Race and the Whistle

Alan Schwartz reports for the New York Times:

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.

Interesting. Tyler Cowen observes that "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." Of course, all other things are unlikely to be equal if you try and field an all-white team. Which, of course, leads to the ever popular question of who makes the All-White Team. My ballot has Okur at center, David Lee and Dirk Nowitzki at the forwards, and Nash and Ginobili in the backcourt.

Comments

Try doing it with only white American-born players. Then you're really up the creek.

Posted by owenz | May 2, 2007 8:58 AM

American born? Off the top of my head, would have to be something like:

C: Brad Miller
PF: David Lee
SF: Mike Miller
SG: Mike Dunleavy (?)
PG: Kirk Hinrich

Backups: Jason Williams, Luke Ridnour, Troy Murphy, Matt Harpring, Chris Kaman, Nick Collison, and Luke Walton

If any of these guys turn out to be Canadian, I don't know what we'll do.

Posted by right | May 2, 2007 9:21 AM

Surely you KIDD?!?

Posted by Otto | May 2, 2007 9:23 AM

The first thing that jumps out at me is the time period of the study (1991 - 2004). So, the cohort of officials being studied includes many who have long since retired. Second, the study was done on the basis of box scores, which do not identify which ref called which foul (and also do not identify the party who was fouled). So, even if a real problem has been identified, it is not clear (1) whether the problem has been eliminated through normal attrition or (2) whether the problem is the work of one or two biased (consciously or sub-consciously) refs, or a more systemic problem (nee Malcolm Gladwell).

Posted by Ephus | May 2, 2007 9:25 AM

Lee over AK-47? Really? What about Peja? I think you need to find room for Gasol in there somewhere to.
American squad: Hinrich, Jason Williams, Chris Kaman, Lee and Harpring - Miller and Murphy in the rotation.

Posted by berger | May 2, 2007 9:25 AM

I glazed over a little at the analysis, but I wondered, is there any foolproof way to say that the black players didn't actually commit more fouls? Maybe the parts where the black refs called somewhat more fouls on white players could count in that respect, I don't know. It's a usual problem with crunching real life data vs. designing an experiment to filter out extraneous factors. Other "bias experiments" such as sending out identical resumes with a "black" and "white" name, or having orchestra players audition behind a screen to prevent the listeners from knowing if the player was male or female have provided disturbing, but more foolproof results, I think.

Posted by JMS | May 2, 2007 9:46 AM

While Truehoop itself is as good as ever, the level of the comments has definitely gone down once the overall espn.com audience got in on the act. The thread on this subject is hilarious in a frustrating way; you can clearly tell that everyone attacking the study there either failed to read the NYTimes article, or failed to understand it.

That said, I dunno how good this interpretation of the data is, and I'm not going to replicate the study by myself. David Stern is absolutely right that the league has better access to better data than any outside party does, but the opacity of the L's refereeing organization - exacerbated by the policy of rapidly fining any NBA figure who dares to voice any sort of meaningful critique of the referees - means that I'm just not inclined to trust anything they say anyway.

Posted by Quarterican | May 2, 2007 9:47 AM

What I don't understand from the article is WHY this study is supposedly better than the David Stern study showing that there is no bias. We are told that some economists said that's so, but how am I to evaluate that claim, other then through the appeal to authority? It seems to me that the league's study should be better, given that they actually have the data on which refs made which calls.

It would be most interesting, actually, if BOTH studies were correct. Which, off hand, I think is possible.

Posted by Al | May 2, 2007 9:54 AM

Oh, and Otto, Kidd is black. Really. And so is Bibby. And Richard Jefferson.

(Actually the Times article states that some players could be classified either way, but that it makes no difference to the study.)

Posted by Al | May 2, 2007 9:56 AM

I really wouldn't be surprised if it were the case that whites caused fewer fouls in general. The stereotypical white NBA player is a slow-moving perimeter player and a soft defender. You're not going to pick up many fouls if you're a spot-up three-point shooter and you consistently hover 6 feet off your man on defense.

Posted by Alex | May 2, 2007 9:57 AM

Sometimes I feel something close to affection for Al. Anyway, this is a tempest in a teapot, at worst. The players don't think that the refs are, on the whole, unfair to black players. Neither do the fans or the refs.

I'd love to know how they factored out Bonzi/Artest/S. Jackson knuckleheadedness, as well. That's got to be good for 5%, right there.

Posted by SomeCallMeTim | May 2, 2007 10:13 AM

Kidd is clearly black. Fifteen years of watching John Stockton get away with murder on the court validates this study.

