Racial bias in NBA foul calls?

[Arthur Dent]

Maybe they did this, but I'd be interested to know whether they took position into account. 1) If there's a big difference in fouls committed by guards vs. post players, and 2) if there's a disparity in the racial makeup of the positions.

They did. Because of the enormous sample size, the authors were able to take into account a large number of variables including position, minutes played, home vs. away, year of game, etc.

[Radio_Bugtussle]

Maybe they did this, but I'd be interested to know whether they took position into account. 1) If there's a big difference in fouls committed by guards vs. post players, and 2) if there's a disparity in the racial makeup of the positions.

I can't believe I'm getting into this, but bad baseball does strange things to me.

The proper way to study this is with a complete review of video of a very large and representative sample of games showing what the official who made each call could see. I also want to see the video of what the other two officals could see.

The study should look into what was called, what wasn't called, did players with reputations for fouling get special attention, did the defense make a more concerted effort to foul poor FT shooters or to avoid fouling guys who are good at the FT line, etc. That's just a start.

[haltz]

Maybe they did this, but I'd be interested to know whether they took position into account. 1) If there's a big difference in fouls committed by guards vs. post players, and 2) if there's a disparity in the racial makeup of the positions.

They did. Because of the enormous sample size, the authors were able to take into account a large number of variables including position, minutes played, home vs. away, year of game, etc.

How so? Let's just say my knee-jerk reaction was actually something. How would the sample-size help a construction problem? Sincere question, by the way.

I can't believe I'm getting into this, but bad baseball does strange things to me.

The proper way to study this is with a complete review of video of a very large and representative sample of games showing what the official who made each call could see. I also want to see the video of what the other two officals could see.

The study should look into what was called, what wasn't called, did players with reputations for fouling get special attention, did the defense make a more concerted effort to foul poor FT shooters or to avoid fouling guys who are good at the FT line, etc. That's just a start.

Yeah, I haven't touched a basketball since I ripped off my warm-ups for the last time my senior year as the ninth man. Funny what bad baseball gets you interested in.

I like your last paragraph, a lot of other variables that wouldn't have crossed my mind. Let's say that white players are better FT shooters on average (no idea if this is true), then that's another variable to account for. I read the study a couple of weeks ago, and if I remember correctly, it was like another foul every couple of games or so. That seems significant, but could be accounted for pretty easily if just one variable isn't accounted for.

Does that make sense to anyone else or am I just rambling here?

[Radio_Bugtussle]

On fouling in the NBA... If I'm guarding Shaq (or Manute Bol back in the day) and he gets the ball in the paint, I'm gonna hammer is brick-laying arse and put him at the line for a pair of FT's.

If I'm guarding some guy whos' 88% at the line, I don't mind a properly guarded* jumper. I'll let him have that.

*I realize that proper defense is a lot to ask in the NBA these days.

[Arthur Dent]

Maybe they did this, but I'd be interested to know whether they took position into account. 1) If there's a big difference in fouls committed by guards vs. post players, and 2) if there's a disparity in the racial makeup of the positions.

They did. Because of the enormous sample size, the authors were able to take into account a large number of variables including position, minutes played, home vs. away, year of game, etc.

How so?

The magic of multivariate regression, a standard statistical technique where you look at how each independent variable correlates with the foul call rate. It's essentially a way to control for intervening variables when you don't have the luxury of doing a controlled experiment. The larger your sample, the more variables you can account for without reducing the uncertainty in your result.

How would the sample-size help a construction problem?

It wouldn't, but where's the construction problem?

Again, it seems like people are extremely anxious to dismiss this study because they don't like the result. The bias found was small but significant. It certainly does not show that any of the referees are racists, so the defensive reaction is not necessary. If I were the NBA commissioner, I would have responded by saying that the results are not surprising since referees are human beings, and that the league will respond by ensuring that they have good, diverse staffing of officials.

