A System for Computer-Aided Probation Decision-Making

Publisher Website
Google Scholar
Abstract
This paper describes the application of a modern theory of decision-making and computer technology to the process of juve ni !e probation decision-making. It is based on approximately four years of research, in three phases: 1. A conceptual model of the juvenile probation process in California was developed that specified the major treatment and disposition points for probation officers. 2. A comprehensive follow-up study was made on 2,290 juve niles referred to probation departments in seven California coun ties. Background and personal information and complete proba tion records over a one-year period were gathered for each juvenile to predict what would happen to him as a result of background characteristics and what disposition-placement alternative he would receive. Two prediction criteria were used: (a) recidivism (the number of repeat offenses) and (b) behavior "improvement- deterioration" (the score on a scale that described each juvenile's behavior after the probation process as improved, staying the same, or getting worse). A conditional probability model to pre dict these criteria on the basis of the information variables was developed and tested. 3. Using the information and results of phase 2, an "on-line" computer system. was developed to aid probation officers to make decisions and recommendations about various treatments for individual youngsters. This system allows information relevant to each youth to be combined and presented to the officer on his request. One component of this system is a Bayesian decision model, designed to reduce the uncertainty of the probation decision-maker by allowing him to select whatever disposition- treatment alternative will have the highest probability of success.

This publication has 0 references indexed in Scilit: