Study sees racial bias in calling fouls

[kidcolin]

[ QUOTE ]
about half way through it now, thought i'd answer this one:

[ QUOTE ]
A technical foul study would be WAAAAY more interesting than this, I don't know why they didn't do that. I wouldn't be surprised if the correlation was 3-5 times stronger than for regular fouls.

[/ QUOTE ]

they do look at flagrants and technicals. The same pattern emerges, but because of their rarity and smaller sample size, they aren't as statistically valid. For flagarants the effect IS statistically signifigant, but it isn't exactly a home run.

[/ QUOTE ]

plus, Sheed is black.

[UATrewqaz]

Was position taken into account? Don't centers commit more fouls than other positions typically? And aren't centers taller than other players, typically? If a high number of centers are black that would explain part of the discrepency wouldn't it?

(the above is just an example of how outside factors can contribute, I have no idea if it's true)

The problem with stats is people want to start drawing wild conclusions without sufficient data, especially when it comes to causality.

Without studies like this Jesse Jackson and Al Sharpton would have to get real jobs and focus on real problems facing the black community (and the NBA dress code is not one of them btw)

[MicroBob]

needle - thank god. i was confused by your initial agreement with him.
Smith's arguments were about as bad as an announcer talking about how great it is that some 0-for-20 player is up there in this spot because "He's due so he's more likely to get a hit than some guy who's been hitting .450 lately."

[Needle77]

[ QUOTE ]
needle - thank god. i was confused by your initial agreement with him.
Smith's arguments were about as bad as an announcer talking about how great it is that some 0-for-20 player is up there in this spot because "He's due so he's more likely to get a hit than some guy who's been hitting .450 lately."

[/ QUOTE ]

I want to walk to the booth that the announcer is sitting in and just pound him into a pulp everytime I hear something that even comes close to this.

And I'm learning that reading uNL, posting in SE, watching ESPN, and 4-tabling just doesn't work for me. I'm considering stopping snide comments, it'll take me some time.

[Rodney_King]

Note that this study was actually not peer reviewed yet.
Great reporting by the NYT, clearly it was "fit to print".

I'd like to see the full text version of the NBA's study; mostly because after school I will be working for the Segal Company, the consultants who administered the study. IF they had found bias in their study, do you think the NBA would no longer use them as benefit consultants? I assume they were involved because they consult on the referee benefit plans. Would they NBA go to a different firm if Segal found the same results as this survey? Personally, I'd doubt it, as it would make the NBA look good by taking steps to fix the problem before it hit the media.

[teamdonkey]

[ QUOTE ]
plus, Sheed is black.

[/ QUOTE ]

i know you're kidding, but it's not whether black players get more technicals, rather if white refs are more apt to call them than black refs.

[ QUOTE ]
Was position taken into account? Don't centers commit more fouls than other positions typically? And aren't centers taller than other players, typically? If a high number of centers are black that would explain part of the discrepency wouldn't it?

[/ QUOTE ]

they have a monsterous data set (267,000 games, and however many fouls that includes) so they were able to factor in a bunch of things like that. Things like position, if they're a rookie or all-star, if they only play a few minutes a game and possibly are in there to foul hack-a-shaq style, if their team is in playoff contention, if they're at home or away, etc. The end result with or without these factors was pretty much the same.

Just finished it, and i'm really not the best person to critique it, but some more thoughts:

-these guys do a good job of crunching numbers but a poor job of drawing conclusions. It's obvious they don't follow basketball and have a difficult time applying their results to what happens in a game.

-I don't have a statistical background so i can't talk with confidence about their methods, but they seem to come up with the same basic answer with several different approaches. In science this usually means their premise (there is some bias by opposite-race referees) has some merit.

-IMO, they find 2 big things: (1) white refs call fewer fouls on white players than black refs do. (2) whiter teams win more often with white refs than they do with black refs. Because referee crews are mostly white, this benefits teams that play more white players.

the problem is, they can't say (1) causes (2). The actual amount of fewer fouls called by white referees is tiny and shouldn't by itself be a big deal. However the actual change in winning % for whiter teams is significant. They spend a decent chunk of the paper trying to explain how (1) causes (2), and fail IMO. I don't have an answer for this either.

an interesting thing to note: this difference isn't what you'd immediately think, it isn't evil white refs calling poor black victems for more fouls... all refs call fouls on black players at about the same rate. It's the rate of fouls on white players that changes. There's also no real way to tell which ref group is discriminating: white refs might be calling fewer fouls than they should, black refs might be calling more, or somewhere in between.

[kidcolin]

I haven't read the report yet. Kiki Vanderwhatever (I HATE him as an analyst) just said that they did the study based on the box scores. Is that true? How do they know which ref made which call? Did they just do some mathematical assumptions?

Barkley called the study BS, but unfortunately used the "more black players" argument. Still, when Barkley is defending white people, take heed.

[teamdonkey]

[ QUOTE ]
I haven't read the report yet. Kiki Vanderwhatever (I HATE him as an analyst) just said that they did the study based on the box scores. Is that true? How do they know which ref made which call? Did they just do some mathematical assumptions?

[/ QUOTE ]

they don't know which refs made which calls (only the NBA has that data). For most of the stuff they looked at the crews as a whole - 0% white, 33% white, 67% white, 100% white.

edit - they also used some fancy statistical stuff i didn't understand to isolate specific refs and came up with the same pattern. One thing they said was no ref (black or white) stood out as particularly biased.

[danvh]

Come on, how many fouls can you draw standing around the 3 point line draining 3s??

[NT!]

[ QUOTE ]
Even if this "study" was performed in a way that made sense, why is it not possible that black players just commit fouls at a higher rate?

[/ QUOTE ]

this is accounted for in the study, because one of the things the regression would look at is the foul rates for the same players in games officiated by different crews.

i haven't had a chance to look at this study yet, but keep in mind that the 4.5% number is between the two greatest extremes (all-white vs. all-black). in terms of statistical significance, if you take a guy who commits an average of 3 fouls per game, it would mean an extra foul per 9 games or so. the rate difference is presumably smaller between racially mixed crews.

in reality the impact of racial bias in foul calls would be more evident among big men. if white big men accounted for a disproportionate percentage of all white players' time on court, the impact of the extra foul calls against big men would also be magnified. since big men commit the majority of fouls, and since fouls against big men are of greater defensive significance, the impact could be larger than it initially seems.

in fact, even if whites received a proportionate amount of playing time at the key defensive positions, my guess is that the greater impact of foul calls against big men would magnify the importance of the extra fouls.

i would guess that the reason the study projects a difference of a couple wins per season to this bias is that the impact of being in foul trouble spreads to the entire defensive outlook for the team, and also affects offensive output if you can't have your best big men out there.

this would be much, much more interesting if we had access to which officials called each foul. other interesting data would include whether the official calling the foul was the one to whom the call 'belonged' (the closest / best positioned official), and the service time / ranking of the official making the call.

in general, unconscious or subconscious discrimination along racial and gender lines has been documented in other areas, so it wouldn't be a shock in the NBA, though of course they wouldn't admit it since it would be bad for business.

i bet it would be much easier to test this in baseball, where there is so much more data.