NCJ Number
176483
Date Published
1997
Length
15 pages
Annotation
This report presents the methodology and findings to date of a 3-year study designed to determine the relative reliability and validity of three approaches to risk assessment currently used in child protective services (CPS).
Abstract
Generally, CPS risk-assessment systems fall into two basic types: consensus-based systems, in which workers assess specific client characteristics identified by the consensus judgment of experts and then exercise their own clinical judgment about the risk of future abuse or neglect; and actuarial systems, which are based on an empirical study of CPS cases and future abuse/neglect outcomes. Two of the more widely used versions of consensus-based systems are the Washington Risk Assessment Matrix, a risk assessment system developed by practitioners in Washington State, and the California Family Assessment Factor Analysis, a derivative of the Illinois CANTS system. The actuarial-based approach identified in this study was the Michigan Structured Decision Making System's Family Risk Assessment of Abuse and Neglect, a research-based tool constructed during a study of 2,000 Michigan families and recently revalidated with a cohort of 1,000 families. An analysis of these three systems found that although none of them approached 100 percent inter-rater reliability, raters who used the Michigan system made consistent risk estimates for a high percentage of the cases they assessed, and inter-rater reliability for the Michigan model was much higher than that achieved by the other systems. The level of inter-rater reliability attained by the California and Washington risk assessment models was well below what could be considered adequate. Suggestions are offered for enhancing the California and Washington systems. Implications of these findings are discussed. 5 figures, 5 notes, and 20 references