NCJ Number
59455
Date Published
1979
Length
47 pages
Annotation
THIS PAPER PRESENTS A DISCUSSION OF METHODS AVAILABLE FOR CONDUCTING MODEL-BASED EVALUATIONS OF TREATMENT-RELEASE CORRECTIONS PROGRAMS.
Abstract
WITHIN THE CORRECTIONS COMPONENT OF THE CRIMINAL JUSTICE SYSTEM, A RANGE OF PROGRAMS AIMED AT THE REHABILITATION OF SELECTED INDIVIDUALS HAS BEEN ESTABLISHED. THESE PROGRAMS ALL SHARE THE FOLLOWING FEATURE: INDIVIDUALS COMMITTED TO A PROGRAM ARE SUBJECTED TO A PERIOD OF 'TREATMENT'; UPON SATISFACTORY COMPLETION OF THE TREATMENT PERIOD, THESE INDIVIDUALS ARE 'RELEASED.' OF INTEREST TO OFFICIALS OF SUCH PROGRAMS IS THE LIKELIHOOD THAT A RANDOMLY CHOSEN PROGRAM CLIENT WILL COMMIT AN OFFENSE AFTER RELEASE. THE LIKELIHOOD OF THIS EVENT IS TERMED THE 'RECIDIVISM PROBABILITY.' THIS PAPER DISCUSSES THE METHODOLOGY FOR CONDUCTING MODEL-BASED EVALUATIONS OF TREATMENT-RELEASE CORRECTIONS PROGRAMS. MODEL-BASED TECHNIQUES PROVE USEFUL IN EVALUATING POLICE PATROL PROGRAMS MIGHT BE HELPFUL IN EVALUATING THESE TREATMENT PROGRAMS. A GENERAL MODEL FOR REARREST PATTERNS OVER TIME WOULD MEASURE TIME FROM INDIVIDUAL RELEASE AND CALCULATE THE PROBABILITY THAT THE RELEASED INDIVIDUAL WILL BE REARRESTED AT A FUTURE TIME. THE MODEL DEPENDS ON THE HAZARD FUNCTION OR PROBABILITY OF FAILURE. THE BEHAVIOR OF THIS MODEL IS THEN EXAMINED UNDER ALTERNATIVE ASSUMPTIONS IN A NUMERICAL EXAMPLE. CLASSICAL AND BAYESIAN ESTIMATION METHODS ARE USED TO ESTIMATE MODEL PARAMETERS. A BRIEF DISCUSSION IS PRESENTED OF POSSIBLE EXTENSIONS TO THE METHODS OUTLINED IN THIS EVALUATION. TABULAR DATA AND REFERENCES ARE INCLUDED. AN APPENDIX PRESENTS THE BLOOM AND MALTZ-MCCLEARY MODELS WHICH INVOLVE AN EXPONENTIAL HAZARD FUNCTION. (AUTHOR ABSTRACT MODIFIED--MJW)