Posted by JP | May 2, 2007 10:14 AM

The flipside to the "sucky white defenders foul less" theorem, though, is that less capable defenders are going to need to resort to the intentional or semi-intentional foul more often than a defender who's better able to contain, direct, or foil his opponent's offense; where Raja Bell thinks/intuits "he's driving, but I can force him left and get him to pick up his dribble," Wally Szerbiack might think/intuit "oh shit he's driving I have to stop him there isn't a shotblocker behind me oh shit" and wind up just reaching in and fouling because there isn't anything else he (believes) he can do.

It'd be interesting, in this regard, to try and figure out the fouling rates (normalized for position as much as possible) of players considered to be good vs. bad defenders; my intuition, which is often wrong, is that if you throw out extreme badness at each end (i.e., the defender who just always fouls and the defender who never does anything to impede his opponent) then Ben Wallace and Pau Gasol are fouling at roughly the same rate. Someone may have already figured this out, of course, I just don't know about it.

Posted by Quarterican | May 2, 2007 10:19 AM

The players don't think that the refs are, on the whole, unfair to black players. Neither do the fans or the refs.

Until now!

Dontcha think NOW there's more of a chance someone's gonna say, after some egregious call, that ref only made that call because the ref is white and I'm black?

Posted by Al | May 2, 2007 10:30 AM

somecallmetim, you probably don't know about the al-howard compact, in which we agree on virtually all matters sporting and disagree on virtually all matters political, but yes, it's easy to feel affection for Al on sports.

this "study" doesn't deserve the label: if you want a real study, you would have to try something like taking a game tape and playing it for a bunch of refs and have them make "calls" as it went, then compare outcomes.

and yes, my very second thought (the first being "this looks like a bogus study") was Al's 10:30 thought: we're going to start hearing about biased refs, in both directions.

i mean, they are biased, but their bias runs to the standard nba template: superstars get the calls, home teams get the calls, teams making a run get the calls....

Posted by howard | May 2, 2007 10:36 AM

BTW, I agree with Quarterican completely that slow, white defenders are probably MORE likely to commit fouls, not less. I mean, isn't a foul by definition bad defense?

I would also dispute Alex's notion that "stereotypical white NBA player is a slow-moving perimeter player". Alex is ignoring the whole big, slow white center category - Eric Montross, or Chris Dudley (both of whom fit into the time period for the study). Nevertheless according to the Time article, the profs tried to account for these types of differences.

Posted by Al | May 2, 2007 10:36 AM

Following up on Alex's comments, kind of, I think this is pretty suspect if the study didn't control for *position*. I can't tell if it did (the article says the study "accounted for" a variety of factors including the stuff about centers being disproportionately white, etc.), but if not I'd think that would be a problem.

I'm not saying I doubt the findings -- I wouldn't be at all surprised if the conclusions were true -- but I do think to get a solid answer you'd have to compare fouls called on white guards versus black guards, white centers versus black centers, etc, rather than all together. Yes? No?

Posted by AJ in DC | May 2, 2007 10:46 AM

I'm sorry, I just have to say this: but I simply cannot believe that David Lee is good enough to be included on this hypothetical "All-White" team. Unfortunately, I simply cannot summon an alternative name off the top of my head.

This irks me, because Lee was a year behind me at a rival high school, and even back then I was insistent that he would never amount to anything in major college basketball, much less the NBA. He struck me as a guy who looked more impressive than he was when he was playing against vastly inferior athletes.

This irks me, because apparently I was wrong. I must defer to the superior NBA savvy of Matt and the commentariat in general.

Posted by Epsilon | May 2, 2007 10:47 AM

Dontcha think NOW there's more of a chance someone's gonna say, after some egregious call, that ref only made that call because the ref is white and I'm black?

Isn't the most unfair ref of all time--Hue Hollins--black? I'm not seeing it. The set of referees is small enough that players can come up with individual motivations for each of them.

, I agree with Quarterican completely that slow, white defenders are probably MORE likely to commit fouls, not less. I mean, isn't a foul by definition bad defense?

Here we might get near or to the great unsayable. If we're willing to believe, for the moment, that black players believe themselves to be more athletic (see Parker, Smush) than white players, might they not play more aggressively, and therefore get more fouls? As for the black ref/white ref change--if refs are former players, might they not be most aware of the fouls that their player-selves felt should/shouldn't have been called? You might have to believe that there's such a thing as "black game" and "white game" for that, though.

I'd love to see a split out of black/white/European.

Posted by SomeCallMeTim | May 2, 2007 10:50 AM

Eric Montross and Chris Dudley may well commit more fouls than the average (black) NBA player, but that does not explain why WHITE officials would call less fouls on them than BLACK officials.

I would be more interested to know why they picked 1991-2004. Would the results be the same if we limit the study to 1991-1998? Would there be no results if we studied only 2000-2004?

As I recall, some of the oldest (and worst, possibly most racist officials) have been weeded out over the last 10 years or so. Not that Violet Palmer is an improvement, but maybe the inherent bias is less for officials born after ~1940?

Posted by dan k | May 2, 2007 10:51 AM

"Up to 4.5%"? I'd have to see a distribution before I call "statistical significance". The 4.5% number sounds like an outlier.

Posted by nal | May 2, 2007 10:55 AM

Link to the study (in pdf)
http://graphics.nytimes.com/packages/pdf/sports/20070501-wolfers-NBA-race-study.pdf

Posted by Ephus | May 2, 2007 11:00 AM

Wow. There are days I'm truly ashamed to be white. Seriously, some of the gymnastics used here to excuse this are laughable.

ephus: Yeah, I'm sure that ALL the racist referee's retired and the problem is completely solved in three years. Brilliant fucking argument man.

JMS: Sure, those damned black people love to commit crimes, so of course they'll commit more fouls (this is a bit more unfair, but think about the stereo-types involved before you say something, will you?)

Howard: What in the hell kind of methodology is that, and what would it prove? Your force a problem by letting the referee's know they're being evaluated, and the criteria when they should know neither. You're also insisting that we study a contrived circumstance that is not analogous to the situation being studied. What you're suggesting isn't a "study" at all, and it would say more about the effects of scrutiny on calls made by referee's than it would on any actual basis's of the referee's themselves.

SomecallmeTim: the "black people do it doo" defense? Really? Is that what you want to go with?

Maybe I'm being unfair. But these are same cop-outs used WHENEVER a study finds racism is the cause of anything. A bunch of white people get up and spout nonsense that has nothing to do with the data, apply all sorts of personal anecdotes and poof: all the racism is gone.

Posted by Soullite | May 2, 2007 11:07 AM

Figure 2 seems to be relavent. A 95% confidence interval? When you're dealing with percentages around 1% to 2%, a 95% CI is too small. There is too much variance in data to warrent their claims of racial bias.

Posted by nal | May 2, 2007 11:14 AM

soullite, yes, you are being unfair. what i and other people are noting has nothing to do with saying that racism is gone; what it has to do with is whether you can judge racist effects from rates of foul-calling.

i can appreciate that my alternate study has problems of its own; in fact, i'm questioning how you can conduct such a study at all. every game is its own unique set of circumstances: what we would need to see is how different refs would call the same game. lacking that, we don't have anything useful at all.

to make myself clearer, i think there's a lot to criticize nba reffing for, but i don't see any basis to assume that racism in making calls belongs in the list. if there was a study that would show that the same play is consistently called differently on a racial basis, i would change my mind, but i don't begin to see how you find that out without, you know, having a bunch of refs look at the same play. i have no idea, though, how that becomes a "contrived" circumstance: the nfl consistently reviews refs calls and every now and then actually acknowledges the refs blew it, so it's certainly possible to go to the tape.

al, as a side note, sometimes a foul is "bad" defense, sometimes it's just not getting your feet set in time, and sometimes it's a good team defensive response to an indvidual's bad defense (i.e., if i foul someone who beat his man in order to force him to make 2 from the line, that was a "good" move as far as my team is concerned).

Posted by howard | May 2, 2007 11:18 AM

Soullite,

I don't know whether these are cop-outs. I think
the discomfort is with using stats as a diagnostic
tool, all by itself.

Its kinda like how docs view self-referred medical
scans: in the absence of symptoms etc., they
consider them a dangerous scam. So if the players
haven't previously considered it a problem...

You don't have to be white to be wary of
statistics. You know, the whole "lies, damned
lies..."

Posted by Raj | May 2, 2007 11:35 AM

Maybe I'm being unfair.

I don't think there's any reason to qualify it. Everyone is trying to figure out why a problem like this went unnoticed and unremarked upon by everyone who loves the game: players, management, and fans. One possible explanation is that the specific explanation offered by the economists is wrong, because it didn't account for other factors. As regards the Hollins comment, it wasn't "black people do it, too," it was "players (specifically '90s Bulls) don't seem to have noticed race as an issue in calls, and have explanations for each individual referee's bias; why would the study change those beliefs by the players?"

Posted by SomeCallMeTim | May 2, 2007 11:45 AM

Oh - well, if the *NBA* says there is no discrimination, then I guess that settles it. Who cares what a disinterested 3rd party academic study says?

