Are NBA referees racist?

I wouldn’t have thought so, but Justin Wolfers, writing with Joseph Price, says maybe yes:

…during the 13 seasons from 1991 through 2004, white referees called fouls at a greater rate against black players than against white players…[the authors] found a corresponding bias in which black officials called fouls more frequently against white players, though that tendency was not as strong.

Here is the paper.  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.  A panel of three independent experts has judged that the Wolfers-Price analysis is more convincing than a David Stern-sanctioned rebuttal that no bias is present.

The NYT web site is slow this morning, try back later if the first link is giving you trouble.


Maybe black players really do foul more often that whites?

Did they account for differences in position? I suspect there are more white shooting guards than power forwards.

Ah, I misunderstood. Never mind.

Where's that coffee pot...?

I'm not sure if there are enough white players in the NBA to make this sort of comparison possible. There generally aren't more than two white players (out of ten) on the court at any given time during a game, and it's not at all unusual for there to be only one or even none at all.

Is a player with one black parent and one white parent classified as black or white? What about with one white grandparent?

You're obviously not from the United States, or otherwise you'd never ask this question. Here in America, a basketball player (or anyone else for that matter) would be legally considered black if he had merely one black great-great-great grandparent (all others being white) even if his physical appearance was completely white.

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.

As the authors note, some effects – fewer blocks and steals – may reflect less aggressive play. Assuming the players respond optimally to the racial composition of the referees, the data could also illuminate the trade-off between fouls and aggressive play: What is the marginal cost of a foul?

Individual-level referee data might shed additional light on peer effects as well: How do referees respond when working in mixed or uniform-race crews? I suspect some social monitoring occurs.

The paper seems to be making a fundamental ecological error of attributing community level behavior to individuals. The study suggests that having more white refs on a three-man referee team leads to more fouls on black players, but it has no means by which to imply that white referees call more fouls on black players.

While not as elegant an explanation, a black referee with two white referees in his group might be more likely to call a foul on a black player, and the white referees might make no distinction.

The proper conclusion of the study is that a community-level effect exists, and another study needs to be performed addressing which of the individuals makes the call to create strong evidence for any individual level bias hypothesis. As the NBA has that data and will not release it, and no grad student probably has the time to watch 1500 basketball games a year to generate the data, I doubt we're going to get an answer.

Here in America, a basketball player (or anyone else for that matter) would be legally considered black if he had merely one black great-great-great grandparent

Is the one-drop rule still on the American law books? I can't think of any currently active legislation that would depend on it. From Wikipedia:

In 1967 the U.S. Supreme Court, in its ruling on the case of Loving v. Virginia, conclusively invalidated Plecker's Virginia Racial Integrity Act, along with its key component, the one-drop rule, as unconstitutional. Despite this holding, the one-drop theory is still influential in U.S. society. Multiracial individuals with visible mixed European and African and/or Native American ancestry are often still considered non-white

Well, a black great-great-great grandmother is probably enough to check the minority box on a college application or the census, I'd be somewhat surprised if basketball referees are conducting blood tests on players so that they can call more fouls on some dude who looks completely white but really isn't.

Tyler S. -- The authors assigned players labels as "black" or "not black" by viewing photographs. They also enlisted a former referee to determine how referees would likely "see" the players. The analysis does not depend on anything like a "one-drop rule."

Regarding ice hockey, I vaguely recall a study that suggested that teams which wear brighter colors are called for more penalties than those which wear darker uniforms. The thought was that perhaps the brighter colored jerseys make infractions more obvious to referees. On the much smaller playing surface used in basketball, with so much less contact, could there be some sort of related effect that outweighs other effects of race? I doubt it, but it's worth considering before dismissing.

A much bigger example of ethnic bias in sports is how baseball managements were long biased against black and white American players and in favor of Latin American players because -- before the Bill James / Billy Beane / Moneyball statistical analysis revolution -- they overrated batting average and underrated on-base percentage, and thus statistically overrated Latins on average.

As I wrote in 2003, after the Toronto Star denounced the new management of the Blue Jays for trading high priced Latin players like Raul Mondesi for cheap white players:

The reason that scientific general managers like Ricciardi are modestly more likely to sign more white players than traditional general managers is because the old, less logical norms for evaluating ballplayers tend to slightly overrate Latin Americans.

For example, players with Spanish names (lumping both foreign and American-born Latinos together) ... are on average less likely to accept walks than whites or African-Americans. "It's not easy for a Latin player to take 100 walks," said Sammy Sosa early in his famous 1998 season.

In 2002, Hispanics had a combined batting average of .264, while everyone else together hit .260. On the other hand, the Hispanic "walk average" was 0.060, while the non-Hispanics' bases on balls ratio was 0.069, a significant 14 percent higher, leaving the non-Latinos with a better on-base percentage.

The patience gap has declined somewhat, from 16 percent in 1992 and 19 percent in 1982, probably because Latinos have largely closed the power gap. Twenty years ago, non-Hispanics hit home runs 42 percent more often than Hispanics, but that difference was only 4 percent last year.

The last 15 years have seen the emergence of Hispanics with excellent batting eyes like Delgado, Edgar Martinez and Rafael Palmeiro...
Still, this huge increase in slugging has not made the Hispanic shortfall in walks disappear.

Nobody is sure why this inequality exists, but it's been around for decades. American Negro Leaguers playing winter ball in the islands back in the 1930s were amazed at the kind of pitches at which their hosts would swing.

In the past, Latinos tended to cluster at the positions where fielding was more important than power, but that does not fully account for the patience gap.

In one of the few sabermetric studies ever done of the discipline disparity, David Marasco looked at American League hitters during 1994-1996. He found that American-born "glovemen" (shortstops, second basemen and catchers) were 24 percent more likely to walk than Latin American-born glovemen, while at the more offense-minded positions the gap was 7 percent.

