Referee Bias Coverage

Last Wednesday, when I first followed up Dwil’s comments about the study that suggested a racial bias among NBA referees, I wrote:

There are a couple of things about this study that may pose obstacles for a clear-minded discussion of it, however. The first is that the sports commentariat is not well-versed in the language of statistical analysis, generally speaking. This lack of facility with that language is what is likely going to allow David Stern to get away with trumpeting the league’s own study (which shows no bias), notwithstanding the obvious fact that the league’s study, which includes no statistical controls for alternative hypotheses, is obviously a joke. The second, as I’ve written about before, is the commentariat’s ongoing difficulty in seeing racism in anything other than Black and White terms. In other words, it’s hard for people to acknowledge the subtler forms of racism (or prejudice more generally). Either somebody goes Imus or John Rocker or Tim Hardaway, in which case their prejudices are obvious, or to raise the issue of race is to “play the race card” and to label something prejudiced is to be divisive or shrill, or whatever.

Having spent the last few days sampling some of the coverage, I am here to tell you that I dramatically understated just how obtuse the sports commentariart could be. Not everyone got it wrong, but the commentariat’s tendencies toward anti-intellectualism and laziness were on full, painful display.

A few examples, with commentary. Here was Charles Barkley with Dan Patrick last Wednesday:

“That might be the most stupid study I’ve ever heard, for two reasons. Number one, there are a lot more black players in the NBA. So, of course, there are going to be more calls by white referees against Black players. But, also, I bet those jackasses, if they wanted to, I bet Black referees call more fouls against black players. For them to come out with a statement like that, is irresponsible and it’s asinine.

“to come out with a survey saying white officials call more fouls against black players is stupid and its not right.”

One has to love the fact that Charles Barkley has, all of a sudden, developed delicate sensibilities about what should and shouldn’t be said (“it’s irresponsible”). And, it’s not clear whether, when Barkley says “it’s not right,” he means that it’s factually inaccurate or that it violates a code of civility that should not be violated. (As an aside, it’s always fascinating when a crowd that normally decries “political correctness” suddenly and un self-consciously embraces it so fervently. But, as I’ve discussed many times before, talking about race, unless it comes in a flagrantly offensive and incendiary form, is considered out of bounds by much of the mainstream sports media).

I don’t know why I should be surprised, but the point Barkley raises here, about there being more fouls on Blacks because there are more Black players falls into the logical fallacy category known as: “moronic.” The study, of course, doesn’t say that the aggregate number of fouls called on Black players as a group is higher. It says that the average individual Black player gets called for more fouls depending on the racial composition of the officiating crew. This, of course, has nothing to do with whether there are more Black than White players in the league. Barkley, of course, was not alone in making this point.

And, in his usual, incisive way, Dan Patrick’s reaction to Barkley’s complete misunderstanding of the study: “Great stuff, Charles.”

The Around the Horn guys (fair question: why bother?) also, naturally, got it wrong. Leading the charge was Michael Smith, who proved that, among other things, he could really use a refresher course in remedial logic:

“what’s the motivation for this kind of a study. I mean, the NBA goes above and beyond to protect the integrity of its officiating. If there’s any thought of a conspiracy, David Stern’d fine somebody or suspend somebody. We just saw what happened with Joey Crawford and Tim Duncan. And, the second thing is, the NBA is 72% Black, so obivously there are going to be a lot more fouls called on Black players than white players, there are alot more black players than white players and the third thing is: what’s the solution, do you reduce the number of black players, do you increase the number of black referees simply to eliminate the fear of a conspiracy. No, that would be ridiculous.”

If you were wondering where the word “conspiracy” appears in the study itself, the answer is: it doesn’t. But, that’s not surprising because I think it’s fair to say that Michael Smith hasn’t actually read the study (it’s here, if you’re interested). And note, like Barkley, the embarrassing recitation of the league’s racial composition, despite the irrelevance of that fact to the study’s findings. In fact, Smith here perfectly illiustrates the Third Law of Punditry: the less you know about something, the more loud-mouthed and shrill your comments about that subject.

