that breaking the analysis down by types of fouls or, at least, by position so that similar kinds of fouls are being studied would help in a comparison of apples to apples. A position by position breakdown as well as one separating foreign from domestic white players might also help to incorporate stylistic differences better.
Barring those kinds of adjustments to studies like this one, I would come back to my earlier observation that there are much more glaring racial differences in awards and All-League teams than in foul-calling if someone is looking to make “discrimination” charges.