[_ptgt]

One concern I've always had when people discuss recidivism rates is that it's sort of impossible to find an effective control group, no?

Suppose you're fresh out of prison with no good employment options. My guess is that, when people talk about recidivism rates, that either look at it nominally or they try to control for socioeconomic status. But apart from socioeconomic status, I can see two major factors that might cause recidivism: 1) the personality that led you to end up in prison in the first place, and 2) the effect on your personality that your time in prison had.

Since the only way to study the personality of the prisoner population is to do so _after_ they go to prison, it seems impossible to isolate either effect. Maybe there's a clever study design that let's you do it.

Failing that though, it's sort of dishonest to assume that rehabilitation attempts or changing the prison environment in some way would be successful in reducing recidivism rates "if we would only try them."

What is the actual evidence on the efficacy of rehabilitation programs?

[canadaduane]

We can compare prison systems across populations. Bastøy prison in Norway has a recidivism rate of just 16%. It seems that all things being equal (i.e. if you can assume that "personality" averages out of the statistical equation when comparing across human populations) then they are doing something very important and remarkable for their people there. See https://www.theguardian.com/society/2013/sep/04/bastoy-norwe...

[_ptgt]

I think it's interesting exploratory work, but you can't just assume that the "personality" of Norwegian prisoners is the same as American ones. There are vast environmental and cultural between the two populations that play a causal role in personalities; to assume that personality differences would "average out" is not a reasonable assumption.

I don't mean to say that the Norwegian model can't or wouldn't work in the US, but it can't be considered any more than "exploratory" from a US policy perspective.