U.S. flag

An official website of the United States government, Department of Justice.

NCJRS Virtual Library

The Virtual Library houses over 235,000 criminal justice resources, including all known OJP works.
Click here to search the NCJRS Virtual Library

Criminal Court Assessment Tool: Development and Validation

NCJ Number
Sarah Picard-Fritsche; Michael Rempei; Ashmini Kerodai; Julian Adler
Date Published
January 2018
43 pages
This report summarizes the development of the Criminal Court Assessment Tool (C-CAT) developed by the Center for Court Innovation, as well as the results of a validation study with a sample of defendants drawn from the Brooklyn Criminal Court in New York City.
The C-CAT addresses a shortage of risk-need assessment tools that identify important needs fueling a defendant's criminal behavior and can be efficiently administered in high-volume settings and inform referral to effective intervention. Drawing on risk-need-responsivity theory and input from a panel of external experts, the Center for Court Innovation developed a comprehensive 183-item assessment tool that covered 18 risk and need domains. It was pilot-tested with 964 defendants who appeared in three misdemeanor diversion programs in New York City. The tool was ultimately reduced to those items statistically associated with recidivism or that were considered key "responsivity" factors (trauma and mental health). This 30-item risk-need assessment became the first version of the C-CAT. A second sample of 928 defendants awaiting arraignment in Brooklyn were reassessed using the original C-CAT tool. This second sample consisted of a representative array of felony, misdemeanor, and violent-level defendants awaiting arraignment in the Brooklyn Criminal Court. This sample was more representative of a general criminal court population. The intent of this testing was to confirm the tool's predictive accuracy and refine the tool based on new data. The analysis of results shows that the original C-CAT tool could and should be improved to fit a more diverse defendant population, and the risk algorithm was revised accordingly. Using the validation sample, the revised algorithm was validated. Standard area-under-the-curve (AUC) techniques were used to assess predictive accuracy. 7 tables, 3 figures, 20 references, and appended supplementary data