Jessica McBride has undertaken just the type of comparative analysis that I think is most illuminating. She has explained her methodology - she looked at all criminal cases and calculated the percentage in which each Justice voted in favor of a defendant's claim.
She finds that Justice Butler voted to do so more than any other Justice but the Chief and that the differences are significant. He voted in favor of a claim presented by a criminal defendant in 58% of the cases. Justices Prosser, Roggensack and Wilcox did so in less than 30% of cases.
A couple of preliminary points are in order. First, the question here is not who was right in these cases. What we are trying to do is determine whether there is a difference in philosophy (or some other factor affecting the likelihood of ruling in favor of a criminal defendant)among various justices.
Second, it doesn't matter if, in these cases, a Justice also voted to deny relief to a criminal defendant as long as we calculate the number in the same way for each Justice. This is because we are not making a claim about any individual case (i.e., the question is not did the defendant win or how much did he win) but simply looking for aggregate differences among the justices.
For those of you are convinced that there is something illegitimate about this type of analysis, it's not all that unusual. To give an example close to home, my colleague Jason Czarnezki (soon to decamp for Vermont) did some work in which he attempted to assess whether justices' voting patterns were affected by the proximity of elections and certain other factors. To so, he coded votes as in favor or against a criminal defendant's claims. He then was able to characterize certain justices' voting patterns in criminal cases as "conformist" (by which he meant having a tendency to vote against defendants which would be in conformity with public opinion) and "moderate" or "nonconformist." He then looked at how these patterns were affected by things like an approaching election.
The assumption is that, if you have enough cases, differences in aggregate numbers mean something. Justice Butler is not three times as likely to vote for the claim of a criminal defendant by chance. There is some type of difference in approach between them. Just what constitutes that difference and what conclusions you draw from it is another matter.
McBride's numbers are, of course, consistent with what any lawyer who is familiar with the Court will tell you. Chief Justice Abrahamson and Justices Butler and Walsh Bradley are easier sells for criminal defendants on appeal.
Two questions remain. Did McBride get the numbers right and what do they mean?
I have to go and judge a moot court practice now so I'll try to get to that this afternoon.