Empirical Evaluation of Prison Classification: The Manitoba Case

Michael Weinrath, University of Winnipeg
Ron Coles, Manitoba Corrections


As classification systems have evolved, some investigators have recommended the use of one instrument for both community based and incarcerated offenders. This Manitoba, Canada based project compared a community/institutional instrument, the Primary Risk Assessment (PRA), with the Institutional Security Assessment (ISA), an instrument intended for use specifically in custodial settings . A validation study was conducted to assess the effectiveness of both instruments using three outcome criteria: proportion of inmates assigned minimum security, institutional misconduct and recidivism rates as outcomes for a sample of 1000+ inmates. A secondary research goal was a comparison of static and dynamic risk factors. The ISA is a research-based tool which is weighted heavily towards static risk factors such as demographics and prior behavior that are not subject to change, while, the PRA is a "Wisconsin model" instrument that makes heavier use of dynamic risk factors such as attitude and education that are amenable to modification.

Results indicated that both instruments were effective in classification but, overall, the ISA performed better than the PRA, while static factors proved more helpful than dynamic factors in accurately classifying provincial inmates. Rather than pick one tool over the other, study findings supported integration of the two instruments to more accurately classify inmates. It is recommended that policy makers follow-up "progressive innovations" with substantive evaluation research to improve management of inmate populations.

(Return to Program Resources)

Updated 05/20/2006