Posted by sherifffruitfly | May 2, 2007 11:48 AM

"my intuition, which is often wrong, is that if you throw out extreme badness at each end (i.e., the defender who just always fouls and the defender who never does anything to impede his opponent) then Ben Wallace and Pau Gasol are fouling at roughly the same rate."

My intuition is the same, because in many cases getting a foul (or at least risking a foul) is a choice made by the player. In most cases a player who is sufficiently out of position will back off and not pick up a cheap foul.

Since fouls are the consequence of freely chosen risks, and from the player's perspective the main consequence of picking up fouls is being forced to sit, most players will make about the same number of risky plays per game, in order to get the occassional block/steal while still picking up the same number of fouls.

Posted by Doug T | May 2, 2007 12:00 PM

After reading the study, I actually find their results more convincing, but wouldn't be so quick to slam people throwing out alternate hypotheses!

But the bottom line is that the NBA has placed a much greater emphasis on getting the call right and video analysis in the last 5 years (most of which was not included in the study). For the older refs, we are also talking about guys that spent much of their careers reffing predominantly white sport while most of the younger guys have called games with predominantly black players.

When you read the paper, their main claim is that white players benefit disproportionately from white refs (although blacks also see a smaller benefit). They also say that white players foul more than black players (but this is mostly due to size), bench players foul more per 48 minutes than starters, and that all stars foul less than non-all stars.

All of their results are also based on the idea that refs are assigned using a vaguely random methodology. If we loosen that assumption, there are any number of theories (including conspiracies) about how ref crews are composed and assigned and any number of them could invalidate the data collected...

Posted by dan k | May 2, 2007 12:11 PM

We don't have to guess, you know. "Personal fouls" is a stat kept by the league. When I'm less busy I'll look up the PF rates for Gasol and Wallace.

Posted by too many steves | May 2, 2007 12:19 PM

I'm sure they included a control variable for position, that's very easy to do.

The effects they find are quite small. I really think there is an issue here with the inherent academic bias to find *something*. These guys maybe a year coding and gathering data. If they don't find any effect, there is no publication, no headlines, and the whole year is wasted. Add to this an increasing tendency of economists to do headline-mongering, and you have a bit of a bias right there. When effects are small, it is very easy to manipulate results in this kind of thing.

Posted by mq | May 2, 2007 12:22 PM

Soulite,

Your comment ("ephus: Yeah, I'm sure that ALL the racist referee's retired and the problem is completely solved in three years. Brilliant fucking argument man.") missed my point about the years of study being 1991 - 2004. I do not know how the results skewed over the course of 14 years, and NEITHER DO YOU, because the study does not present the data. It is possible that the data was consistent throughout the 14 years. It is also possible that the problem was more pronounced during a distinct subsection of the study period. I think it is a failing of the presentation not provide that information (which the authors obviously had). That information would not only have provided a trendline, but also shown the amount of year-over-year noise. A low amount of noise (either because there is a clear trendline or because the numbers remained constant throughout the survey period) would tend to support the hypothesis that referees were more likely to call fouls on opposite race players and less likely to call fouls on same race players.

The second big problem I noted (which you do not address) is that the data set used in the study does not identify which of the refs called each foul (that data is not provided by box scores). For games with three or zero white refs, this is a lesser point, although important to help determine if it is a systemic or individual problem. For games with one or two white refs, this entirely confounds the analysis, because we simply do not know the race of the ref who called the foul.

Posted by Ephus | May 2, 2007 12:23 PM

By the way, here's a much larger case of self-defeating ethnic bias in professional sports that has gotten little publicity:

http://www.isteve.com/Sports_Baseball_Hidden_Ethnic_Bias.htm

Posted by Steve Sailer | May 2, 2007 12:53 PM

It says a lot about the arbitariness of racial definitions that Okur is white, but Kidd is black.

Posted by Noah | May 2, 2007 12:53 PM

Sometimes I feel something close to affection for Al.

BTW, I had been trying to come up with some joke about this and the Stockholm Syndrome, but just couldn't.

Posted by Al | May 2, 2007 1:03 PM

Also, nobody's brought up Timmy Duncan/Joey Crawford yet. Any takers for discussing the underlying reason for Crawford's T's? Ahem.

Posted by Al | May 2, 2007 1:12 PM

Also, nobody's brought up Timmy Duncan/Joey Crawford yet. Any takers for discussing the underlying reason for Crawford's T's? Ahem.

I have thoughts on how I have a hard time imagining racial prejudice playing into that one, but I can't think of a remotely defensible way to state them. Ahem.