Like Barkley, Smith has also suddenly discovered political correctness:

“the fact that we’re discussing this study – that’s offensive, that’s an
insult to every referee in the NBA, to their professionalism, because now they’re being painted with this broad brush….give them the benefit of the doubt…they’ve risen up the ranks not by being racist or prejudiced.”

Bill Plaschke, of the LA Times, followed up Smith’s indignation about the revelation that referees are, after all, human, by playing the race card:

“this is an insult to black referees…you don’t think they’d notice this. They
police themselves.”

No, Bill, that’s not a reach on your part. That’s a responsible interjection of race into the conversation. That’ll be worth keeping in mind next time Plaschke accuses someone of “playing the race card.”

On Wednesday’s PTI, Kornheiser and Wilbon struck a more reasonable tone. Here’s part of their exchange:

Kornheiser: “my feeling is that it probably reflects some subconscious prejudices that we all have…now it seems to me that we could use this as a learning tool. You could call all the referees in and say to the Black refs ‘says this about you’ and you could say to the white refs ‘says this about you’ maybe we can do better.”

Wilbon: “Tony, the key thing you said is ‘prejudices we all have.’ I have found NBA circles, including NBA officials, to be more tolerant than anyone else I’ve covered.

Fair enough. Unfortunately, by last Thursday, Wilbon had apparently gotten the talking points from the league office and went on the war path against the study. Kornheiser introduced the Thursday segment by noting a key point of criticism about the study:

Kornheiser: “one of the major sticking points is that the study used box scores, not tapes of the games to determine which refs called which fouls.

Wilbon: “this is why I think it’s bogus – if you don’t tell me, for certain, this is
supposed to be a study, not just anecdotal evidence who made the call…did the white referee make the call on the Black player..because now your dealing with conjecture, not fact…or, I don’t want to hear this…”

And later, Wilbon added:

“Black players said, ‘we don’t see this – Kobe Bryant, Derek Fisher – we don’t understand this. Because we don’t have this problem.”

Note the attempt by Wilbon to use the language of statistics to impugn the study. Telling a social scientist he or she has only “anecdotal” evidence to support a claim is the equivalent of insulting someone’s mother in the school yard. But. clever though he is, Wilbon gets the terminology wrong. The study, however flawed it may be, is not based on anecdotal evidence, but massive statistical evidence. Wilbon may believe that the racial composition of the officiating crew is an imperfect proxy for knowing who made the actual calls, but since the NBA won’t release that data, social scientists will be stuck making approximations. And, as it turns out, the study’s findings all hold up even when you eliminate mixed-race crews. In other words, when we know with 100% certainty the race of the official making the call.

On his radio show last Thursday, Stephen A. was a bit more measured. After reminding his listeners of his bona fides as an NBA expert, Smith said:

“I have never, not one time, walked into a game, had this perception that white referees had something against black players…Now, that doesn’t eliminate the reality of biases…maybe [a referee] would be more comfortable if a white player talked to him a certain way as opposed to a Black player…the individual biases that people have inside of themselves – I’m quite sure they exist and the article alludes to that.”

But, Smith, too was bothered by the study:

“But, as a Black man, and one of the few that’s over the airwaves, what bothers me about a report like this…all I am saying is that as a Black man, when we bring issues of race out to the public, make sure it counts. Make sure we can say “look, this is clear. This is not one of those cases and that’s where I am uncomfortable with this.”

This comment perfectly illustrates the point I’ve made here and before: the difficulty of talking about race in any but the most simplistic terms. I actually give some credit here to Smith for at least wrestling with that problem. On a less impressive note, Stephen A.’s producer, Mike, did not cover himself in glory in his discussion of the study. Mike spent much of the show disparaging academics and intellectuals, demonstrating that he was a regular guy by saying he didn’t understand “all these charts and graphs.” But, Mike, who holds a law degree, also failed the elementary logic test: “in a league that’s 85% black (sic), how could more calls NOT be committed against Black players? And, to his discredit, Stephen A. chimed i: “exactly.”

Stephen A. also criticized the study for accusing the referees of “racism” and, at the same time, saying that the study contradicts itself because it only talks about unconscious bias which, in Smith’s view, cannot be racist because racism only exists where it’s self-conscious.

