I have not read the study and thus can't comment on its rigor, but this conclusion is not surprising; a similar study found a same-race bias in NBA officials. Hopefully the umps don't prove to be corrupt as well.

there are black pitchers?

2.1 million sounds like a lot of pitches.

When did Time buy out BASN?

It doesn't happen all the time — in about 1% of pitches thrown — but that's still one pitch per game, and it could be the one that makes the difference.

2.1 million sounds like a lot of pitches.

I think it's around 2 full MLB seasons.

They should do a control study where they blindfold the umps.

Please release the full study. I don't disbelieve, but these things sometimes are REALLY tough to analyze properly, and poutside analysis and corroboration would be a very good thing.

but more importantly, <u>Getting Questec data would be even (MUCH!) better</u>.
There is no end to the useful studies that could be made with full accounting of how umpires' strike zones correlate with a bajillion variables. Anybody know if MLB will release the data sometime?

2.1 million sounds like a lot of pitches.

Nahh, that's like a month and a half for Livan Hernandez.

did this study control for Letroy Hawkins?

The article states that the effect goes away when Questec is used.

2.1 million pitches divided by 3 years divided by 30 teams divided by 162 games is about 144 pitches per team game... sounds about right.

If you're a baseball umpire, you may be a rapist.

Blacks may not have some of the necessities to be, let's say, a strike-calling umpire.

So in the extreme case, white starter vs black starter, with a white umpire, the white starter can expect to get maybe 1.5 strikes more than he "should" get? I can live with that.

Anybody who saw the width of Livan Hernandez's strike zone against the Braves in the 1997 play-off game knows this cannot be true.

Hamermesh’s team analyzed the calls on 2.1 million pitches thrown in the Major League

I wonder if that's a typo or if Bud's program to completely meld the AL and NL is achieving success...

I have not read the study and thus can't comment on its rigor, but this conclusion is not surprising; a similar study found a same-race bias in NBA officials. Hopefully the umps don't prove to be corrupt as well.

That NBA study was a joke. Just like I'm sure this one is.

Anybody who saw the width of Livan Hernandez's strike zone against the Braves in the 1997 play-off game knows this cannot be true.

Eric Gregg was the ump. I can't comment on what Livan Hernandez is for purposes of this study because I've been yelled at for saying that Latinos were Black in the past . . .

Something something Joe Torre?

well, then we know that really fat guys (Gregg was easily over the stated 315 lbs., I believe) favor hispanics in their strike zones...

Did Gary Sheffield certify that these findings are for "real" black players? I'd hate to see some Mexican like Orlando Hernandez skew the data.

That NBA study was a joke. Just like I'm sure this one is.

That NBA study was not a joke; it found a slight tendacy in the foul-calling patterns of referees to favor their own race. I DID read that study, and it's reasonably conclusive. Again, I can't comment on this one, but am not surprised. People tend to be inherently racist; take this test and the results may surprise you. Or not.

That NBA study was not a joke; it found a slight tendacy in the foul-calling patterns of referees to favor their own race. I DID read that study, and it's reasonably conclusive.

The problem with both studies is that the effect of the bias is so minor that it's within the margin of error.

That NBA study was not a joke; it found a slight tendacy in the foul-calling patterns of referees to favor their own race. I DID read that study, and it's reasonably conclusive. Again, I can't comment on this one, but am not surprised. People tend to be inherently racist; take this test and the results may surprise you. Or not.

The NBA study was a joke. It did not account for which ref called what fouls. It looked at groupings of refs and used statistics to make its claims. It clearly stated in the paper how they gathered there data, and nowhere in that paper did it state they either watched games to judge refs calling plays, or did they go through referees individual stats and calls.

The thing was a joke used to generate publicity. I'm sure this is much of the same. I haven't read it, but I bet they didn't look at the variance in a ump's strike zone. They probably didn't compare him to Questec. They probably didn't compare him across nights or in the same game, or even against the same pitcher later in the season.

That NBA study was not a joke; it found a slight tendacy in the foul-calling patterns of referees to favor their own race. I DID read that study, and it's reasonably conclusive. Again, I can't comment on this one, but am not surprised. People tend to be inherently racist; take this test and the results may surprise you. Or not.

The study didn't look at what official called what foul. All it did was look at box scores, and the refs doing the game. So if a black ref doing the game with two white refs and calls a foul on a black guy, this foul was 'credited' to a white crew. Awesome methodology there.

I didn't read the NBA study in its entirety, but it seemed to be rather well-done from what I saw. Of course, the fact that it didn't know which ref made which call was a limitation, but there's only so much they could do with the data.

Guts, your link didn't work; was that the test where all the "good" things are on the right and the "bad" things are on the left, and then the "white" is on the right and the "black" on the left, and then they start reversing it all, and you have to click in all these different places? I thought that test was horrible; it seemed to have a lot more to do with how coordinated one was than how racist one was. (And patient, the gorramn thing went on forever.)

The study didn't look at what official called what foul. All it did was look at box scores, and the refs doing the game. So if a black ref doing the game with two white refs and calls a foul on a black guy, this foul was 'credited' to a white crew. Awesome methodology there.

Just re-read that study, they also judged if a ref was white or black by "appearance". That introduces a nice bias into the study at the get go.

Just because something is printed in a peer-reviewed journal doesn't mean it's good. There's lots of #### out there. This NBA ref study is one of the higher profile ones.

Beat me to it Jimmy.

I think it is better to just ignore any thing coming out of academia where sports is involved. Another example is the study from Wharton that says 500 college basketball games with a point spread higher than 12 were fixed over the last dozen years. Shocking stuff there: big favorites don't cover as often as they should. Anyone with a real world's experience in gambling could give the reasons. Instead, it is because the games are 'corrupt'.

