Stop the Redundant Interviews! Using Institutional Data to Predict Inmate Risk & Needs

• In Event: The Role of Risk/Needs Assessment in the Correctional System

Abstract

Dr. Meredith led the team that developed a series of 3 risk and 8 needs algorithms for both male and female inmates to replace existing paper-and-pencil risk/needs assessments used to a guide treatment plans. State-operated criminal history repository, prison, probation and parole operations databases were linked to conduct extensive research on 39,000 Georgia inmates released 2008-2009 and 27,000 inmates administered personal interview assessments in 2012-2013. Data reduction, item response theory and logistic regression techniques were used to create each scale of public safety risk upon prison release and treatment need during the prison stay. All scales were validated on a historical cohort of 17,000 inmates released from prison in 2010, where the personal interview was replaced by institutional data predictions. Validation results indicate equal and superior performance in predicting outcomes of interest: new arrest activity, technical violations of community supervision, drug test failure on parole. Findings suggest that agencies can replace hour-long personal interviews, often replete with questions asked at multiple junctures, with key measures from institutional data.

Author

• Tammy Meredith, Applied Research Services, Inc.