[haltz]

The magic of multivariate regression, a standard statistical technique where you look at how each independent variable correlates with the foul call rate. It's essentially a way to control for intervening variables when you don't have the luxury of doing a controlled experiment. The larger your sample, the more variables you can account for without reducing the uncertainty in your result.

How would the sample-size help a construction problem?

It wouldn't, but where's the construction problem?

I think your response A, answers the question you pose in response B. Since my statistics knowledge is limited to cramming for a few college exams, that's exactly what I was after. If I didn't think you knew more about this stuff than I do, I wouldn't have asked. And like I said, it was an honest question.

Again, it seems like people are extremely anxious to dismiss this study because they don't like the result. The bias found was small but significant. It certainly does not show that any of the referees are racists, so the defensive reaction is not necessary. If I were the NBA commissioner, I would have responded by saying that the results are not surprising since referees are human beings, and that the league will respond by ensuring that they have good, diverse staffing of officials.

I don't think so. People are generally skeptical of studies or whatever that produce uncomfortable results, and for the record, I don't think the study would be necessarily describing racism either way. What I am interested in is finding out what all studies of this type are actually describing, which is what I was getting at in my previous questions. It generally takes a preconceived notion to even decide to study something, coupled with my generally skeptical nature, and voila, you get a couple of questions.

[Arthur Dent]

If I didn't think you knew more about this stuff than I do, I wouldn't have asked. And like I said, it was an honest question.

I'm no statistics expert. I took the class in college and have read some things that use statistical techniques. That's about it.

I'm obviously coming off as snippy which I don't mean to be.

Again, it seems like people are extremely anxious to dismiss this study because they don't like the result. The bias found was small but significant. It certainly does not show that any of the referees are racists, so the defensive reaction is not necessary. If I were the NBA commissioner, I would have responded by saying that the results are not surprising since referees are human beings, and that the league will respond by ensuring that they have good, diverse staffing of officials.

People are generally skeptical of studies or whatever that produce uncomfortable results, and for the record, I don't think the study would be necessarily describing racism either way. What I am interested in is finding out what all studies of this type are actually describing, which is what I was getting at in my previous questions. It generally takes a preconceived notion to even decide to study something, coupled with my generally skeptical nature, and voila, you get a couple of questions.

Skepticism is great, but with certain topics like race, people use it as an excuse to dismiss the truth. I heard the same response to the ESPN study about Bonds and to that implicit association test I posted a while ago. People know (or want to believe) that race is not a factor, so obviously any measurement that shows otherwise must be due to some kind of error.

[haltz]

Skepticism is great, but with certain topics like race, people use it as an excuse to dismiss the truth. I heard the same response to the ESPN study about Bonds and to that implicit association test I posted a while ago. People know (or want to believe) that race is not a factor, so obviously any measurement that shows otherwise must be due to some kind of error.

Once again, you're painting with a broad brush here. I think there are a ton of decent questions to be asked about this study, because I'd imagine that it's tough to quantify. Also, what are the results actually telling you? (Not to get back into it, just reiterating my original knee-jerk.)

I should have put a disclaimer on the entire thing actually because I agree with you in principle, and figured that questioning the test would perhaps lead to this conclusion.

As for the Bonds stuff, I am pretty sure that race does have something to do with it. If I thought it was worth it, or felt like having the argument at the time, I would have chimed right in. I truly do not understand the seething hatred of Barry Bonds.

[greenback44]

*I realize that proper defense is a lot to ask in the NBA these days. :oops:

I have real trouble watching college ball after watching a few NBA games. The defense and officiating are so much better that they aren't playing the same sport really.

[Arthur Dent]

Once again, you're painting with a broad brush here. I think there are a ton of decent questions to be asked about this study, because I'd imagine that it's tough to quantify.

Certainly. I guess I see the response (not yours specifically) as being about trying to find a reason why the study can be immediately dismissed without so much as reading it.