I thought that test was horrible; it seemed to have a lot more to do with how coordinated one was than how racist one was. (And patient, the gorramn thing went on forever.)

Not only that, but it also accounts for practice. This is one of the hardest variables to account for in any testing situation, especially in psych and sociology. Anytime you give subjects tests repeatedly, you are automatically influencing your data because you are allowing practice.

It builds together associations, and then does a poor job of breaking them. It also uses the same faces and words over and over, so you don't even start looking at the category. Your brain starts linking certain things with "right hand" not "Good or Bad". In my mind, that test would've been the same if they would reverse the pictures and the words from the get go. Again, when you start linking these, you introduce a bias that is dependent on how you introduce the data. Shitty test.

Shibal, why is it that big favorites don't cover as often as they should? I'm sincerely curious.

That NBA study was not a joke; it found a slight tendacy in the foul-calling patterns of referees to favor their own race. I DID read that study, and it's reasonably conclusive. Again, I can't comment on this one, but am not surprised. People tend to be inherently racist; take this test and the results may surprise you. Or not.

The link doesn't work, but I found the test anyway. And the test is complete horseshit. For those of you who won't want to bother wasting your time with it, the methodology is as follows:

Establish a left-hand association with European-American faces and "good words," and a right-hand association with African-American faces and "bad words." Then jumble everything up, and proclaim the test takers racists when they take a longer time to associate African-American faces with the same hand as "good words."

This is an example of a test written to get specific predetermined results. Thanks for posting the link, very interesting.

EDIT: Looks like I was beaten to the punch.

I think it is better to just ignore any thing coming out of academia where sports is involved.

You'd be ignoring a lot of very good work.

-- MWE

I think it is better to just ignore any thing coming out of academia where sports is involved.

It's not just sports, though. There's lots of junk coming out of academia about everything, and a lot of it is just that, junk. You just have to read through things yourself and judge. Don't depend on ESPN, Time, or any other "mass media" to get your research from. They aren't in the business of spreading knowledge, they're in the business of making advertiser dollars.

Establish a left-hand association with European-American faces and "good words," and a right-hand association with African-American faces and "bad words." Then jumble everything up, and proclaim the test takers racists when they take a longer time to associate African-American faces with the same hand as "good words."

Gladwell, in Blink (the best selling book), says the test is valid. "What does it mean? Does this mean I'm a racist, a self-hating black person? Not exactly. What it means is that our attitudes toward things like race or gender operate on two levels. First of all, whe have our conscious attitudes. This is what we choose to believe. These are our stated values, which we use to direct our behavior deliberately. ..... But the IAT measures something else. It measures our second level of attitude, our racial attitude on an unconscious level- the immediate, automatic associations that tumble out before we've even had time to think. We don't deliberately choose our unconscious attitudes.

The study didn't look at what official called what foul. All it did was look at box scores, and the refs doing the game. So if a black ref doing the game with two white refs and calls a foul on a black guy, this foul was 'credited' to a white crew. Awesome methodology there.

And when black crews are more lenient to black players than white crews? You're missing the forest for the trees.
The Wolfers-Price study is very strong.

If Gladwell believes it, then it probably isn't true. Gladwell loves to get publicity writing about interesting findings, even when they lack any rigor whatsoever. He also tends to present intensely studied, nuanced concepts with superficiality. Decent writer I guess, but not a rigorous scholar.

Gladwell, in Blink (the best selling book), says the test is valid. "What does it mean? Does this mean I'm a racist, a self-hating black person? Not exactly. What it means is that our attitudes toward things like race or gender operate on two levels. First of all, whe have our conscious attitudes. This is what we choose to believe. These are our stated values, which we use to direct our behavior deliberately. ..... But the IAT measures something else. It measures our second level of attitude, our racial attitude on an unconscious level- the immediate, automatic associations that tumble out before we've even had time to think. We don't deliberately choose our unconscious attitudes.

Gladwell's a journalist, not a scientist.

The IAT that I took just tested associations. I could have mapped anything to the words good or bad, and then reversed them. It would have given the exact same result. The minute I saw the test, I knew what it was doing. It started out associating European with 'Good'. What would have happened if these were reversed? If they were reversed would I have gotten every word wrong because instead of me thinking "Good words -> right hand" I would have thought "White -> Good"?

If the IAT wanted to be better, but still not totally robust and not passing the smell test, it should have done the test with European -> Right Hand, the Good -> Left, the European OR Good -> Right. But, with European and Good both mapping Right individually, and then linking them, and then reversing one, there methodology goes right out the window. They take 5 minutes to teach you an association without thinking ("Go as fast as you can!"), and then reverse and expect you to get it right on the first try? Sure.

It's an interesting book and a quick read.

...The disturbing thing about the test is that it shows that our unconscious attitudes may be utterly incompatible with our stated conscious values. As it turns out, for example, of the 50,000 African Americans who have taken the Race IAT so far, about half of them, like me, have stronger associations with whites than with blacks. How could we not? We live in North America, where we are surrounded every day by cultural messages linking white with good. "You don't choose to make positive associations with the dominant group," say Mahzarin Banaji, who teaches psychology at Harvard University and is one of the leaders in IAT research. "But you are required to. All around you, that group is being paired with good things. You open the newspaper and you turn on the television, and you can't escape it."

The Wolfers-Price study is very strong.

The result may or may not be strong. The methodology sucks.

One pitch out of a hundred sounds much smaller than the variance between umpires, or between the same umpire on different days.

