Thursday, December 20, 2007

TN Sentencing Report

A friend tipped us off to this latest report on sentencing in TN, a very nice example for those of you who either need to do these reports or do them better. I'm biased, of course, because the report is the first case of a state report that I'm aware of that has actually presented the state's "conventional number preferences," a concept that I've published on with Andy Wiseman in WI and Brian Ostrom and colleagues pioneered in MI. What are CNPs?

In a 10-year sentence, there are 120 possible month sentences for judges to give and prosecutors and defense to recommend. If you study it though, you'll find that the overwhelming majority of sentences are multiples of 6 and are limited usually to 6, 12, 18, 24, 36, 48, 60, 72, 84, 120 months. So much for "reasoned sentencing" if it just amounts to things judges feel comfortable dividing. What makes multiples of 6 sacred or more effective? Imagine if the multiples were 4. In a 120-month possible sentence, instead of those above, you'd have 4, 8, 12, 20, 32, 40, 48, 64, 80, 120. In each case, you'd be cutting off months of bedspace with absolutely no less justification for the sentence than you have with multiples of 6. If we'd started with those sentences, we wouldn't even notice the difference. (Not to mention what this does say about the links between sentencing goals and the sentences that are supposed to achieve them and about the fact that, with or without institutionalized guidelines, judges are using structured sentencing, usually without even realizing it. So much for judicial opposition or claims of "discretion.")

As I said, the examples of this to this point have been academic articles. The TN report is a pioneer in bringing the concept to the attention of the public and to the practitioners. Let's hope others realize its implications and utility for policy decisions as well. And congratulations to the state and the author for being so forward-thinking and ground-breaking. And so brilliant to cite me.

No comments: