Friday, November 9, 2007

Faith-Based Prison Evaluation

The academic literature has been thin regarding evaluations of faith-based prisons in the U.S. Mainly because these prison programs aren't that old, it's been difficult to find thorough evaluative research by which to measure the astoundingly low recidivism rates these programs claim.

Via Michael Connelly's Corrections Sentencing blog, the Urban Institute has completed a study of Florida's two Faith - and Character - Based Institutions (FCBI) that deal specifically with faith-based prison initiatives. A copy of the report is available here.

I've only scrolled through the Executive Summary and abstract, but it appears interesting on two fronts. First, on the generalizability and selection bias criticisms of the program, the UI's report found that "the impact analysis component of this evaluation, however, only found evidence of selection bias for one subset of the male FCBI participants, suggesting that the FCBI model is generalizable to a broader population of inmates."

This would challenge the prevailing finding that these programs, while successful, tend only to benefit those already predisposed to faith and organized religion. In other words, there is a certain amount of "preaching to the choir" in that all inmates who attend FCBI's volunteer and are therefore either interested in furthering their religiosity or intent on "working the system" for more nefarious reasons. This latter point is addressed in the report, which notes, "even those inmates who sign up for the FCBI strictly to be closer to family or to be in a more peaceful prison environment nevertheless benefit from their time there."

On the second front, recidivism, the UI reports a more negative relationship. While there appears to be a slightly lower recidivism rate amongst inmates who complete the FCBI in comparison to general population inmates, six months after release, the differences disappear at the twelve month mark. A lack of statistical significance on this front would also call into question program advocates claims of "92% success rates" and so on.

The overall conclusion of the report calls for more research in both these crucial areas, and I would agree. There really hasn't been enough time to conduct longitudinal studies regarding recidivism, and we definitely need more data on selection bias and "market share", if you will, regarding whether these faith-based alternatives can work on statistically significant groups of inmates.

First amendment concerns also need to be addressed, especially since the report notes that prison staff feel the inordinate amount of contact the inmates have with "religious volunteers" might be undercutting their authority and jeopardizing institutional security.

No comments: