One of the most striking innovations in the criminal justice system during the past thirty years has been the introduction of actuarial methods – statistical models and software programs –designed to help judges and prosecutors assess the risk of criminal offenders. Predictive algorithms are currently used in four major areas of the U.S. criminal justice system: pretrial and bail, sentencing, probation and parole, and juvenile justice. These algorithms consider a small number of variables about a defendant – either connected to her or his criminal history (previous offenses, failure to appear in court, violent offenses, etc.) or socio-demographic characteristics (age, sex, employment status, drug history, etc.) – in an effort to predict a defendant’s risk of recidivism or their likelihood to fail to appear in court if they are let out on bail.

This document is a workshop primer from Data & Civil Rights: A New Era of Policing and Justice.