Résumé
This study provides an overview of prediction and classification assessment instruments used by many North American probation departments, and examines the construction of several risk prediction models. Using a sample of Texas offenders, the author identifies three dependent variables used in developing effective prediction models: successful probation, re-arrest, and probation revocation. Testing the validity and reliability of predictor variables, the author indicates that four index scales, education, employment, substance abuse, and criminal history, are the most effectively used predictors of recidivism. The poor performance of the Wisconsin assessment instrument is identified, and the author suggests that replacing that model with a battery of assessments grounded in theory. These would be more effective because they take into account both offender characteristics and the community in which he or she resides. The author suggests that significant improvements can be made to offender risk prediction instruments if such instruments are linked to criminology theory.
Contenu
1. Introduction. – 2. Overview of prediction and classification. – 3. Research design and methodology. – 4. Descriptive analysis of the felony cohort. – 5. Testing the validity and reliability of predictor variables. – 6. The construction of rearrest, revocations and successful probation prediction models. – 7. Discussion and conclusion.