People do find smaller signals in bigger noise, but it's not trivial. An old physicist joke is that half of the "five sigma" results that get published are wrong. I'm not sure of the mechanism, either: many more than 1% of pitches are borderline strikes or balls. It seems like if "unconscious racism" were a real effect, it would show up at the same order of magnitude as the set of pitches it affected.

If anybody runs across a preprint, please link to it.

The result may or may not be strong. The methodology sucks.

The methodology could be stronger, but multivariable regression analysis using a very large sample of data and controlling for dozens of factors is quite good.

Just re-read that study, they also judged if a ref was white or black by "appearance". That introduces a nice bias into the study at the get go.

Digging this deep... you yourself may be biased?
It seems very unlikely that any such appearance "bias" would harm the study's findings.

Crap! So I wasted my time reading that book!?!?! I expected more from a staff writer of the New Yorker, and a former business and science reporter for the Washington Post.

If anybody runs across a preprint, please link to it.

Just search for the author's name in Google.

Well, without reading the study, I can't say I've learned anything about NBA/MLB official racism in this thread, but I have learned a lot about people not understanding how scientific studies work.

I'm less interested in studies on race than interested in what it says about people that "can live with" racism not directed toward them or need to respond with illogical snide remarks toward any attempt to test, investigate or discuss the issue.

You'll notice, there hasn't been a single post on the subject of the opposite knee-jerk type--that umpires are crazy supremacists.

“Across all of these specifications we find that black players receive around 0.12-0.20 more fouls per 48 minutes played (an increase of 2.5-4.5 percent) when the number of white referees officiating a game increases from zero to three.”

Care to explain this, Jimmy P?

How many black umps are there? Not to dismiss this study out of hand, but the stated difference could be simply the result of one or two black umpires refusing to let Maddux, Glavine, and the like redefine the strike zone.

The methodology could be stronger, but multivariable regression analysis using a very large sample of data and controlling for dozens of factors is quite good.

I'm not questioning regression analysis, it is strong. Their data collection sucks. Regression analysis doesn't mean a thing if the data you're using it horeshit.

Digging this deep... you yourself may be biased?
It seems very unlikely that any such appearance "bias" would harm the study's findings.

Biased? Probably, I'm human. I'm not going to be one that says I'm 100% without bias, that'd be lunacy. But, when you do a study, you have to try and not interfere with the data as much as possible. Injecting your own thoughts into it (like deciding who is and isn't white or black) introduces a bias. An example: if you saw a color pitcher of Derek Jeter, what race is he? Alex Rodriguez? Michael Jackson? (Alien is not an answer on the last one)

E-X, It seems that you think people have a tendency to avoid having their worldviews rocked. And so they disregard (in various ways, snide remards, for example) information that doesn't conform to those world views. Is that what you're saying?

Care to explain this, Jimmy P?

I haven't questioned the conclusions of the study in this thread. I've questioned the entire study (the study being the NBA one, I have not read the MLB ump one) because their methodology was horrible. Now, the only place their methodology actually would work, would be to compare the games called by three refs of the same race. At that point, you'd have a sensible study from how they've designed it. But, that's not how they've printed their study. Even if you extrapolate the data that I've proposed from their study, you may run into a sample size problem because how many games are called by an officiating crew comprised of one race?

I recently became aware of some of the complexities of race and identity. Since then, I feel that a person's identity and own self-perception of that identity is so complex that studies of race might not have very much validity.

Haven't read the study. Basketball refs probably have always been gym rats. Gym rats, as a whole, are less likely to associate "bad" with "black". And, alternately, are more likely to associate "good" with "black." Therefore, a higher frequency of calls against blacks isn't likely

In such a study I would like to see self-percaptions of race, IAT results and other indicators of racial attitudes prior to the game.

I wonder what happens when you remove CB Bucknor from the dataset?

I don't think he's racially biased, but his strike zone is essentially a random variable in a very small sample of black umpires.

Bucknor's not black. He's Jamaican.

Jimmy, if the authors of the NBA study assigned a race to each referee based on appearance before seeing who their referee teams assigned fouls to, how does that introduce bias into the study? This wasn't bad methodology on their part. They worked with the best data that they had available to them as all social scientists are forced to do.

Also, why is only the data from referee teams that are all of the same race relevant? If there's a statistical significant difference in foul calls between a team that has two white and one black and another team that has two blacks and one white why isn't that result also meaningful? Granted, it opens up another possible explanation: that white/black refs make calls differently depending on the race of the referees that they're teamed with.

btw, I agree with Raskonikov that Gladwell isn't rigorous in his analysis in either Blink or The Tipping Point (I don't think he'd claim that he is) but that doesn't mean that he's wrong.

That test is obviously crazy. It's not racist to say that, it's just logical. But I question, on first impression, whether there are enough black pitchers or umpires in MLB for the results of this study to be distinguishable from random.

Does this study distinguish between blacks from America and ones from Latin America? Between light-skinned Americans and Latin Americans? It's always seemed to me that latino pitchers--regardless of racial makeup--are more likely than any others to develop a reputation in the msm for lacking control, or "choking" in key spots. A study of whether that has to do or doesn't have to do with umpires' biases could be interesting if done properly.

And are the controls for "pitcher's tendency to throw strikes" adequate or even correct? That could be the reason for the "anti-Asian" bias easily, since a good proportion of the Asian pitchers who have come here have tended to be nibblers--again, at least to hear the msm report matters.

A few years ago, a white Australian tennis player, Lleyton Hewitt, accused a tennis official of making calls against him during a match because the official was black and Hewitt's opponent, James Blake, was (at least in part) black. Hewitt was raked over the coals for making this complaint, of course. He was called a racist and severely rebuked.

