Improving the Accuracy and Fairness of Pretrial Release Decisions: A Multi-Site Study of Risk Assessments Implemented in Four Counties

The implementation of pretrial risk assessments into pretrial decision-making has emerged as a principal component of efforts to increase the efficiency, consistency, and accuracy of pretrial release decisions. However, risk assessments have proliferated rapidly in pretrial settings without strong evidence on their effectiveness in achieving the goals of evidence-based pretrial release reform. This study proposes to address this gap in the literature by conducting a multi-site, quasi-experimental investigation of risk assessment-informed pretrial release decisions relative to decisions made without risk assessments. The objectives of this study are to investigate whether use of pretrial risk assessments: 1) improves risk management decisions relative to practice as usual; 2) improves the fairness of pretrial release decisions for racial minorities relative to practice as usual; and 3) reveals judge-level variation in pretrial release decisions through administrative court and jail records from 2014 through 2018 for four Indiana counties currently piloting a risk assessment instrument (Indiana Risk Assessment System Pretrial Assessment Tool; IRAS- PAT) in pretrial case processing. The study will employ a multiple nonequivalent comparison group design with propensity score matching to approximate an experimental design. Defendants who receive pretrial risk assessments during the pilot implementation periods will be matched separately to two comparison groups consisting of defendants with case records prior to the start of the pilot project and defendants with case records during the pilot project but who did not receive a risk assessment. Primary outcomes will include both pretrial release (i.e., release on recognizance, bail bond issued, bail amount, and pretrial detention) and pretrial misconduct (i.e., arrest and failure to appear) outcomes. Weighted multivariate analyses will assess for differences in outcomes by risk assessment condition and examine defendant race and adherence to structured guidelines (i.e., decision grids) as a moderator of effects (i.e., Objectives 1 and 2). Multi-level modeling will be conducted to examine judge-level variation in pretrial release decisions (i.e., Objective 3). Findings will inform whether implementation of structured risk assessments reduces rates of pretrial misconduct while improving the fairness of pretrial release decisions for racial minority defendants relative to practice as usual. "Note: This project contains a research and/or development component, as defined in applicable law," and complies with Part 200 Uniform Requirements - 2 CFR 200.210(a)(14). CA/NCF