Hey, the comments format is better!

Posted by Quarterican | May 2, 2007 1:22 PM

Wow. Finally an explanation for the success of the Larry Bird celtics.

Posted by Uncle Horace | May 2, 2007 1:42 PM

It says a lot about the arbitariness of racial definitions that Okur is white, but Kidd is black.

What would the case be for labelling Mehmet Okur as black?

By the way, Sailer's link above is actually kinda interesting, but it also contains the following gem for those who enjoy ridiculing him as race-obsessed:

Look, I noticed back about 17 years ago, while I was reading Bill James' "Baseball Abstract 1986," that Hispanics walked less on average than other players. It was obvious just from casually scanning the statistics.

Classic! But seriously, check out the link anyway.

Posted by Steve | May 2, 2007 2:02 PM

"To see what is in front of one's nose needs a constant struggle." - Orwell

I realize that respectable people try to be as ignorant as possible about patterns of human behavior, but it makes you so boring...

Posted by Steve Sailer | May 2, 2007 2:15 PM

What would the case be for labelling Mehmet Okur as black?

Presumably it wasn't this, but some perception along the lines of "Arabs aren't white, Okur is Turkish, and Turkish is like Arab"...which, uhm, sort of fits with the way lots of people think about these things, but has no logical consistence (which exactly fits with the way lots of people think about these things). Me, I learned All About Race from the 1962 World Book Encyclopdia, so I know that Turks, Arabs, Persians, and Indians (dots) are all Caucasoid.

Posted by Quarterican | May 2, 2007 2:18 PM

The 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.

http://www.isteve.com/Sports_Baseball_Hidden_Ethnic_Bias.htm

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

Posted by Steve Sailer | May 2, 2007 2:28 PM

Please! Any All-White Team has to start Yao at the pivot. That's just common sense.

Posted by Hedo at the five? | May 2, 2007 2:36 PM

What I don't understand from the article is WHY this study is supposedly better than the David Stern study showing that there is no bias. We are told that some economists said that's so, but how am I to evaluate that claim, other then through the appeal to authority?

Listening to the experts on a subject really (aka "appeal to authority") doesn't seem like a bad way to evaluate an issue.

Posted by Matt Weiner | May 2, 2007 2:36 PM

Steve Sailor: In order for your study to deserve the same level of attention as the NBA one, you would have to show that Latin players were preferred simply because they were Latin (whatever that means--born in L.A.? Latin last names? Does Eric Chavez count? He can barely speak spanish). The excerpt you posted merely suggests that they were given priority because of their performance (and, yes, sabermetrics suggests that their performance was overvalued). It doesn't show that Latin player A gets more money than white player B with the same statistics. I don't see what the problem is.

Posted by bc | May 2, 2007 3:23 PM

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

Maybe, just maybe, it's because accusations of actual, active, direct racism are generally newsworthy, whereas accusations of poor recruiting judgment that loosely correlate with ethnicity are not nearly as newsworthy.

I mean, I understand that this is difficult for you to grasp, but did it occur to you that pre-moneyball baseball GMs picked Latino players because they were biased in favor of middle infielders with good speed and high batting averages, not because they went around saying "I gotta get me one a those Dominican shortstops. I hear they're really athletic and swing the bat well."

I mean, your article really is quite interesting, but the only actual bias it describes is a bias in favor of batting average over on-base percentage, not an actual bias against American-born white and black athletes.

Posted by LaFollette Progressive | May 2, 2007 3:24 PM

Look, in discrimination law, it's called "differential impact" when an objective factor affects different ethnicities differently in hiring decisions. The Supreme Court has called for close legal and bureaucratic scrutiny of factors causing differential impact in employment, such as the use of objective standardized tests (Briggs v. Duke Power, 1972). Congress made this statuatory in 1991. Thus, a firm has to spend a lot of money and time to document the "business necessity" of using a standardized cognitive test in hiring, or it will be sued for discrimination by the EEOC and DOJ.

Bill James has been telling us for decades that the baseball old guard's emphasis on batting average over on-base percentage is irrational, and it's been even more obvious for decades to anyone with a decent knack for pattern recognition that Latin players' batting averages are inflated relative to on-base percentages. So, relative to black and white American players, Latin players have been differentially benefiting from an irrational hiring standard

And yet, over all these years in which sabremetrics was enjoying a good press, who mentioned this in public?

Posted by Steve Sailer | May 2, 2007 4:50 PM

In contrast, imagine how much indignant media publicity there would have been if new wave statisticians like Bill James had found that the baseball old guard was inadvertently discriminating _against_ a minority by using the wrong statistic?

