Measuring Desistance via Decay: Making Criminal History Dynamic

• In Event: Rutgers at 50: Reentry and Reintegration

Abstract

There is a well-established link between past and future criminal behavior. Despite this criminological “fact,” scholars have demonstrated that the risk of recidivism declines with time; that is, virtually all individuals eventually stop committing crime. However, measuring the point at which an individual’s past is no longer predictive of future criminal offending has mostly eluded researchers. This element of desistance should be considered by criminal justice system practitioners when determining the importance of criminal records for future risk classification. While today’s risk assessments consider a wide variety of dynamic factors, a common critique is their overuse of static criminal history indicators. Specifically, that these items are weighted heavily, preventing an individual’s ability to reduce their risk score over time, and an inability to display desistance. Using a large sample of Washington State reentrants (N~400,000), we use a dynamic scoring procedure, termed a ‘decay function’, to reduce an individual’s criminal history scores annually. Study findings indicate risk models built with decay functions provide a more accurate prediction of reoffending and have theoretical and policy implications regarding desistance and redemption, and racial/ethnic bias reduction thought to be inherent to criminal history indicators.

Authors

• Zachary Hamilton, University of Nebraska Omaha

• Jennifer J. Tostlebe, University of Nebraska Omaha

• John Michael Ursino, University of Nebraska Omaha

• Baylee Allen-Flores, University of Nebraska Omaha

• Alex Kigerl, University of Nebraska Omaha