Of course, the word “racism” doesn’t appear in the study, but that gets back to the whole reading thing, so…

Some commentary, especially in print, did a better job of taking on the subtlety of the findings. David Steele, in the Baltimore Sun last Thursday wrote:

“But biases are in the nature of our society, not just today, but always and probably forever. You can’t, no matter how much you try, discount the idea that subconscious prejudices come into play. Pro-you, anti-them, or whatever has been drilled into your head in your years on this planet.

In the heat of the moment, it’s not inconceivable that an official, no matter how full of integrity he is, sees Shaq and Vlade Divac collide in the lane and forms the instinctive thought — faster than he can blow breath into the whistle – “That enormous black guy can’t get away with that.”

Or vice versa: An official sees Kirk Hinrich strip Dwyane Wade going full-steam toward the basket and, in the blink of an eye, concludes, “No way a white guy can pull that off cleanly.”

Steele himself demonstrates the difficulty of discussing sub-conscious bias when he notes that:

The study hints at racial profiling. In real life that’s borne out by real people
describing real experiences of being pulled over for DWB (driving while black). But the next condemnation of FWB (fouling while black) in the NBA will be the first.

Of course, if the study is concerned with the phenomenon of implicit association (referenced by Henry Abbott’s excellent True Hoop entry on the study), then we would not, in fact, no it when we see it.

Steele seems at once both unable and able to recognize that fact:

Joel Litvin, the NBA’s president of basketball operations, told that the paper’s conclusions are “flat-out wrong.”

Still, we are living where we are when we are. So, jumping to that conclusion is also flat-out wrong.”

And, Harvey Araton, in the New York Times, gets to the heart of the matter:

Let me say right from the start that I am not here to champion a 13-year study by a University of Pennsylvania professor and a Cornell University graduate student that found that white referees in the N.B.A. had called fouls at a greater rate against black players than against white players, and a corresponding but lesser bias by black refs against whites.

Not as a basketball issue, anyway.

Regarding the impact on the outcome of N.B.A. games and the integrity of the sport, my many years of being inside the arena tell me there are way too many variables beyond mere data solely extracted from box scores to draw meaningful conclusions from this. But as a study of the human condition, it is fascinating, and who among its critics qualifies as the authority on how the brain stores information and makes that split-second call? (my emphasis).

That last question by Araton is pivotal, because much of the reaction to the study (and I’ve spared you much idiocy, I promise, by- no surprise – Peter Vecsey, for example) is, in fact, born of an unthinking arrogance on the part of the sports commentariat. It’s the same arrogance that has produced such visceral disdain for the sabermetric revolution and other more statistically oriented approaches to sports. Namely, that anyone who suggests that our sports “experts” own eyes are not the final arbiter of all truth in sports is ill-informed, out of touch with reality and common sense and not fit to be part of the conversation with real sports guys. I wrote about this issue at some length last November, in a post on David Berri et al’s Wages of Wins.

One guy who really got it right last week was Boog Sciambi, talk sports radio host on Miami’s 790 The Ticket. Sciambi actually thought to talk to both Alan Schwarz, who wrote the article for the Times, and Justin Wolfers, the Wharton school professor who co-authored the study.

Some interesting points from Sciambi’s conversation with Schwarz:

1) Schwarz emphasizes that the study focuses on unrecognized, subtle racial bias: “which exists in otherwise perfectly fair-minded people…In general people will base decisions on racial details of which they are not even aware…in split second, high-pressure situations.”

<> 2) Sciambi also raised the issue of race as a sometimes non clear-cut factor, in the following interesting exchange:

<>Sciambi: “I think that white America specifically is uncomfortable with the phrase ‘race
is a factor.’ If it’s not ‘race is the factor’ or ‘it is racist’ they don’t want to hear it. For example, I think that race is a factor in how Barry Bonds is covered. The majority of the people covering him are old white guys and the fans that go to the games are majority white. I think it is a factor. People don’t like hearing that.”

