Penn/Cornell NBA study: Blacks called for more fouls

[Jim Hyla]

[quote BCrespi] However, ignorance like this, in which analysts refuse to even read what they are commenting on, much less, understand that a couple percentage points above chance might be statistically significant drives me moderately insane.  I guess I need to pick my spots better with which shows I listen to.[/quote]

Why worry about these guys? Our president seems to get along quite well doing the same thing.::cry::

[Ken70]

P. 29: "Table 3 is instructive, showing that the rate at which fouls are earned by black players is largely invariant to the racial composition of the refereeing crew."

P. 4: "Moreover,in light of the mismatch between the composition of the players (around four-fifths of whom are black) and their evaluators (around two-thirds of referees are white in our sample), an own-race preference may drive an aggregate bias against blacks (or for
whites)."

In plainer words:

> The data revealed no difference in the rate of fouls called on black players between an all black crew or all white crew or any combination in between.

> The only difference was found in rate of fouls called on white players, black crews called at higher rate than white crews

> Point above is the only evidence authors have for their hypothesis of racial disparities in foul calling.  They generalize the white player discrimination data to broadly assert race makes a difference on part of all refs against all players (despite their own data that shows it makes no difference in the case of black players).

> Since they have now shown a race bias to exist, and since over 80% of NBA players are black, it is therefore shown that blacks are disproportionately discriminated against in the NBA.

Can you imagine these ideologues teaching your children?  Happens every day in PC academe.  Shame on Cornell for employing this guy. We can only hope tax dollars didn't pay for this screed.

[oceanst41]

Danny Ainge was being interviewed on WEEI in Boston now and just said the study isn't false and it holds some weight. This coming from the Celtics front office statistics he said.

He added that the only 100% sure statement he could make was that Bird and Jordan got all the calls. ;-)

[Tom Lento]

[quote Ken70]P. 29: "Table 3 is instructive, showing that the rate at which fouls are earned by black players is largely invariant to the racial composition of the refereeing crew."

P. 4: "Moreover,in light of the mismatch between the composition of the players (around four-fifths of whom are black) and their evaluators (around two-thirds of referees are white in our sample), an own-race preference may drive an aggregate bias against blacks (or for
whites)."
[/quote]

I should point out several things:

1) Table 3 does not include controls, and as such it is nearly meaningless, since after including controls there does appear to be a statistically significant relationship between racial composition of officiating crews and fouls called against black players. There certainly appears to be a relationship between race of referee and fouls called against black players, and it's clearly illustrated in Figure 2 (includes controls, methodology may be questionable but there are proper ways to do what they claim they did so I'll assume they did things right).

2) By the same logic, I'm not sure how they can make any claims about whether black officials call disproportionately more fouls on white players, since from what I can see this is not checked as a dependent variable in the multivariate regression analyses. Figure 1 appears to be their big hit there, but it's not clear if they included controls in the stats used to generate those charts, and I don't see the regression tables which show a full-scale analysis of ratio of white referees in the officiating crew as an independent variable *except* where it's interacting with the race of the players. Maybe I missed it, or maybe it was buried in the full-scale table of coefficients, which if you ask me is a hideous way of presenting your results, or maybe they coded their interaction to somehow show the relationship for both whites and blacks, although that's not really standard practice, at least not in my field.

3) My gut response is that I don't like their approach, although I'd have to do way more thinking about it to really tell you why I don't like their methodology or how I would do things differently. It may be perfectly fine, in which case I simply dislike the way they presented the data (unless I'm missing something, which is also entirely possible). At this point I'm not sure I trust their main results, and based on their regression results and other output figures, whether or not this paper will be accepted for publication is largely a toss-up in my mind. Assuming a more rigorous look shows that their methods and results are sound, I'm sure they'll get a revise and resubmit and they'll be told to do a better job of clearly illustrating their results with the appropriate tables and figures. Note I didn't check their theoretical framework, which might be enough all on its own to get accepted/rejected at some journals (I am not an economist, so your mileage may vary).

Can you imagine these ideologues teaching your children?  Happens every day in PC academe.  Shame on Cornell for employing this guy. We can only hope tax dollars didn't pay for this screed.

Holy leaping logic, Batman! One academic submits something for publication that *may* be a bad piece of research which apparently does not match your own personal world view, and he's an ideologue who's fully representative of everyone in academia. And this in the same post where you complain about them mis-representing statistical results. . . wow.

Note two things:

1) This article has not yet been published, so it might not pass the peer review for the journal, and even if it does it may be a publication mill anyway, in which case nobody in the field will pay much attention to it. Furthermore, the paper may be perfectly sound - I took a much better look at their results than you did, apparently, although it was admittedly a quick glance at the tables and figures, and I still can't say whether or not their findings are any good.

2) According to the NYT article, the Cornell researcher is actually a graduate student, and as such was not exactly "hired" by Cornell (edit - grad students are occasionally considered employees of the university, but it's not like we're really employed here). The only professor involved is at Wharton.

[ugarte]

[quote Tom Lento] (edit - grad students are occasionally considered employees of the university, but it's not like we're really employed here)[/quote]
How to determine if grad students are employees of a university depends on whether they are trying to unionize or attempting to get around worker's compensation laws.

[Jeff Hopkins '82]

I want to know how never calling fouls on Shaq impacted these studies.