Now, any thoughts on why this NBA study made it big in the NYT, but this quite obvious long-term systematic irrational ethnic discrimination in baseball has, as far as I know, never been mentioned in the NYT?

Hmmmhmmmhm ... I wonder what the reason could be?

Indeed, that Latin players benefited from the use of dumb statistics was, as far as I can tell, _never_ mentioned in the press until the 2003 "White Jays" crusade by the Toronto Sun against G.M of the Blue Jays, Bill James' acolyte J.P. Ricciardi, for dumping high priced Latin players with low on-base percentages like Raul Mondesi, in favor of cheap Moneyball players, who turned out to be overwhelmingly white.

The Toronto Star newspaper ran a series of articles on June 28, 2003 under the heading, "White Jays: In a city of so many multi-cultural faces, Toronto's baseball team is the whitest in the league. Why?" Nineteen of the 25 players on the Blue Jays' opening day roster were white Americans, three were African-Americans, and three were Latin Americans. (The average major league team has about three American blacks and seven Latinos.)

In late 2001, the Blue Jays hired as general manager J.P. Ricciardi, who is a follower of maverick statistics analyst Bill James. The Star documented that Ricciardi's quantitative acumen has had what civil rights litigators call a "disparate impact:" "Of the 39 players Ricciardi has since acquired through trades, free agency or waiver claims, 36 of them -- 92 percent -- are white," the newspaper reported.

The Star's sports reporter Geoff Baker claimed that this "raises the issue of whether the Jays truly need to be more representative of the city they play in at a time when they are satisfying fans by winning." ...

It's easy to criticize the Star's series, yet, although blinded by its political bias, it was groping toward the germ of an important idea. There's a revolution going on in how teams evaluate talent, and this growing sophistication is, on the whole, likely to benefit previously overlooked U.S. players at the expense of Latin Americans with flashier batting averages and stolen base totals.

Defending the Star's much reviled "White Jays" series, columnist Richard Griffin wrote the next day, "Jays General Manager J.P. Ricciardi along with Oakland's Billy Beane and other new-wavers believe in building offense through patience at the plate and taking no chances on the bases."

My UPI article "Baseball's Hidden Ethnic Bias" was published in response to "White Jays."

Steve--I'm still not entirely sure what point you're trying to prove in your posts, but I think it's something along the lines of liberal msm promotes some sports stories involving discrimination but ignores others?

You ask, "how much indignant media agitation would there have been over the last 20 years if Bill James, et al, had shown that baseball's old guard was persisting in using an obsolete statistic that discriminated _against_ a minority?"

I'm guessing not much for the simple reason that using power and and batting averages as metrics for choosing baseball players is prima facie a perfectly logical screening method, especially considering the majority of gm were former players, not statisticians.

Also, this is fairly obvoiusly not a case of "disparate impact." From first google hit: "Disparate Impact: Even where an employer is not motivated by discriminatory intent, Title VII prohibits an the employer from using a facially neutral employment practice that has an unjustified adverse impact on members of a protected class."

Guess what? On-base percentage is not a "facially neutral employment practice" in the sport of baseball. In fact it's pretty important. So regardless of what that moronic Star article posited, this is not a case of disparate impact. I think all you proved here is that the Toronto Star is not a very good paper, or at least lacks good editors.

"Business necessity: If the plaintiff establishes disparate impact, the employer must prove that the challenged practice is "job-related for the position in question and consistent with business necessity."

Whether you chose power metrics or the new moneyball statistical analysis, I don't you'll convince any court in the land that the relevant statistics are not "job-related for the position in question."

Speaking as an evolutionary psychologist - the underlying assumption behind this study may well be biologically incorrect.

The underlying assumption is that race has no relevance to performance in professional sports, and therefore that differences between races are due to prejudice. This assumption is incorrect - obviously - since there are big race differentials in terms of sports participation, presumably related to the fact that races on average look different and (therefore) have differently-structured bodies.

Since the effect size described here is so small, I would guess that there is still an element of uncontrolled confounding at work.

I believe that confounding cannot be fully-controlled by multivariate statistical analysis (mainly due to the imperfectly true assumptions of those techniques). What is needed are bottom-up controlled bioscience-type studies, which select subjects for inclusion using highly stringent criteria (ie. smaller studies which match subjects, analyzed using simple comparative statistics).

In other words, answering this question needs a more controlled 'biological' approach than this multivariate economics approach.

Re: Comments. All I can say is "Wow!" I'd like to bring this one back to B-ball if no one else objects. I often find sports being overanalyzed in the comment section of this blog. Point of fact, the NBA has always treated players differently in the way they are officiated. All one needs to do is re-watch the 1998 NBA Finals (Bulls v. Jazz) and MJ's game winning shot (I believe in Game 6) and his blatant push-off on Bryon Russell. Or how about the number of freethrows D Wayde shot in last year's finals. Superstars are often treated diffently. So really what we're saying is that officiating is more art than exact science (otherwise every single play would be reviwed, and wouldn't that be fun to watch). So why are we arguing about how an "exact" science is being applied to an inexact profession. It's just basketball people! It may be fun to debate this topic, but every single player and true fan pretty much have the same response...."What?! Who cares?!"

Chicagoan made a point: Jordan Rule and Ben Wallace Rule.Both afroamerican

I just saw a study which states that black people get more wet than white people when they go out in the rain.

It would be really nice the see the statistics year by year to see if there were any outlying years and see if there is some kind of trend.

Maybe the hard fouls should be tallied. That would certainly make your brown eyes blue. More likely blue eyes blackened.

Remember the Olympic Dream Team. Their racist bullying went unmentioned in the 1992 Olympics.

Comments for this post are closed