<>Schwarz: “I think you have to be careful where you diagnose racism and where you do not because while it may be true at every turn, if it dilutes the conversation and makes
people calloused to it, it ultimately might backfire in terms of what you’re trying to
teach them…:

Sciambi: “I don’t think it makes people callous. I think it threatens them.”
As Stephen A.’s and Michael Smith’s comments, above, demonstrate, it appears that it’s not just White America that struggles with the distinction that Sciambi makes here.

3) Sciambi also specifically asked Schwarz whether the authors were calling NBA referees racist. Schwarz’ response:

“no, they are not in the business of labelling. What they found was a statistically significant meaningful numerical difference in the number of fouls that were called by each group on each group. What radio hosts and other folks want to interpret those as being, whatever noun and adjective they want to assign to them that’s your business.”

The following day, Sciambi interviewed Wolfers. Of note was that Wolfers blasted David Stern, calling him the biggest source of the misconceptions about the study (see my comment at the top). Wolfers told Sciambi that he asked the NBA whether there were any relevant variables theNBA thought he should include in his study that migh change the findings and that the NBA never responded and noted, about the NBA’s own study, which it refuses to release: “if I had facts that could help my organization, I would be the first to send it out to everyone I could…”

Wolfers also explained what the study did and did not conclude:

“we’re not calling anybody racist. In fact, I’ve talked to referees and they strike me as people who are unbelievably committed to getting the calls right.” And: “I think the NBA is a remarkably color blind organization. But, even in a place with such great training for the referees and such a commitment to excellence that we could find some evidence of bias, that makes you worry about the rest of society. We’re really not calling anyone a racist – we’re worrying about implicit biases…”

Sciambi repeated to Wolfers the comment he made to Schwarz, about the difficulty of talking about race in nuanced terms:

“people want it to be – no pun intended – black or white – they don’t want it to be gray. They want it to be “this guy is officiating with a clan hood on,” or they don’t want to hear it.

Wolfers response:

“our study is subtle.”

Too subtle, it appears for much of our sports punditocracy.


14 Responses to “Referee Bias Coverage”

  1. Another wonderful piece, dwil.

    Wilbon criticizes the study for its reliance on “anecdotal evidence” (which, by the way is false) but then cites anecdotal evidence to prove his point!! (“Black players said, ‘we don’t see this – Kobe Bryant, Derek Fisher – we don’t understand this. Because we don’t have this problem.”) My brain hurts.

    Also, not sure if you’ve mentioned this previously, but here’s an interesting test your readers might want to take. Offered by Harvard Univ., it’ called “Project Implicit” and gives some insight into implicit biases that each of us might carry.

  2. khweaz Says:

    Great article, I was thinking the same thing listening to the ESPN pundits discuss. Although, I thought Hollinger did a good job describing the study through the use of statistics. According to Hollinger, Lebron James, the leauge leader in MPG, will have committed roughly 11 “extra” fouls for the whole entire season. So it really is not the end of the world.
    However, the fact that everyone is refuting it so easily is absurd. I would be willing to bet that Wild Bill Plaschke has not been hunting around academia looking to analyze statistical studies. I guess the real blame is by having sports writers and personalities provide “expert” advice to the public. If the study is going to be refuted it should be done by legit scholars or at least someone who has read it.

  3. I have an issue with this whole “story,” that being the sensationalist way that the NYT – a liberal rag looking to cause a stir if there ever was one – summarizes the study. Schwarz writes in the second paragraph: “during the 13 seasons from 1991 through 2004, white referees called fouls at a greater rate against black players than against white players.” This is the “story” that every media outlet ran with.

    Now, if you actually look at the study, on pg. 36, Table 3, there is some interesting information — information that did not receive much attention from the drive-by media — showing the difference in foul rates called on white and black players by white and black referees. It shows there is NO difference in the foul rates for black players by black and white refs. What it does show is a difference in the foul rates for white players.

    So as far as I can tell, this can lead to one of two conclusions: either white refs favor white players by calling less fouls than black refs, or black refs disfavor white players by calling more fouls than white refs. The point is: there is NO WAY to determine which conclusion is correct.

