I have a hard time agreeing that black pitchers are "disadvantaged." Simply put, there aren't that many black pitchers in the major leagues. I can only think of a few off the top of my head:

Sabathia Willis Ray King Tom Gordon Darren Oliver

There might be like like 10 more, but even still (I wish I could actually find a number on how many black pitchers are in the Majors right now but I couldn't), that's a tiny sample size. One pitcher's poor control (Willis maybe) would have a HUGE effect...

That gives us way too small of a sample size to see if umpires are biased. I'd chalk the black/white difference up to a few black pitchers who don't have great control, especially once you notice that white pitchers got more called stirkes from all races of umpire. In addition, there are only a few non-white umps in the majors, which means that one umpire's tendencies could make a big difference in the study. I'd like to see it broken down ump-by-ump....

I just feel taht the stats taken on minority have way too small a sample size in terms of both players and umpires, which hurts the study in general. It's not unimaginable that umps are slightly biased in terms of race (by accident), but this study doesn't tell me much....

The sample size is much larger than that. It's 3 or 4 full seasons of pitch by pitch data, meaning it has more than 2 million observations. There are 27 black pitchers in the sample and they represent over 50,000 pitches thrown. The paper uses conventional tests of statistical significance to show that despite your concerns about sample size, the results are statistically significant. And the point people keep raising about white pitchers having better control is irrelevant, because it is controlled for through the statistical methods. If it is true that black pitchers have worse overall control, the fixed effects control for each pitcher's individual tendencies to "wildness".

"I don't want to play golf. When I hit a ball, I want someone else to chase it."

by buffalobillsrul2002 » Sat Aug 18, 2007 11:52 pm

A. Where does the study ever control for a pitcher's "wildness"? I read it and I didn't see it (maybe I missed it, I'm just asking you to point it out to me; maybe even quote that part in your post so I can see it )

B. It is 3 years of data, but still, there are a small number of pitchers who have a large effect on the study for both the Blacks and Asians (notably SP that pitched all of those 3-4 years).

C. So umpires are racist against their own race? Seems kind of odd to me....

buffalobillsrul2002 wrote:A. Where does the study ever control for a pitcher's "wildness"? I read it and I didn't see it (maybe I missed it, I'm just asking you to point it out to me; maybe even quote that part in your post so I can see it )

B. It is 3 years of data, but still, there are a small number of pitchers who have a large effect on the study for both the Blacks and Asians (notably SP that pitched all of those 3-4 years).

C. So umpires are racist against their own race? Seems kind of odd to me....

A. There are three different ways they controlled for this in the study. First, they control for what the count was when the pitch was thrown. Wild pitchers will be often be in counts where they are behind, so this variable controls, to some degree, for wildness. Second, they report in footnote 10 that they also estimated models with walks per inning as a control and found it did not change the results. Third, they use a statistical method known as "fixed effects". This method essentially uses each player as their own control; a fixed effect is a statistical technique that controls for unmeasured features of an observation. In other words, things that are inherently part of each pitcher's make-up (wildness, toughness, etc.) are controlled through this technique.

B. Even a pitcher that averaged 100 pitches per game and pitched 35 games in all three years represents less than 20 percent of the black sample...and the model still achieves significance, despite your concerns.

C. No, umpires are biased against the opposite race. That bias is moderated and possibly even eliminated or reversed when they are monitored by Questec or are in situations where they may be more likely to consciously self-monitor their behavior. Since there are far more situations where an umpire's subconscious biases appear to be operating, however, the net result is that pitchers who are of the opposite race of the umpire get fewer strike calls, controlling for other factors. Over the course of a season, a team that tried to match its pitcher to the race of the home plate umpire would gain a 2-4 win advantage (theoretically; there would be huge practical problems in ding this). Also, since there are far more white umpires than black umpires, minority pitchers end up with worse results, and effectively earn lower salaries.

"I don't want to play golf. When I hit a ball, I want someone else to chase it."

Alright, I really had a hard time understanding the count table as I couldn't figure out what probit and LPM meant (I know very little about statistics, and the study didn't really explain it in a great way). The same thing goes for the "fixed effects" part, so I'll take your word for it on that.

I still won't bend on the point about the minority pitchers. 27 pitchers really isn't very many, especially once we consider taht some of the pitchers probably only threw a few innings each before being sent back down to the minors. Even if say, 3 or 4 pitchers make up half of the study (which I find entirely possible if there are 3 SP who pitched all 3 of those years. Willis + Sabathia are 2, although they may be the only 2...) Also, there are only a few black + hispanic umpires each, which means that the minority umpire-related statistics are greatly influenced by each umpire...

You're right on point C after further review of the study....

The relevant sample size is not the 27 pitchers. It's the 50,000+ pitches. And the statistical methods control for the fact that these pitches are highly correlated because they came from just 27 pitchers, violating the normal assumptions of independent observations. And, even with those corrections, the results are statistically significant.

"I don't want to play golf. When I hit a ball, I want someone else to chase it."

you can use data to come to whatever conclusion you are looking for.

i like how the response to the REVERSE racism(opposite of what this study is trying to prove) in terminal counts (graph posted in this thread) is that umpires are more scrutinized in that situation and their subconscious kicks in is that a scientific fact or principal i'm unaware of?

if the numbers had just kept showing a tiny percentage point of "racism" would the study ever mention that explanation above? would the researchers say "this is surprising because we expected the results to be reversed because of the scrutiny and subconscious"? no they wouldn't even think about it because the data would already be showing what they wanted and there would be no need for excuse explanations.

GotowarMissAgnes wrote: And, even with those corrections, the results are statistically significant.

to people who want to believe

Exactly. I'm pretty sure it would be possible to write a paper with the exact OPPOSITE conclusion of this one and find plenty of statistically significant phenomena to support the claim.

What nonsense. This same type of human behavior is observed in hundreds of other studies. In fact, it's a well known result called the "Hawthorne effect" in social research. When people are observed and know they are being observed, they often alter their behavior to conform to expectations. The results found are exactly what any good social scientist in this field would have expected based on decades of prior research.

If you really think you can write a paper with evidence supporting the other side, be my guest. I'm always willing to be convinced by good evidence. Since you can't even correctly understand the paper at hand, however, I'm eager for the hilarity that would result.

"I don't want to play golf. When I hit a ball, I want someone else to chase it."