NCJ Number
80477
Journal
Federal Probation Volume: 45 Issue: 3 Dated: (September 1981) Pages: 25-32
Date Published
1981
Length
8 pages
Annotation
Major findings are presented from research that examined the effectiveness of devices for probation outcome prediction used in new Federal procedures for classifying probationers.
Abstract
Four base-expectancy-scale prediction models were selected for validation and comparative evaluation: (1) the Modified California BE61A, developed by the State of California; (2) the Revised Oregon Model, developed in the U.S. Probation Office for the District of Oregon; (3) the U.S. Parole Commission's Salient Factor Score; and (4) the U.S.D.C. 74 Scale, developed in the U.S. Probation Office of the District of Columbia. These models were selected for study because each was already being used in several districts as a classification tool. All four models are heavily dependent on items relating to prior criminal record. In addition, all the models have social or economic stability variables, such as employment history, residential stability, and drug or alcohol involvement. Geographical and size criteria were used to choose nine Federal probation offices throughout the country for validating and comparing the models. A systematic sample of 300 offenders was drawn from each of the district listings for 1974. A risk score was computed for every offender according to the scoring directions for each of the models. All four of the models were found to be valid for making risk assessments in the Federal probation system. It is recommended, however, that U.S.D.C. 75 be implemented systemwide as a caseload classification tool because of its relative predictive power, lower administrative costs, and its lower bias to an offender's race or sex. Tabular data and 10 footnotes are provided.