  4. Let’s just say it: the pundits are philistines. They are. Not only do they have no understanding of the intricate details of a study like this, but they have no understanding of the reasons academics perform studies like this. However, they excel at yelling after the fact, with an arrogance that suggests they are smart and the people that devote their lives to these studies are fools.

  5. Jason, your conclusion is correct – it is either one or the other (white refs favor white players, or black refs disfavor white players). The point of JWeiler’s post is that, throughout the media coverage of this study – and not just mainstream media, but on too many blogs, too – that conclusion is not what they are reporting as the point of the study.

    Look at the quotes taken from Barkley, Wilbon, and Smith – all seem to misintepret the data of the study.

    That’s the “subtly” that the press is missing. It’s a shame, too, because this could be the rare opportunity to take an honest look at race. After all, we don’t know if the “bias” that exists in the refereeing that this study discovered is coming from the white refs or the black refs….

  6. SML – you’re right, the MSM isn’t reporting the “either one or the other” story. What they are reporting is the “white refs call more fouls on black players than they do on white players” angle . . . when at the same time, the study also shows that BLACK refs call more fouls on black players than they do on white players.

    This is the angle of the NYT piece:

    “A coming paper by a University of Pennsylvania professor and a Cornell University graduate student says that, during the 13 seasons from 1991 through 2004, white referees called fouls at a greater rate against black players than against white players.”

    This is also the angle that the MSM ran with, which obviously is meant to accuse white referees of being biased (read: racist), seeing as they call fouls at a higher rate against black players than they do against white players. But if it is also true — as I pointed out above — that black refs also call fouls at a higher rate against black players than they do against white players, then I have to ask: what is the point of the lede from the NYT article, other than to flame up another “racial” issue?

    You agree there are two possible conclusions: either the white referees are biased (not calling “enough” fouls against whites), or the black referees are biased (calling “too many” fouls against whites). Now please tell me why you think the “white refs are biased against blacks” angle got the lede in the NYT.

  7. jweiler Says:

    As the authors explain in the study, the reason that there is a higher foul rate among white players (per 48 minutes), is that a higher percentage of white players play big-man positions relative to other positions and big men are more prone to fouling than other players. In other words, Whites play a higher percentage of their minutes at foul-prone positions. When position played is controlled for (as it should be in a study about foul rates), white officiated crews do call fouls at a higher rate against Blacks than they do against whites. So, conspiratorial inferences about the Times are misplaced in this instance. Schwarz is a careful author and he got the data right.

  8. Jweiler – you missed the point.

    When position played is controlled for (as it should be in a study about foul rates), white officiated crews do call fouls at a higher rate against Blacks than they do against whites.

    You are as guilty as Schwarz of trumping one angle over another to suit your agenda. If you completed your sentence, it would also read: “and black officiated crews do call fouls at a higher rate against blacks than they do against whites. Additionally, black officiated crews call fouls against whites at a higher rate than white officiated crews.”

    Black players are not at issue here. The rate at which white officials and black officials call fouls against black players is exactly the same.

    White players are the issue. Again, either black refs are calling “too many” fouls against white players, or white refs are calling “not enough” fouls against white players. And the point is: there is NO WAY of knowing which it is. Therefore, even if you accept a bias, there is no way of knowing who the biased party is.

    Just as easily, the Times could’ve led with: black referees call more fouls on black players than they do on white players. Or, black referees call more fouls on white players than white referees call on white players. Now you tell me why they didn’t.

  9. jweiler Says:


    You’re simply wrong. As Schwarz noted in the article, there is an own-race bias for both Black and White referees. That was clear and reported in every single discussion of the study I have read or heard. But, the effect of white referees calling fouls on Black players is greater than the effect of Black referees calling fouls on white players. And, to repeat, adjusting for player-position, white referees and black referees do not call fouls on Black players at the same rate.

    So, the reason why the Times didn’t lead with your suggestion is that the effect is larger, and the overall impact on the game is, just as obviously, larger, since there are many more White referees than Black in the NBA, and many more Black players than White.

