Predicting Violent and Nonviolent Recidivism of Incarcerated Male Offenders

• NCJ Number: 200559
• Journal: Aggression and Violent Behavior: A Review Journal Volume: 8 Issue: 2 Dated: March-April 2003 Pages: 175-203
• Author(s): Wagdy Loza
• Date Published: March 2003
• Length: 29 pages
• Publication Format: Article
• Publication Type: Literature Review
• Language: English
• Country: United States of America

Annotation

This article examines the challenges inherent in predicting the violent and nonviolent recidivism of incarcerated adult male offenders.

Abstract

The United States penal system relies on clinicians to predict whether offenders are likely to re-offend once released into society. Decisions by parole boards are often based upon these predictions of recidivism. As such, a degree of controversy surrounds the practice of recidivism prediction. Opponents argue that there is no evidence that clinicians can reliably and accurately predict who is likely to recidivate and that this type of prediction violates offenders civil liberties. The author reviews the problems surrounding accurate predictions, the variables associated with recidivism prediction, and the process of completing prediction reports and communicating their findings. Among the problems associated with prediction are base rate problems; a lack of specificity regarding the etiology of, and even the definition of, violence and aggression; a lack of feedback needed to test the accuracy of predictions; and a failure by clinicians to account for environmental or situational variables. Also widely noted as a challenge in the prediction of recidivism is the lack of agreement on the psychological instruments used to assess the likelihood that an offender will re-offend if released into society. Among the variables examined to predict recidivism are age, criminal history, childhood behavioral problems, mental illness, substance abuse problems, and educational attainment. Finally, the author discusses how clinicians complete prediction reports and communicate them to decisionmakers. Suggestions are offered for how clinicians may increase the accuracy of their predictions, including the necessity of having a complete understanding of all the myriad cultural, contextual, structural, and psychological variables involved in predicting recidivism.