But since that time, tennis has adopted a really good replay system. It has almost completely eliminated the possibility of racial bias or judicial incompetence affecting a tennis match. (And it has gotten rid of the John McEnroe-like antics of the kvetchers in that sport.) It's a shame that MLB is too obdurate to consider a limited replay challenge system for all of its calls, and using Questec to call all balls and strikes.

My feeling is that for every degree of human nature which is racially biased, human incompetence is ten times higher. In other words, while bias may account for some degree of missed calls in baseball, human frailty accounts for many times more.

Does this study distinguish between blacks from America and ones from Latin America? Between light-skinned Americans and Latin Americans? It's always seemed to me that latino pitchers--regardless of racial makeup--are more likely than any others to develop a reputation in the msm for lacking control, or "choking" in key spots. A study of whether that has to do or doesn't have to do with umpires' biases could be interesting if done properly.

And are the controls for "pitcher's tendency to throw strikes" adequate or even correct? That could be the reason for the "anti-Asian" bias easily, since a good proportion of the Asian pitchers who have come here have tended to be nibblers--again, at least to hear the msm report matters.

Vaux, I think these factors that you mention would be problems for the study if the question was "Who race gets more strike calls?" but it doesn't sound like that's what they did. They took those black pitchers and looked at whether they got more strike calls with a black ump or a white ump and then they did the same thing with white pitchers.

well, I think umpiring has improved since questec so while MLB doesn't have a replay system it has made progress.

also, someone mentioned earlier that this is a small effect and that raises a sabermetric question that I don't know the answer to. How many (random) ball/strike changes is equivalent to a run? We know that, as a rule of thumb, 10 runs over the course of a season is equivalent to a win but how many addition strikes does it take to remove a run?

Forget the study and the inflammatory headline -- as one or two people pointed out above, real bias exists because of unconscious factors that influence our behavior.

I think that's the hard one for most people to accept -- at least when it applies to you, personally.

As for myself, I think I have an affinity for Black folks, but yesterday I was outside of my usual environment and was walking up the street behind three young Black men, all bigger than me, who were slowly walking up the street in the same direction. When it became apparent that I was going to have to walk right through and past them, the temptation was great to walk around them to the edge of the street, or to cross the street altogether, even though it was midday and I was on the main street of the town with plenty of people around.

I walked through their midst, and the only response was that one asked me if I had a light, I said no, and we all went on our way.

This anecdote doesn't relate directly to the link, but both scientific studies, human nature, and my own personal experience tell me that Americans have unconscious fear-based behavior when it comes to interacting with black folks they do not know.

I don't know why that's so hard to admit.

I'm not going to be one that says I'm 100% without bias, that'd be lunacy. But, when you do a study, you have to try and not interfere with the data as much as possible. Injecting your own thoughts into it (like deciding who is and isn't white or black) introduces a bias.

How would you do it?

Shibal, why is it that big favorites don't cover as often as they should? I'm sincerely curious.

SS, there's a few factor involved. #1 reason for me is that usually power numbers are used to make the initial line. If Duke (80 power rating) is playing North Carolina (79 power rating) on a neutral site, Duke would be favored by 1. If Duke is playing Florida State (65 power rating), Duke would be favored by 15. Once the game gets comfortably in hand for Duke, they may pull their starters or sit on the ball without regards to scoring. While the alumni may care who covers, Coach K won't. The incentive is to win...that's it.

Another reason is that the betting public usually bets on the stronger team, so sportsbooks will shade that number a bit higher to draw in the 'smart money' and get more balance in the action. Bettors love Duke, so the betting line may be 16 or 17 when it really should be 13 or 14.

E-X, It seems that you think people have a tendency to avoid having their worldviews rocked. And so they disregard (in various ways, snide remards, for example) information that doesn't conform to those world views. Is that what you're saying?

Exactly. Of course, I'm subject to the same, and have to be careful that I don't simply react snidely to any indication that institutional racism isn't persuasive. Concrete evidence to that fact is pretty sparse though...

It's not even just a disparity in treatment (that people are more likely to scientifically weigh studies that don't address issues they are trying to avoid), look above, there's a small, but steady stream of people trying to find little nitpicks with the study, and it turns out those nitpicks are covered in the methodology.

As Pops asked, the key question is "How would you do it?"

And I don't think that's being sincerely addressed. After all, if we really were interested in the study, but didn't like the methodology, we wouldn't dismiss the findings, just ask for a for a revision in the methodology.

We can even look at one of the precise examples--the natsiness directed toward the researchers by using their own determinations to decide race for the purposes of the study.

Not to be rude, but I believe this shows an extreme shallowness on the part of the critic in understanding the complexities of race in American society.

The idea that only monoethnic ref crews are worthy of testing is bizarre--why not look at mixed crews and compare them as well?

Furthermore, we have this:

But, when you do a study, you have to try and not interfere with the data as much as possible. Injecting your own thoughts into it (like deciding who is and isn't white or black) introduces a bias. An example: if you saw a color pitcher of Derek Jeter, what race is he? Alex Rodriguez? Michael Jackson? (Alien is not an answer on the last one)

And yet people do make these determinations every single day. Yes, treatment according to ethnicity is certainly not uniform, but once again, that's a topic for more testing, not less, and you are basically tossing out any quantitative study whatsoever. After all, no population is uniform and no one fits perfectly into a gender/race/class/sexual orientation box. There are transgendered folks, but that doesn't mean it's idiotic to try to parse for gender in testing.

Jeter, A-Rod and Michael Jackson are certainly not going to face the same experiences that Bernie Mac will, but it's pretty easy to guess what most people in the society would judge them as in terms of ethnicity.

Not to be rude, but I believe this shows an extreme shallowness on the part of the critic in understanding the complexities of race in American society.

And check out that headline.

Time stinks.

This anecdote doesn't relate directly to the link, but both scientific studies, human nature, and my own personal experience tell me that Americans have unconscious fear-based behavior when it comes to interacting with black folks they do not know.

This is certainly supported by some research:


Cunningham, W. A., Johnson, M. K., Raye, C. L., Gatenby, J. C., Gore, J. C., & Banaji, M. R. (2004). Separable Neural Components in the Processing of Black and White Faces. Psychological Science, 15, 806-813.

In a study of the neural components of automatic and controlled social evaluation, White participants viewed Black and White faces during event-related functional magnetic resonance imaging. When the faces were presented for 30 ms, activation in the amygdala--a brain region associated with emotion--was greater for Black than for White faces. When the faces were presented for 525 ms, this difference was significantly reduced, and regions of frontal cortex associated with control and regulation showed greater activation for Black than White faces. Furthermore, greater race bias on an indirect behavioral measure was correlated with greater difference in amygdala activation between Black and White faces, and frontal activity predicted a reduction in Black-White differences in amygdala activity from the 30-ms to the 525-ms condition. These results provide evidence for neural distinctions between automatic and more controlled processing of social groups, and suggest that controlled processes may modulate automatic evaluation.



Ronquillo, J., Denson, T. F., Lickel, B., Lu, Z. L., Nandy, A., & Maddox, K. B. (2007). The effects of skin tone on race-related amygdala activity: An fMRI investigation. Social Cognitive and Affective Neuroscience, 2, 39-44.

Previous work has shown differential amygdala response to African-American faces by Caucasian individuals. Furthermore, behavioral studies have demonstrated the existence of skin tone bias, the tendency to prefer light skin to dark skin. In the present study, we used functional magnetic resonance imaging (fMRI) to investigate whether skin tone bias moderates differential race-related amygdala activity. Eleven White participants viewed photographs of unfamiliar Black and White faces with varied skin tone (light, dark). Replicating past research, greater amygdala activity was observed for Black faces than White faces. Furthermore, dark-skinned targets elicited more amygdala activity than light-skinned targets. However, these results were qualified by a significant Interaction between race and skin tone, such that amygdala activity was observed at equivalent levels for light- and dark-skinned Black targets, but dark-skinned White targets elicited greater amygdala activity than light-skinned White targets.

Several comments:

1. The effect is very small - about one pitch per game, according to the study - and while one pitch *can* be significant, it would be very hard to measure that impact on the game.

2. The effect goes away when considering other factors instead of Questec - the two others mentioned are full counts and well-attended games.

3. While the article says that the study controlled for "all" outside factors, the article sheds no light on the nature of those controls, and I wouldn't be surprised at all to see subtle leakage of those other factors into the study. Age-related or experience-related effects would seem to be one possible leakage - many of the older/more experienced umpires are white.

-- MWE

I saw some of them hanging around with Albi the dragon. They must be racist.

I can't believe that link didn't work. Sorry, everyone. Try this one. Not the best test, as pointed out, but interesting.

I'm sure neither study contains a perfect data set or a perfect methodology; no social science study does. Wolfers-Price made the strongest case they could, based on the data they had. They used data from boxscores; the NBA has data on which ref called which foul, but they're "unlikely" to release it. What they could use is, I think, strong enough to draw some conclusions about the biases inherent in the refs and, dependent on this study's rigor, the umps.

I don't think anyone is suggesting that Tim McClelland is spending the offseason passing out KKK flyers or anything. The study merely says that when Glavine is facing off against Dontrelle, McClelland (note - McClelland is the first ump I thought of, not singling him out for any reason) simply gives a marginally bigger strike zone ot Glavine, to the tune of 1.5 extras strikes per game. This is, as others have presented here, not a totally new idea.

If you disagree with the study, that's fine - I disagree with studies all the time. But the Wolfers-Price one, I think, is pretty well done with the limitations they had. Economists have a hell of a hard time getting data....

Those studies and Wilson's personal testimony support the conclusion made by Gladwell in Blink about unconscious associations. I'm thinking now that the arguments against Gladwell were mere ad hominem attacks. Yah! As if a journalist is inherently disqualified from making sound conclusions.

Paper


FAQ

Those studies and Wilson's personal testimony support the conclusion made by Gladwell in Blink about unconscious associations. I'm thinking now that the arguments against Gladwell were mere ad hominem attacks. Yah! As if a journalist is inherently disqualified from making sound conclusions.

I don't know Blink, so I don't know Gladwell's precise conclusions, but the abstracts I posted suggest several mediating factors which are important. Although most whites may have an automatic fear response to unfamiliar black faces, people differ in the amount of inhibitory control they can exert over this response. So some may be better able at consciously overriding their primitive impulses. Either way, I'm not sure what the relevance is to baseball or basketball officiating, with many other factors at play.

A few years ago, a white Australian tennis player, Lleyton Hewitt, accused a tennis official of making calls against him during a match because the official was black and Hewitt's opponent, James Blake, was (at least in part) black. Hewitt was raked over the coals for making this complaint, of course. He was called a racist and severely rebuked.

But since that time, tennis has adopted a really good replay system. It has almost completely eliminated the possibility of racial bias or judicial incompetence affecting a tennis match. (And it has gotten rid of the John McEnroe-like antics of the kvetchers in that sport.) It's a shame that MLB is too obdurate to consider a limited replay challenge system for all of its calls, and using Questec to call all balls and strikes.

None of which changes the fact that Lleyton Hewitt remains a racist a$$hole, as our friend and Lleyton countrymate Mr. Coorey would be happy to point out.

When the NBA quit complaining and released its superior data, it supported Price-Wolfers. Oops!

Good catch, JCB. Stern is really ####### up recently.

My feeling is that for every degree of human nature which is racially biased, human incompetence is ten times higher. In other words, while bias may account for some degree of missed calls in baseball, human frailty accounts for many times more.

Why make this connection? It's absolutely irrelevant whether incompetence is a great factor or not. The question is whether racial bias is a factor at all.

Otherwise, you've completely thrown out the spirit of fair competition by waving through racial bias as a way to decide competition simply because it's not as big a factor as incompetence.

It's like saying, "Why are you complaining about me slugging you and stealing your wallet? You hit your head harder on that low doorway, and you waste more money paying your creditors each month."

One of Gladwell's conclusions: "It measures our second level of attitude, our racial attitude on an unconscious level- the immediate, automatic associations that tumble out before we've even had time to think. We don't deliberately choose our unconscious attitudes."

most whites may have an automatic fear response to unfamiliar black faces, people differ in the amount of inhibitory control they can exert over this response. So some may be better able at consciously overriding their primitive impulses.

Noting the use of the word "automatic", I think these conclusions support each other. (While the study you cite goes further.)

And like you, I don't know in what specific ways these findings would influence referee and umpires' decisions.

It's not even just a disparity in treatment (that people are more likely to scientifically weigh studies that don't address issues they are trying to avoid), look above, there's a small, but steady stream of people trying to find little nitpicks with the study, and it turns out those nitpicks are covered in the methodology.

This is how the scientific method works. We are looking at the results and brainstorming alternate hypotheses to explain them. We could obviously make more educated guesses if we were to read the entire thing, but I would blame laziness and my refusal to install Acrobat on my home computer for the failure to do so, not something more sinister.

Am I missing someone, or are Kerwin Danley and Chuck Meriweather the only black umps in MLB?

Bucknor.

EDIT - CB, of course; the idea presented of throwing out his calls as "random" is RDF.

I don't know anything about tennis or this Hewitt fellow, but if a black player said that a white ref was racist, the black player wouldn't be "raked over the coals," because the media doing it would be afraid of being accused of being racist.

On my local public access channel the other day, there was a show where a black man was saying that all white people were liars, evil, and the spawn of the devil. This racist program runs on the television, and nobody does anything about it. A program where a white man said such things about black people? I really, really doubt that it would be run. I really doubt it.

A few things before I read the study. One, I'm not sure I would call the results of the study evidence of "racism," as opposed to simply "racial bias" (I guess the latter attracts more attention and sells more magazines). I doubt that if there IS in fact any bias, that it is conscious among most or all of the umpires. Not that racism has to be conscious prejudice or discrimination (depends on what definition you want to use), but that is generally the connotation I think.

On the other hand, if subconscious discrimination or prejudice to the degree that these reseachers purportedly found evidence of is "racism," then I suppose that virtually everyone is a "racist" in one context and to one degree or another. IOW, I think that almost everyone has subconscious racial biases regardless of how they view race on a conscious level.

Two, as far as point spreads and basketball game outcomes, there IS no reason for the median result of a game NOT to be the pointspread, by and large. What the other poster was talking about is when a team that is 15 points better than their opponent, the point spread will be around 9 or 10 points, because of garbage time at the end when one team has a large lead. But still the results of the game should fall around that point spread. I said "by and large" because the other poster was (somewhat) right with regard to his second explanation. The public does tend to prefer favorites so that if a team should be a -10 point fave, the linesmakers might make them a 10.5 or 11 point fave on the average. I said "somewhat" right because the linesmakers are only going to be biased toward the fave by a small amount (usually less than .5 point) otherwise smart and reasonably smart bettors will simply bet the dogs and change the line anyway. And linesmakers don't like to :open" with a bad line and then have it "corrected" (because the sportsbooks who pay the linesmakers don't like that).

Finally, one extra strike or ball per game is probably worth around .05 runs which is worth around .5% in win percentage, or around .8 extra wins per season. Not a whole lot. If they mean one extra strike and one fewer ball (for the pitchers who have an umpire advantage) then you can double that (the pitcher has an extra 1% chance of winning the game). I don't know whether to call that a substantial advantage or not. I suppose you can call that whatever you want.

And it is NOT the size of the effect that determines whether something is "significant" or not. It is the size of the effect as compared to the standard error which is usually principally a function of the size of the data sample. It also has very little to do with the variation in strike calling "talent" (the size of their strike zones) among umpires, although that does add some extra (not much) sample error into the equation. IOW, if I looked at a million pitches and each pitch were a ball or strike in around a 50/50 ratio, one standard error would be .05%, assuming that all umpires were the same. If umpires were not the same (which they are not), then that one standard error would be slightly, but not much, more. I think that is the case.

If I were of conspiracy mentality, I'd say headlines like this one are designed to strip the word 'racism' of any real meaning.

The effect is the same, regardless.

Can anybody make head or tails of figure 3 in the paper? The probit function seems to be all over the place depending on the count. What I was hoping was that they would include a table of their beta coefficient (the one that measures how important it is that the ethnicities match) vs the count.

They talk about how the ethnicity preference shifts with various changes in presumed scrutiny, but I would actually much rather see some proof that the effect is robust -- ie, that they get the same or similar coefficient measuring starts on odd numbered or even numbered days of the month, or something like that. I notice that whether the pitcher is home or away is not one of the variables they control for, although that seems like it would be a very large effect in terms of getting calls or not. They do some checking of whether the home team's winning percentage is affected by the ethnic matchup -- here they oddly find that the home team starting pitcher's ethnicity does matter, but the visiting team starting pitcher's doesn't (?).

