Revalidation of the Federal Pretrial Risk Assessment Instrument (PTRA): Testing the PTRA for Predictive Biases

Abstract

The Pretrial Risk Assessment (PTRA) instrument was developed for use in the U.S. federal pretrial system. Specifically, this instrument was constructed to help federal officers assess the likelihood that defendants will commit pretrial violations including being re-arrested for any or violent crimes, failing to make court appearances, or having a revocation while on pretrial release. While previous studies have demonstrated the PTRA’s predictive validity, these efforts primarily used development and validation samples and did not investigate the PTRA for predictive bias by defendant demographic characteristics. The current research evaluates the PTRA’s capacity to predict various forms of pretrial violations on 85,369 defendants with officer completed PTRA assessments. Bivariate and multivariate models were estimated by race, ethnicity, and sex. Results show that the PTRA performs well at predicting pretrial violations as the area under the receiver operating characteristic curve ranged from 0.65 to 0.73 depending upon the subsamples and outcomes being predicted. Moreover, the PTRA predicted new criminal arrest activity equally well between non-Hispanic whites and blacks, while for Hispanics and females, findings show the instrument validly predicting re-arrest activity with some evidence of overprediction depending upon the outcome being examined.