Prior Record and Recidivism Risk
An individual’s prior record can have a pronounced impact on the punishment he or she receives for a new offense, substantially increasing the likelihood and duration of an incarceration sentence. Not only does prior record contribute to mass incarceration, but prior research has consistently shown...
| Main Author: | |
|---|---|
| Format: | Electronic Article |
| Language: | English |
| Published: |
2019
|
| In: |
American journal of criminal justice
Year: 2019, Volume: 44, Issue: 3, Pages: 353-375 |
| Online Access: |
Volltext (lizenzpflichtig) Volltext (lizenzpflichtig) |
| Journals Online & Print: | Drawer...
|
| Check availability: | HBZ Gateway |
| Keywords: |
| Summary: | An individual’s prior record can have a pronounced impact on the punishment he or she receives for a new offense, substantially increasing the likelihood and duration of an incarceration sentence. Not only does prior record contribute to mass incarceration, but prior research has consistently shown that criminal history mediates race effects and exacerbates disparities. In guidelines jurisdictions, criminal history enhancements are partially or primarily employed as proxies for risk prevention. But for the most part these scores were not developed empirically, and, to date, whether scores are valid predictors of risk has gone unexplored. This paper uses survival analysis and area under the curve analysis to examine the predictive efficacy of the Pennsylvania Prior Record Score using a sample of offenders sentenced in Pennsylvania and followed-up for 3 years after release ( n = 130,758). The results show that some of the Pennsylvania PRS categories fail to accurately distinguish among offenders based on their likelihood of recidivism. Further, some of the key score components that increase the PRS (and the punishment imposed) have marginal effects on the predictive efficacy of the score, often only increasing the prediction accuracy by a single percentage point. By re-engineering the PRS categories and sub-components, this jurisdiction could recommend less punishment in some cases without any apparent increase in risk to public safety. |
|---|---|
| ISSN: | 1936-1351 |
| DOI: | 10.1007/s12103-018-9460-8 |