I'm going to duplicate this study (using my own methodology). It is not that hard to do. If I don't get it done while this thread is alive, I'll post the results on The Book blog.

I have a database of players' race/ethnicity. I'd have to go through all the umpires and guess from recollection. Can people post the black and hispanic umpires here? The only ones that I know of that look black are Bucknor, Danley, Diaz (maybe he is Hispanic), and Meriwether. The only other one that I think is Hispanic is Hernandez. Any others? I wonder why the authors did not look at the Hirschbecks and the Jewish pitchers (Marquis and Levine?)?? ;)

That's a good idea, mgl.

How are umps against FAT pitchers? I always felt Bartolo Colon got unfairly squeezed.

Great, mgl, that should be interesting. I would think Jewish players are counted as white, since it's much harder to distinguish from appearence.

I just read the beginning of the study (the abstract I think). The found that 31.5% of called pitches (balls and called strikes) were strikes for non-matched umps and pitchers and 32.1% for matched pairs. That is a difference of .6%. For around 1 million called pitches, the standard error of the difference between one set and another where one set is around 700,000 pitches and the other is 300,000, is around .10%. So a difference of .6% is highly significant. That is assuming that there are not any sample biases in the various groups. For example, maybe in the matched pairs there happened to be better (more strikes) pitchers. These can easily be controlled for and apparently the authors do that. If we assume that this difference is the advantage that a typical white pitcher gets (since 87% of all umpires are white), we get an extra .3 strikes (and .3 fewer balls) per 100 pitch game (for a starter) with a white umpire and the same in reverse with a non-white umpire, or an average of .44 more strikes and less balls per start, or around .25% extra wins per start, or around .08 extra wins per year. The black pitcher loses around the same, .08 wins per season, and the Asian and Hispanic pitchers really get shafted and lose around .1 wins per season (32 starts). Of course their ERA should barely be affected (less than .02 runs) unless those extra .6 strikes are really "chosen" by the umpires when it matters the most. Anyway, I'll read on...

Reading some of the study, this sticks out. Black umpires called 1729 pitches from Asians, 522 were called strikes. Isn't that way too small of a sample to make any sort of conclusions? And who is to say a guy like Chuck Meriweather hates Chinese food and takes it out on Asian pitchers, painting the other 3 black umps in the study as closet racists as well?

In 1729 pitches, you have a standard error of about 1%, so any difference of less than around 2% is not that significant (technically you have to add the variances of the two samples to get the error variance of the difference between the two samples). Regression should give you around the same thing.

No studies like this tell you whether one observation or another causes the effect, only that there is an overall effect. If you then want to analyze and discuss possible causes, whether one observation (umpire in this case) is more or less likely than another to be causing the effect, that is another story. It does not change the results of the study which is trying to ascertain whether a particular effect exists int he aggregate.

One thing I like about this study is that the authors implicitly suggest that this is a Bayesian problem which it is. IOW, when we expect a particular result for one reason or another, if the results of sampling suggest that that result exists, then the significance of those results are stronger than of we did not expect those results. For example, if we reach into our pocket and flip a coin a hundered times and come up with 70 heads and 30 tails, why don't we conclude that our coin in biased? After all, we are at a very high significance level (p=99.8% or something like that). It is because we strongly expect that our coin is NOT biased even before we do our test (100 flips). It is really a Bayesian problem with an a priori probability. The probability that we have a biased coin before we flip it is very small. That influences the conclusion that we draw from the results of our test. Similarly, we should not be surprised if there are racial biases (subconscous I assume) in umpiring (at least I'm not). That also influences what we conclude from the results.

Anyway, I thought that this was a well-done (as far as I can tell - not being a statistician) and fascinating study. I am particularly intrigued (if not skeptical) by the reverse effects we see when the umpire is under scrutiny and not under scrutiny. If I am reading the results correctly, it appears that when under scrutiny (high attendacne, Questec park, or a "terminal" count), the umpires exhibit severe reverse racial bias and when not under scrutiny, they exhibit severe racial bias. That seems a little strange to me, but then again, I have no experience with these kinds of cognitive and sociological effects. Maybe that is typical. If they are, then this study only confirms what we would expect.

I was kidding about the Jewish umps and players of course. Then again, I think it is obvious to the umps who the Jewish players are and vice versa (a Jew can spot another Jew from a mile away - jewdar). I am Jewish of course (therefore I can make these kinds of jokes), at least by heritage.

And I actually thought about the Jewish thing, mgl. Though not for too long.

How about including Jewish umps/players (including hitters) in your study MGL? That would be interesting as well.

There's a couple of fundamental problems I have with this study:
1) They exclude hitters. It might have been too difficult to do, but why not use a) black pitcher vs white hitter with a black ump, b) white pitcher vs black hitter with black ump, and so on.

2) I don't like the assumption that umpire's biases are dominated by race. Initially, maybe I can see it. But these guys get to know the players fairly well. If a white pitcher is a jackass, is he really going to get an edge over a black or hispanic pitcher if the ump is white? I doubt it.

3) if the umpire was biased, it should show up a bit more than just one or two pitches a game. I'm not a stat guy and can't tell you if anything is significant or not, but that number doesn't pass my smell test.

This study is interesting and not at all surprising.

I'm not really a stat guy either, Shibal, but I'll give it a shot.

1) The hitter is not relevant to whether or not a called pitch is a ball or a strike. Also, since is in an inherent, reactionary bias, the umps simply isn't looking at the hitter when the call is made; he's staring right at the pitcher. Adding the hitter would be interesting, but since the demographics of position players is much more diverse than that of pitchers/umpires, the results would not be as clear.

2) Umpires can't love all the players, and can't hate them all either. Over the data of 1 million pitches, these personal feelings are subtracted, which is one of the reasons why these studies are done in aggregate. Again, they're looking for an unconscious bias; liking or disliking a player is certainly conscious, unless there's some sort of pheromone effect.

3) The effect is very small, and does only effect a small number of calls per game; this is why it only shows up in the aggregate data. If it affected more calls, I think it would be pretty obvious to fans, MLB, and the umps themselves that they were biased, and there would be some measure to correct it. As for significance, as mgl points out in one of his posts, the effect found is well over the margin of error for this sample size, making the results siginificant in a statistical sense. Again, this is a tiny, unconscious bias, and shouldn't be expected to be easily seen. The NBA study have 13 years of data, and this one has over 1 million pitches.

Then again, I think it is obvious to the umps who the Jewish players are and vice versa (a Jew can spot another Jew from a mile away - jewdar).

Speaking as a white Jew of Eastern European descent, I can tell a fellow Ashkenazi from two miles away. When it comes to the Sephardim or mixed race Jews, or Oriental Jews or any other possible Yiddim, that's a harder call. I usually err on the side of intelligence. If he has a Yiddishe keppulah, (I hope) he's a Jew. If he's a moron, (I hope) he's not a Jew. Alas, I have found a few MOT's are not geniuses.

"My feeling is that for every degree of human nature which is racially biased, human incompetence is ten times higher. In other words, while bias may account for some degree of missed calls in baseball, human frailty accounts for many times more.

Why make this connection? It's absolutely irrelevant whether incompetence is a great factor or not. The question is whether racial bias is a factor at all.

Eraser, my intention with that comment was to say with the implementation of certain technologies, we can get rid of the mistaken calls, taking out human bias and human frailty. I don't think most of the time that an umpire misses a strike call, he misses it because of his subconscious racism. I think most of the time he misses that call it's because he was not properly positioned behind the plate -- such as when a catcher will set up just outside of the strike zone and the umpire moves in behind him and fails to adjust for that.

Someone else (in a thread some weeks ago) said this on Primer: the home plate ump should announce the ball and strike calls, but those calls should be determined by a Questec-like technology (one that tells him if the pitch was a strike or not). As far as I know, that can be implemented now and with far less inaccuracy than with an umpire making the decision. I like this suggestion.

On all other calls, I think the umpires should make the calls, just as they do now. However, I think the managers should have the right to challenge an umpire's ruling on a non-ball/strike question. When a challenge is made, a replay umpire should review the film and overturn the call only if the decision is clearly wrong, as in the 1985 Don Denkinger call. If a call is not overturned, the challenging manager should lose his right to make another challenge that game. If the challenge is sustained, the manager should retain his right to challenge another call in the same game. (As long as the manager is correct and the field umpires decisions in that one game are overturned, a manager would have no limit on how many challenges he could file.)

My belief is this kind of a challenge system would speed up the game and greatly reduce the incentive of managers to throw Earl Weaver-like tantrums, which occur all to frequently when there is a questionable or bad call. In tennis, I have noticed the crybaby antics are way down since replay challenges have been introduced.

As far as I know, there is only one Jewish pitcher, Marquis, pitching in MLB right now. Jason Hirsh is listed in several sources as being Jewish. He went to a catholic high school and college I think. Schoenweiss is half Jewish I think, but I'm not sure if he is still pitching (DL?). I'm pretty sure that J. Hirschbeck (does Mark umpire any more? I don't think so.) and whichever other umpires are Jewish (I think none, actually) are aware of the Jewish pitchers. A few hitters, the most notable being Berkman, Youk, Green, and Ryan Braun, are Jewish (also Ausmus, half Jewish I think, and maybe one or two others).

mgl, the list below comes from this website:

Matt Ford - P, Brewers
John Grabow – P, Pirates
Shawn Green - OF, Dodgers
Gabe Kapler - OF, Red Sox
Al Levine - P, Tigers
Andrew Lorraine - P, Brewers (inactive)
Jason Marquis - P, Cardinals
Scott Schoeneweis - P, White Sox
Justin Wayne - P, Marlins
Kevin Youkilis – 3B, Red Sox
Rich Rifkin - F, Athletics

P.S. I just noticed how out of date this list is (Shawn Green "Dodgers"). I guess the proprietor of Jewishsports.com is not keeping up his website any longer.

I don't know anything about tennis or this Hewitt fellow, but if a black player said that a white ref was racist, the black player wouldn't be "raked over the coals," because the media doing it would be afraid of being accused of being racist.

What do you think the reaction would be if a gentile tennis player, upset with a call made by a Jewish line judge, suggested that the Jew be 'put in a crematorium and burned' for having made a ball call? When Vitas Gerulaitis actually said this in 1978, he was fined for 'gross and obscene behavior.' I don't have any recollection of a big deal being made of it by the public, B'Nai Brith, or the media. I think there was far more of an outcry when Hewitt made his (rather stupid) comment.

To answer your question, Vaux, 'if a black player said that a white ref was racist,' such a comment would quickly be ignored, because such cries of racism are all too common and very often not justified. Crying wolf only carries so much weight for so long. Now it is mostly weightless.

With that IAT test (if anyone is interested FYI) I'm left handed tending towards ambidextrous and it came back with a moderate preference for African Americans, i.e. the opposite to what would be expected if the test was designed with the non cack-handed in mind...