Indeed, the issue was never raised in the press until the Toronto Sun launched a major attack -- their 2003 "White Jays" series -- on Moneyball analysis for causing more whites to be hired. (My UPI article "Baseball's Hidden Ethnic Bias" was a response to that series.)

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."

Posted by Steve Sailer | May 2, 2007 5:15 PM

Actually, I am pleased with the study. Most of the referees, black or white, are probably a bit older than most posters here. Everybody has prejudices. They get taught to you when you're a kid, and they never go away. The older you are, the more likely it is that you were exposed to significant prejudice. We learn that overt racism is wrong, but we never erase the seeds that were planted.

The levels of racially motivated foul calling are probably real - a 95% confidence interval of 2.5-4.5% is meaningful. But these levels are so low (even lower for black refs toward white players) that they are probably just a residue of subconcious prejudice that involves no malice. Compared to the overt, conscious racism in hiring decisions in many industries, it's nothing. It will probably be an unmeasurable quantity soon, if it isn't already. No method of redress will be more effective than waiting.

Posted by Njorl | May 2, 2007 5:21 PM

By the way, is there any real evidence in the study that the disparity between white and black officiating is caused by the white officials discriminating? How do we know the black officials aren't the ones who are discriminating? Or that perfect refs would fall in between the two. That the problem is caused by white prejudice is everybody's assumption, but that sounds more like a characteristic 21st century prejudice than something that can be be objectively proven in this study.

Posted by Steve Sailer | May 2, 2007 5:49 PM

Actually, I think subconscious prejudice that involves no malice is still very disturbing, because it's harder to fight than outright bigotry, but it still hurts the subjects of the prejudice for no good reason. If we could be sure that this was declining to zero soon, it'd be OK, but I think we need more data for that.

And of course the important thing isn't how this effects the NBA, but how similar subconscious prejudices might be at work somewhere else. It can be hard to detect this sort of thing, because there won't be any obvious decisions determined by race; but if marginal calls tend to go for white people and against black people, it still hurts black people's life chances. And that's bad. And yes, some explicit affirmative action might be called for to counteract it.

Posted by Matt Weiner | May 2, 2007 5:55 PM

Sailer, READ THE SECOND PARAGRAPH OF THE LINKED ARTICLE.

Posted by Matt Weiner | May 2, 2007 5:56 PM

Oops, third paragraph.

Posted by Matt Weiner | May 2, 2007 5:57 PM

Agree with Njorl. Even if the study is valid, it's quite a small effect.

Moreover, the idea that a team could gain an advantage by getting more white players is ridiculous. Anybody who saw the GS-Indiana trade could see that Indiana's acquisition of two white players was a debacle, whereas GS's acquisition of two black players was incredibly beneficial. I don't think this example is a coincidence.

Posted by Al | May 2, 2007 6:02 PM

I think Steve Sailor's comments are best summed up by Futurama.
Prof. Farnsworth: "The problem with both parties is that they give all your money to the less fortunate."
Fry: "Man, the less fortunate get all the breaks."

Not to say that professional baseball players--of any stripe--are disadvantaged, but I'll be more concerned about the disparate impact on white MLBers when white people make up less than 95% of our economic and governing elite.

Posted by bc | May 2, 2007 6:14 PM

Sailer, quite a few people have discussed the changing demographics of baseball, and you're hardly the first to note the free-swinging nature of many "Latin" players. Billy Beane's sabermetrics have, if anything, received too damn much attention in the press. If there were evidence that baseball general managers had been consciously discriminating against white players, that would be big news. There isn't.

For starters, this is "discrimination" based specifically on skills that are only relevant to baseball. We aren't talking about an issue with widespread repercussions on economic opportunity. Like, say, perceptions about intelligence and race.

But more fundamentally, it's difficult to see how the "differential impact" standard would apply here. Administering an IQ test during a job interview is a fairly arbitrary hurdle, unless one's job actually requires one to solve logic puzzles within strict time restrictions. It makes sense for the military, which recruits people straight out of high school and sets them loose with multi-million dollar equipment, to screen out people who can't follow directions. But there are good reasons to be skeptical about widespread use of cognitive tests in hiring.

Batting average and stolen base statistics, on the other hand, are job performance metrics that inarguably relate to the overall performance of the business. These stats are analogous to sales quotas on Wall Street. The "business necessity" speaks for itself. It's the sort of "discrimination" that someone like Jesse Jackson could make into a big media issue if he really wanted to, as I suppose you are trying to do in your role as the blogosphere's leading white race-obsessive. But I don't see any realistic grounds for government intervention.

Old-school baseball managers were using the wrong metric, and nobody really understood this until the problem had already started to correct itself. Now that IS interesting. But is it really grounds for the government to regulate baseball teams to make sure their general managers aren't boneheads? Are you a Dodgers fan, by any chance?

The "White Jays" article was foolish, and it received some well-deserved scorn. The revolution in baseball sabermetrics has been exhaustively covered. The link to ethnicity is just an interesting sidenote. You seem to be suggesting that there is a conspiracy of silence. I'm sorry your UPI article wasn't picked up by more newspapers, but then I don't think anyone has taken the UPI seriously since it was purchased by the Moonies, so don't take it personally.

You also might have a clearer view of what's in front of your face if you occasionally removed those flesh-tone glasses you wear.

Posted by LaFollette Progressive | May 2, 2007 6:18 PM

The economists write: "Unfortunately our framework is not well-suited to sorting out whether these
results are driven by the actions of black or white referees."

They then go on to offer a much weaker level of analysis than the rest of their paper somewhat more implicating white refs than black refs, but as they would be the first to admit, it's hardly conclusive.

That the NYT jumped all over this as evidence of white racism is just more evidence of the deep hunger in the media for examples of whites behaving badly that was so evident in the NYT's shameful behavior in promoting the Duke lacrosse hoax.

Posted by Steve Sailer | May 2, 2007 6:19 PM

As a long time critic of the government's disparate impact lawsuits, I'm obviously not calling for the EEOC to sue for discrimination baseball teams that don't implement Moneyball. What I'm documenting here is the kind of media bias that helped cause the NYT to cheer on the Duke frame-up.

For decades, it was was obvious to any sophisticated baseball fan with eyes in his heads that baseball teams were systematically overrating one ethnic group -- which had big implications becaue it entered into expensive decisions such as whether to build huge sleep-in recruiting "academies" in the Dominican Republic -- but _nobody_ would mention it in the media until the lame-brains in Toronto got up on their high horse about it and demanded the "White Jays" go back to old guard stupidity.

Why the media silence out of all the zillions of words written about baseball statistics? Because it was a minority that was benefiting from old guard stupidity, and there's very little market for facts like that just because they are true.

Posted by Steve Sailer | May 2, 2007 6:33 PM

The most ridiculous aspect of this debate is that black American baseball players are better at getting on base than white American baseball players (and better at hitting for average, and better at hitting for power...). The four best OBP guys of the last two decades are Barry Bonds, Frank Thomas, Rickey Henderson, and Wade Boggs. Only one white guy, and he was a white guy who had no other skills besides getting on base. The reason why you see fewer black hitters who draw a lot of walks is that there's a comparatively small number of black baseball players overall (compared to other sports, that is).

"Latin American" is not a race, it is a national origin, which is different. Javy Lopez sucked at walking. Does this tell us anything about race? Obviously not.

Posted by JP | May 2, 2007 7:34 PM

Before criticizing my article, you might want to try reading it:

http://www.isteve.com/Sports_Baseball_Hidden_Ethnic_Bias.htm

As I wrote: "Indeed, African-Americans have provided many of the most patient hitters and highest percentage base stealers, such as Joe Morgan, Bonds (whose unbelievable-sounding 198 walks last year broke Ruth's single season record), and Rickey Henderson (who broke Ruth's career walks record)."

As I further wrote: "Nevertheless, because the Latin tendency toward not walking is clearly not a racial difference (Latin ballplayers come in all colors), their free swinging may change in the future. For example, if Latin players are failing to get walks merely because they lack role models at an impressionable age, that's not set in stone."

Posted by Steve Sailer | May 2, 2007 7:42 PM

Nope, not going to read any of your stuff, Sailer.

If you're going to concede those points, then you had no reason for introducing the issue into this debate.

Posted by JP | May 2, 2007 7:53 PM

"For decades, it was was obvious to any sophisticated baseball fan with eyes in his heads that baseball teams were systematically overrating one ethnic group "

Really? I don't remember reading that anywhere. Can you point to any sophisticated baseball fans (besides yourself, of course) who watched games and thought that latinos were overrated? And any evidence that teams were engaging in this discrimination "systematically"?

Posted by Mavis Beacon | May 2, 2007 8:30 PM

"Really? I don't remember reading that anywhere."

Of course you didn't read that anywhere. That's my point.

Nobody publicly drew the implications of

1. Bill James' emphasis from the late 1970s onward of the underrated value of drawing walks combined with

2. The Latin hitter's traditional mindset that "You can't walk to America" ... because you just don't say things like that.

