By U.S. Department of Justice Office of Justice Programs

The First Step Act of 2018 (FSA) mandated the development and implementation of a risk and needs assessment system in the Federal Bureau of Prisons (FBOP). The FSA also required that the U.S. Department of Justice (DOJ) review, validate, and publicly release the risk and needs assessment system on an annual basis. The National Institute of Justice (NIJ) contracted with Dr. Rhys Hester and Dr. Ryan Labrecque as consultants for the annual review and revalidation of the Prisoner Assessment Tool Targeting Estimated Risk and Needs (PATTERN). This document is the fourth review and revalidation report, following USDOJ (2021a), USDOJ (2021b), and USDOJ (2023). The previous review and revalidation reports included FBOP release cohorts from fiscal year (FY) 2009 through FY 2018. The current report analyzes a subsequent cohort of FY 2019 FBOP releasees and evaluates PATTERN for its predictive accuracy, dynamic validity, and racial and ethnic neutrality, as mandated by the FSA. This study analyzes one-, two-, and three-year recidivism outcomes, assesses what proportions of change in risk scores and levels are influenced by the current age item, and provides additional descriptive information on individual items, risk scores and levels, and outcomes by race and ethnic group. Finally, this report provides updates on the actions taken by NIJ and DOJ in the past year and the ongoing efforts to review and improve PATTERN. The FY 2019 cohort study finds that PATTERN remains a strong and valid predictor of general and violent recidivism at the one-, two-, and three-year follow-up periods, with Area Under the Curve (AUC) statistics ranging from .745 to .776. Comparisons of recidivism rates by risk level category (RLC) and predictive value analyses by risk level grouping also continue to indicate that such risk level designations provide meaningful distinctions of recidivism risk. In addition, individuals can change their risk scores and levels during confinement beyond mere age effects. Those who reduced their RLC from first to last assessment were shown to have the lowest recidivism rates, followed by those who maintained the same risk level and those with a higher risk level, respectively. While the findings continue to indicate PATTERN’s accuracy across the five racial and ethnic groups analyzed, there remains evidence that the instruments predict differently across those groups, including overprediction of risk of Black, Hispanic, and Asian males and females, relative to White individuals, on the general recidivism tools.

Washington, DC: U.S. Department of Justice Office of Justice Programs 2024. 36p.