NCJ Number
147395
Date Published
1991
Length
63 pages
Annotation
This is a "how-to" manual on procedures for validating risk-assessment instruments used in community corrections.
Abstract
The introduction provides a historical perspective on risk-assessment instruments and their validation, followed by a discussion of validation issues. These include the need for validation, the definition of risk assessment, and the properties of good risk assessment systems. Needs scale validation is also discussed in the introduction. A section on study parameters considers criterion variables, the follow-up period, sampling parameters, and database requirements. The section on analysis addresses the approach to validation and the frequency of validation efforts. A stepwise procedure for conducting a validation study tests the instrument against the entire sample to determine the degree to which offender groups are separated relative to outcomes. The validation study also examines whether existing cutoff scores are the optimal thresholds for defining risk groups. Additionally, the test illustrates how sample cases are distributed among the existing risk classifications. In discussing the use of validation study results, the monograph focuses on the distribution of cases, equity issues, and where to go for assistance. Results of recent validation studies are used throughout the monograph to illustrate specific points and to clarify the discussion of issues. 16 tables, 3 figures, and a 20-item bibliography