  10. I could be misinterpreting, but pg. 30 of the report reads:

    “There are also two ways in which these own-race biases may emerge: they may reflect referees favoring players of their own race, or alternatively disfavoring those of the opposite race[…] Table 3 is instructive, showing that the rate at which fouls are earned by black players is largely invariant to the racial composition of the refereeing crew. By contrast the rate at which fouls are earned by white players responds quite strongly to referee race […] suggesting that the impact of the biases we document is on white players, who are either favored by white referees, or disfavored by black referees.” emphasis mine.

  11. Jason-
    My two… the report states that the effect is large enough that it has a calculable outcome on games – 1.8%, yet when calculated as the ability to pay for quality players the outcome adds another percentage point. An then, there’s this from p. 24 of the study:

    “Indeed given that the winning margin has a standard deviation of about 12 points and is approximately normally distributed, it is not surprising that only a half-point shift in average winning margins would be sufficient to yield the substantial changes in the winning chances of one or another team.”

    I think that says a lot about opposite race bias and its palpable effect on NBA games when the majority of officials are white and the majority of the athletes are black.

  12. “subtlety” and “punditry” are oxymorons ***

    (*** = 1% significance test)

  13. Okay, we accept the findings of the study. But what, exactly, is their value? Here are the issues:

    1.a) The fact the NBA is primarily black does count for something. It puts a great limit on the scope of damage any prejudice may cause. Every team in the league has a bunch of black guys and a few white ones. Most of the time there’s a contentious foul, it’s a black player going against a black player! This prejudice, therefore, cannot affect team competitiveness – if you fill out a team with all white players, do you really believe that would win you a title? Shawn Bradley versus Tim Duncan? The significance of refereeing calls falls rather short of the significance of good shooting or good rebounding.

    1.b) The study finds a prejudice, it does not find conspiracy. Thus, we’re talking about a tendency as opposed to aforethought. If a refereeing bias is instinctive as opposed to pre-planned, it’s weaker. So we’ve identified a weak refereeing bias?

    1.c) Something my history teacher told me – there are no articles or books without bias, it’s just a question of how strong the bias is and in which direction it goes.

    1.c)i) In the NBA, there is a home advantage in refereeing calls. Call it the crowd factor, but in general teams do get some “home” calls, everyone is cognizant of this.

    1.c)ii) In the NBA, there is a superstar advantage in refereeing calls. As Rasheed Wallace once said, if you even breathe on Dwyane Wade, you get a foul called against you. The next time Vince Carter gets an offensive foul called against him will be the first time. Everyone is also aware of this.

    1.c)iii) It is my humble opinion that racial refereeing bias is not as significant as either of the above two. That is based purely on the fact that those involved in the NBA and those most fervent in watching it have never felt even a hint of it. Whereas we’ve all felt the stench of a superstar call.

    2.a) This is a statistical study and while this has its advantages, it also carries deficiencies in being purely technical. We know white refs make more calls on black players than black refs do. But that does not necessarily equate to white refs making more BAD calls. We identify more volume of calls, not greater volume of bad calls. You wanna show prejudice, you should study game tape and add up all the bad foul calls made. One way you can easily “hide” significant prejudice statistically is by giving all the calls one way in crunch-time situations when the game is close, but also calling fouls the other way in situations where the team you favour is in the lead. Thus you’d have statistically similar penalties, but drastically different in terms of importance in the flow of the game.

    2.b) This is my biggest issue – why is it explained as white referees having bias against black players? Given that we have no way to see into the mind of any refs, why is it not equally possible that black refs are favouring black players and not making calls where there’s a foul? This again goes to the heart of numbers failing to represent the full picture and to identify the driving element here. You can send all the white refs to racial integration and tolerance classes and still see the same bias in calls.

    So, really, why is this so damning? People are biased? That’s hardly news at all. If they’re strongly biased, it becomes apparent and can be elimited. If they’re subtly biased but not enough to be visible to an observant eye, what does that matter and how can it possibly be fixed?

    We need to remember that statistical significance is not always equivalent to real-world significance.

  14. Vlad,

    I may not agree with your points. But I’ve got to give you credit for your outlining format.

Leave a Reply

Please log in using one of these methods to post your comment: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s

%d bloggers like this: