To test the possibility that in-prison behaviour and cognition provide information useful in predicting future desistance from crime, two in-prison variables and nine pre-prison and demographic control variables were correlated with post-prison release success in a group of 1,101 male inmates released from federal prison. A Cox regression proportional hazards survival analysis revealed that fewer disciplinary infractions and lower criminal thinking predicted future desistance, as measured by the absence of post-release arrests or a longer time until first arrest for those who were arrested, net the effects of the pre-prison variables and demographic measures. When disciplinary infractions were subclassified as aggressive (fighting, assault, threatening) or non-aggressive (disobedience, theft, use of intoxicants), only the non-aggressive category achieved significance. Likewise, when criminal thinking was subdivided into proactive and reactive criminal thinking, only the reactive dimension achieved significance. These findings suggest that behaviour and cognition assessed in prison may have value both in predicting desistance upon a person’s release from prison and in clarifying the nature of post-prison release success.
Glenn D. Walters
Canadian Journal of Criminology and Criminal Justice, September 21, 2020