"Study of N.B.A. Sees Racial Bias in Calling Fouls"

How can you stand by a study that weighs a white ref calling that foul with the same weight attributed to a white ref calling a charge on Dwyane Wade when he barrels into Andres Nocioni?

 

Which is my problem with how the framers have presented the study. If they want to market this thing as a bias study about the subconcious, fine. Don't make it sound like this bias is affecting the bottom line. But that's exactly how they've sold out.

 

The star/rookie bias has a lot more impact to the game than black/white.

I agree that the study shows racial bias in decision making. And it needs a basketball study to prove it?

 

well since i had some time to kill i decided to try to calculate the effect. the baseline is an all black crew (no extra fouls called on black players) with the .16 extra fouls called going to the all white crew, and then linearly changing along the way (.053 extra with one white ref, .107 extra with 2 white refs). used 68% white refs to determine average ref crew racial composition.

used .9 ft's made per PF called (from this last season) to determine the point effect of fouls (although that would be too simplistic). then used 2.7 extra wins per year per extra point of point differential, which seems to work for most years.



so that bolded "0.27" at the bottom, means that every extra black player game you average would subtract 0.27 wins from your season total. and by black player game, i mean in a typical game you have 5 players on the court so 5 player games per game. if black players get 80% of your minutes, you average .8*5=4 black player games. so if the league average is, say, 70% of the court time for blacks, that would be 3.5 player games on average. so even if you had an all-black team, you would only be 1.5 black player games above average and would therefore only lose about .4 wins over the season versus an average team.

even if it was an all-white league and you had an all-black team, you'd only lose 5*.27=1.35 wins over the course of the season.

and that is a) the most extreme example you could have, and b) assuming all of the .16 bias is on the white ref side of things. in other words, the effect is tiny, but i have a lot of free time, so i can spend time figuring out that things don't matter.

 

Doesn't matter if you think it's a waste of time, it isn't. Thanks for the help and thanks for your work.

I'd say the same thing to Berri, et al if they weren't trying to package their findings the wrong way.

 

That was some good work Francis, I'm not sure how accurate the linear extrapolation is so the effect may actually be smaller than you have it. Then again I don't think fouls equate to wins the way you have it set up either.

As for the argument that none of this matters, think about a close seven game series, with a bunch of games going to the wire, we'll take Dallas/GS because that's the first thing that pops into my head. Would a single biased foul call matter in that series? The cynical answer is no, because soooo many things bias the refs in a playoff series, home court, players whining, coaches complaining, etc. The logical answer is yes, a seemingly innocent foul in the middle of an ultra tight series could swing it either way, it could cause guys to lose their cool and get ejected, it could bring on foul trouble and stars sitting, leading to a larger impact.

Personally I don't think race is an issue with the NBA, I do think refs are a problem with this league, race is just not part of the problem.

 

clearly when the Rockets play games at the Utah arena, majority of the calls go their way... stupid refs

 

Here's a quote from their paper describing how this could impact a team's chances of winning in a single game:

So, in a situation where one teams is "15% less black" than another team, but otherwise they are equal, they estimate that their chances of winning go up about 3.4% with an all-white officiating crew versus an all-black crew. And for just changing the race of one of the refs, it would change their chances of winning by about 1%. Increasing your chances of winning a given game by around 1% might earn you an extra win over the course of a season. Whether or not that's significant depends on your perspective, I guess. Sometimes, 1 game determines who gets home court advantage in the playoffs.

NBA teams spend so much time and energy in various areas trying to get any kind of edge they can over the opponent -- be it use of advanced statistics, special physical training, motivational gimicks, etc. Hell, on top of all that stuff, maybe it actually pays to play the white guy on the bench more minutes than the black guy, all else being equal. Even if it gives you just a fraction of a percent better odds of winning a game, maybe that's enough to make it the tie-breaker in deciding which player to put in the game.

Anyways, I just think it's funny how all the sports experts and radio hosts are referring to this as a big waste of time. I mean, considering all the crap they usually spend hours during the day ranting and raving about, and they want to call a serious academic study pointless and stupid?

 

i don't feel like reading. what's .226 in the .15*.226 formula?


by their way, 3.4%*82 games = 2.788 wins for 15% more uh, blackness in a team. since one player game would be 20%, that comes out to 3.7 extra wins for every extra black player game. the way i did it, 82 games of the whole white crew vs whole black crew differential would be .39 wins (as opposed to the .264 i said earlier with random crew makeup). so they get an almost factor of 10 greater magnitude than my, admittedly, simplified method.

maybe i missed a decimal, but i don't see how .16 fouls per 48 minutes can account for almost 2.8 wins in a season with just 15% more black minutes.

like Hollinger said, that's 10 extra fouls a season. 10 fouls = 2.8 wins? i don't see it.

 

It's a coefficient which determines how much more likely a team with more black players will win when reffed by an all-black crew versus an all-white crew. You multiply it by %-difference in black players to get improvement in their chances.

Hmm, I checked your calculations, and they make sense to me. It seems like the difference in how it impacts the win total is a factor of ten off. Strange.

 

1. It seems like a lot of people are interpreting this as some kind of personal attack on the NBA. I can understand the NBA doing it, since a public misunderstanding could hurt their image and profitability. But, a lot of journalists also seem to be prejudicially hostile to the research, which surprises me. I don't see the research as an attack on the NBA at all. It is a study that reveals something about human nature. The NBA merely provided the researchers with a well-controlled environment and a built-in data-keeping system. That they found a small measure of ref bias that affects game outcomes doesn't make the NBA a seething cauldron of racism -- they probably compare pretty favorably to similar subconscious decision-makers on the street, in business, and elsewhere. Don't be threatened by it.

2. The attempts to belittle the methodology of the study have been weak-sauce. Half the complaints I've heard of so far I know have already been considered by the researchers and dealt with to whatever extent. Another fraction are spurious complaints from people not understanding statistics too well (which is understandable). The rest were probably considered by the researchers and I just hadn't read about it yet. Guess what? These guys do this stuff for a living. They are better at it than we are. The research will undergo peer review. If someone is going to tear it apart (which is possible), it'll be other academics, not the NBA and not journalists. I hope that the research makes as much buzz when it gets the thumbs up/down from the academic community so the general public will know whether it was valid or not, ultimately.

 

Ok, I digged a little further. Perhaps this will clarify the discrepency:

They find that having an extra black player impacts not only the fouls called, but also the points scored and turnovers committed. So, you need to take into account all three to understand how it impacts the final point differential. Also, they say you need to also consider the impact on the other team (they effectively have one extra white player, relative to their opponent).

 

Amen brother! I'm so sick of the black white paradigm.

All these supposed "highly educated" people who's reality only consists of two colors.

 

Hmm... interesting study. I never thought about this aspect before, mainly due to majority of the refs being white and players being black. So any racial bias won't actually matter much upon first inspection.

But I'm both surprised and appalled at a lot of the responses though. Sure, most of us aren't statistians, but some of the arguments against the study just defies common sense. It's kind of weird on an NBA forum, where we pretty much use stats all the time for arguments.

 

Because it's irrelevent to the study. Over 14 years, the white refs and black refs are going to have a similar set of fouls to potentially call.

 

We need a referee in every minority - Hispanic, Asian & Purple!

 

The purple ref would have helped Rodman back in the day.

 

Maybe the Dick Bavetta Tech on Juwan could be traced to this?

 

Thought I would share this.. its a John Holliger blog on this topic.. I didn't see it in the thread, so if it is.. I apologize for the repeat!

link