Monday, June 29, 2009

Was Mr. Ricci Treated Like a Statistic?

In an interview broadcast by NPR, Mr. Ricci, the lead plaintiff in the New Haven firefighters' case, said that New Haven did him a wrong because it treated him like a statistic. Mr. Ricci may have gotten things backwards:
If Ricci had been treated like a statistic, he (probably) would not have been treated like a statistic.
That's because statistical analysis, properly done, might have shown that it was no great surprise that only White firefighters were eligible for a promotion to the vacant positions.

If we assume that New Haven was acting in good faith, New Haven was unduly impressed by the statistic that only White firefighters were eligible for a promotion. So if Mr. Ricci was a victim of statistical argument, the problem was perhaps that he was victimized by very bad statistical argument.

However, the Supreme Court of the United States did not do any better than New Haven did.

Just as you can't show that some set of results is suspicious just by eyeballing the statistics, you can't show that the results are non-suspicious just by eyeballing the statistics.

But that's what the Supreme Court effectively did when it rejected New Haven's characterization of the test results as "raw racial results."

The question comes down to this: Whose (statistical) ignorance do you prefer, New Haven's or the Court's?

  • The Court gave its own ignorance the preferred position, by putting the burden on New Haven to show that the results of the test were racially discriminatory or suspect.
  • The above are off-the-cuff reactions based on a partial reading of the opinion. So I may yet end up retracting, or eating, my off-the-cuff sentiments.

    &&&

    The dynamic evidence page

    Coming soon: the law of evidence on Spindle Law

    No comments: