Two preeminent theoretical models: A proportional hazard rate analysis of recidivism

This was a three-year follow-up study of survival in the community without arrests or a parole violation of 480 male graduates of the only boot camp for adults in Arkansas. The purpose was to determine the relative viability of elements from social selection and social causation models in predicting this survival. Using a proportional hazard rate analysis, the findings indicated that elements of both models were significant predictors of survival. Attachment to caregivers, physical and sexual abuse during childhood, and various personality deficits were found to be useful predictors from the selection model, whereas beliefs, religiosity, caregiver monitoring, and various aspects of peer affiliations were predictors of survival from the causation model. Conceptual, methodological, and intervention implications of the findings were discussed. (C) 2002 Elsevier Science Ltd. All rights reserved.