Improving the Accuracy and Fairness of Pretrial Release Decisions: A Multi-Site Study of Risk Assessments Implemented in Four Counties, Indiana, 2015-2018 (ICPSR 37829)

Time Period(s) View help for Time Period(s)

Date of Collection View help for Date of Collection

Study Purpose View help for Study Purpose

Risk assessments have emerged as a principal component of evidence-based reform efforts to increase the efficiency, consistency, and accuracy of pretrial release decisions. However, as an evidence-based strategy, risk assessments have proliferated rapidly in pretrial settings without strong evidence of their effectiveness in achieving these goals. Further, although risk assessments have the potential to reduce well-documented, pervasive racial disparities in pretrial case processing, whether risk assessments improve the fairness of pretrial release decisions relative to practice as usual remains to be seen. The overall goal was to assess differences in the accuracy and fairness of pretrial release decisions made with and without risk assessment information. In particular, the study objectives were to investigate whether use of pretrial risk assessments: 1) improved risk management decisions relative to practice as usual; 2) improved the fairness of pretrial release decisions for racial minorities relative to practice as usual; and 3) revealed judge-level variation in pretrial release decisions.

Study Design View help for Study Design

Administrative court and jail records were procured from four Indiana counties that piloted a risk assessment instrument (Indiana Risk Assessment System - Pretrial Assessment Tool (IRAS-PAT)) in 2016 to inform pretrial decision-making. Using a multiple nonequivalent comparison group design, defendants who received a pretrial risk assessment during the 1-year pilot period (Pilot condition; n = 2,631) were matched to two separate groups of pretrial defendants without risk assessments. The first comparison group consisted of pretrial defendants who were booked during the 1-year pilot period but did not receive a risk assessment (Pilot Comparison condition; n = 1,580). The second comparison group consisted of pretrial defendants who were booked during the 1-year period prior to the pilot period (Pre-Pilot Comparison condition; n = 3,185).

Sample View help for Sample

Across all three groups and four counties, participants represent unique individuals who were booked into each county's local jail between 2015 and 2017 and had formal charges filed in court. All participants had a case disposition by December 31st, 2018.

Universe View help for Universe

Pretrial defendants who were booked into jail in four Indiana counties during the study period.

Unit(s) of Observation View help for Unit(s) of Observation

Data Type(s) View help for Data Type(s)

Mode of Data Collection View help for Mode of Data Collection

Response Rates View help for Response Rates

Not applicable.

Presence of Common Scales View help for Presence of Common Scales

None

Original Release Date View help for Original Release Date

2021-02-25

Version History View help for Version History

Weight View help for Weight

There are four sets of propensity score weights in this dataset. Propensity score weights should be applied according to each comparison (i.e., pilot vs. pilot comparison OR pilot vs. pre-pilot comparison). The data are designed to be subset by comparison group. There are two variables in the dataset for this purpose (SAMPLE_PILOT and SAMPLE_PREPILOT). Two-group propensity score weights were generated for comparisons using the SAMPLE variable. Three-group propensity score weights were generated for comparisons using the ADHERENCESAMPLE variable, designed for sensitivity analyses. Associated syntax for these weights can be found in the accompanying Syntax Documentation.