The Bosque Boys

Today's Thought: This explains why the all-white teams keep winning the NBA championship.

Yet another group with time on their hands, culling statistics from the air, most likely to prove a preconceived notion. It would stand to reason that fouls would be called more frequently against black players simply because they are the majority in the NBA.

Want to bet there are more Hispanic players with calls against them in Major League Baseball than in other sports?

I think way too many penalties are called against our white brothers in hockey. And don't get me started on the fines against short people in horse racing.

According to the story the analysis did account for differneces in time on the court

To investigate whether such bias has existed in sports, Mr. Wolfers and Mr. Price examined data from publicly available box scores. They accounted for factors like the players’ positions, playing time and All-Star status; each group’s time on the court (black players played 83 percent of minutes, while 68 percent of officials were white); calls at home games and on the road; and other relevant data.

Dang, pg, now I'll be forced to take it seriously. I wonder if the study differentiated between regular personal fouls and taunting/unsportsmanlike fouls. I ask because the trashtalkinginyourface style of basketball came from the urban bb courts.

Photognome you rained on our parade a bit here. But thanks for playing it straight.

On the other hand, even if the number-crunchers took all this into account, racial bias against African Americans in the NBA remains a far-fetched pursuit.

And, perhaps more importantly, it is revealing that the NYT did not address those questions in a more direct and upfront way.

Farmer, I think you give the NYT too much credit. I doubt if they have a reporter who can read scholarly literature and analyze statistical analysis. When reporting science, and other areas, most reporters do not have much depth of knowledge and skill.

Gardener & Farmer

I have given a quick read to the Price & Wolfers article as well as re-reading the NYT article.

I do not have access to the press release so can't speculate on how the story was 'spun' before the Times got ahold of it.

The Times did call in academic experts for help but I do not see evidence of a good critique from a statistician. The research article itself shows evidence of a major strength and a major weakness.

Strength of statistical modeling (what to expect they are economists - right)

Weakness - no insight into the pre-requisites or weakness of 'statistical hypothesis testing'. All indications are that this was an exploratory study not a hypothesis testing study. As such the appropriate mind set is exploratory data analysis not hypothesis testing. If you confuse the two you get a 'fishing expedition.
What does this mean:
While the authors do not make a big deal of it they do use the concept of 'statistical significance' in narrative and more-so in tables. For a refresher this asks the question of "how likely are these results to have been due to chance rather than to a real relationship. Usual cut-offs are 5% or 1%. So if you have 20 findings significant at 5% you must assume that at least one is a chance finding. To control for this researchers are 'supposed' to plan there hypothesis test comparisons before they look at their data, only testing for relationships for which they have already presented sound reason to expect. In exploratory data analysis you apply the "bonforoni correction" (my apologies to senor bonforoni's family for butchering his name) You divide your significance level (say 5%) by the number of comparisons you make to get and honest look at whether these are indeed real relationships. Table two of the study presents 31 comparisons, 25 of which are said to be 'significant' at 1%. Applying the EDA (fishing expedition) correction for 31 unplanned comparisons gives a criterion of 0.034% if you want to assert that none of the apparent relationships are due to chance alone.
One can not tell from reading the article how many of these were planned vs unplanned comparisons. The 'scientific' answer is to use this and similar studies to plan comparisons in future studies from which statistical hypothesis test conclusions can be drawn.
(this aside present in appreciation of all you tax payers who supported me through grad school ; -))

Spin The NYT article presents this study as prejudicial behavior of non black refs against black players. Reading the research article you find two interesting things

The ref's behavior was for, not against - they called fewer fouls against players similar to themselves, not more against dis-similar players
The 'favoritism worked both ways black refs called fewer fouls against black players than against non black players.

Even more interesting - Player behavior was different depending on similarity / dis-similarity player to ref -- steals/turnovers were less frequent on part of players who were dis-similar in color to refs.

I think we have to admit that all of us from time to time act differently toward those different (in culturally defined ways) from us, and, we can act differently in the presence of same.

Red Corvette speeding ticket effect -- the researchers also note previous research indicating that black jerseyed players are called for fouls more often than other uniforms.

I do not think this is a "far-fetched pursuit". I am very concerned about the practice of science education by press release - we don not know what thee research will look like after peer review and publication -but, I suspect we can predict that media spun 'facts' will be remembered by the 'masses' rather than the properly critiqued interesting apparent relationships.