What you do say is that the Duke lacrosse team represents blah blah blah ... Who cares about the truth? The point is to write fashionable things to establish your high standing in the moral pecking order.

Posted by Steve Sailer | May 2, 2007 8:43 PM

Sailer, the crucial flaw in your argument is that it was most emphatically NOT obvious to everyone that Bill James was right about walks being undervalued. If teams had believed this, they would have followed through on that knowledge. Baseball teams want, above all, to win. To insinuate that 20-odd professional sporting franchises deliberately followed counterproductive strategies as part of some politically correct scheme to benefit a minority group at the expense of white Americans, if this is indeed what you are trying to claim, is one of the most brazen lunacies I've ever heard.

It was not until the past 10 years or so that this idea was widely accepted by people who actually make personnel decisions in major league baseball. And, even now, not every team has really taken the Moneyball analysis to heart, and the basic principles have yet to prove themselves in the postseason.

Furthermore, it hardly needs to be said that very few people in America view every single aspect of life through the prism of race. The Oakland A's and Toronto Blue Jays did not build winning franchises by rejecting folk wisdom about the superiority of Latin players and promoting heretofore undervalued white and black players. To the limited extent this happened at all, it was an accidental side effect of their decision to follow a different method of evaluating player performance. The wrongheaded reaction of the Toronto paper gives you ample fodder for a column railing against political correctness. Your conspiracy theory and efforts to link this to the Duke case, on the other hand, go substantially further toward the opposite pole of crazy.

Posted by LaFollette Progressive | May 2, 2007 10:13 PM

Not that this has cropped up in this thread much, but the lack of understanding regarding simple ceteris paribus logic has been astounding in this debate. Listening to three separate sports radio programs this afternoon (don't ask, I did a lot of driving around today), I heard each of them state something to the effect of "of course there are more foul calls on African-Americans, since they make up 70% of the players." Then Charles Barkley on TNT and Michael Smith on Around the Horn turned apoplectic making the same point. Well, duh. But read the study and think a little bit about what regression does. Yes, interaction terms are slightly higher order thinking, but still, come on...

Posted by RatChoicer | May 2, 2007 10:44 PM

Oh, and I almost forgot. All this talk of "The NBA has better data, so I believe them" is contrary to standard efforts of determining the truth. My guess is that the NBA has better data, too. But I don't believe them unless they make the data public and let others take a crack at the analysis.

As to the NBA's claim that the data are private and/or concerned with personell matters, I call BS. This data is private in the way that the number of times Kramer enters Seinfeld's apartment in a "humurous" fashion is private. All someone has to do is go back and watch the games and code foul calls. That the NBA has done this doesn't give it claim that the data are private, just that it is the only one that has done the data collection.

Anyway, this whole thing is (obviously) bothering me because people aren't really thinking about what the study says. They are just falling back to their preferred "NBA is good/NBA is bad" stance.

Posted by RatChoicer | May 2, 2007 10:50 PM

Oh, and I almost forgot. All this talk of "The NBA has better data, so I believe them" is contrary to standard efforts of determining the truth. My guess is that the NBA has better data, too. But I don't believe them unless they make the data public and let others take a crack at the analysis.

As to the NBA's claim that the data are private and/or concerned with personnel matters, I call BS. This data is private in the way that the number of times Kramer enters Seinfeld's apartment in a "humurous" fashion is private. All someone has to do is go back and watch the games and code foul calls. That the NBA has done this doesn't give it claim that the data are private, just that it is the only one that has done the data collection.

Anyway, this whole thing is (obviously) bothering me because people aren't really thinking about what the study says. They are just falling back to their preferred "NBA is good/NBA is bad" stance.

Posted by RatChoicer | May 2, 2007 10:51 PM

I just read through the comments on this blog about Sailer's "Hidden Ethnic Bias" theory. There was a lot of argument, but no one challenged Sailer's facts, which turn out to be false.

As it turns out, by his own statistics, he's full of crap. According to his stats, Hispanic players had an OBP of .306; non-Hispanics had an OBP of .308. Meanwhile, from year to year in since 2000, the standard deviation of the entire league's OBP has been .006, three times greater than the difference cited by Sailer.

What a foolish theory. I'll be posting a more about this later on my blog (with additional data), but for now if you're interested, check out:

http://www.jedreport.com/2007/05/freeslinging_fu.html

The most important thing to remember is that when you're debating with guys like Sailer, never, never, never accept their factual assertions at face value.

Posted by jedster | May 7, 2007 9:39 PM