December 13, 2005

Lies, Damned Lies, and "Statistics"

I’m late to dissecting the Knight-Ridder hit piece on Judge Alito, which claims to discern an ideological tilt in his decisions. Hugh Hewitt has uncovered one glaring flaw: the reporter excluded cases from the survey because Alito didn’t write the opinion. By that token, assessing Chief Justice Rehnquist’s views on federalism should exclude his dissenting vote in Gonzales v. Raich, since O’Connor and Thomas wielded the dissenters' pen. Common sense says otherwise.

But I noticed another, although lesser, flaw: The survey doesn't tell us how many of Alito's opinions are for a unanimous panel. That's a flaw Judge Frank Easterbrook identifies in Cass Sunstein’s study purporting to identify the biases of reputedly "conservative" judges. As Easterbrook has suggested to statistician John Lott (see page 3 of this Lott paper for the partial Easterbrook quote), the fact a judge frequently votes in favor of one type of party or another in a string of cases doesn't tell us much if most of those cases are unanimous. Unanimous decisions may reflect straightforward questions dictated by circuit or Supreme Court precedent. Therefore, they are a poor basis for inferring bias. Indeed, if a large proportion of a judge's opinions are unanimous, it may reflect his or her fidelity to the rule of law. If the Third Circuit is tough on criminal defendants, and Judge Alito follows precedent, then the fact he doesn't dissent very often from his brethren suggests Judge Alito is faithful to the law of the circuit--not that he's "biased" against defendants.

Links to this post:

Create a Link

<< Home