Modeling Recidivism Times: Comment on Bloom

Abstract

Bloom (1979) discussed a new mathematical model that has considerable potential for facilitating program evaluations based on analyses of recidivism data. However, he overlooked a rich literature in other fields, such as survival analysis and reliability theory, dealing both with his particular model and the general type of process of which recidivism may be considered an example. Use of results from that literature could have streamlined his discussion. More fundamentally, Bloom did not carry his consideration of the behavioral mechanisms underlying his data to the point where his particular model specification, and no others, followed as a logical necessity. This limits the generalizability of his approach. One possible formulation of underlying relationships leading directly to Bloom's model is suggested.