A Propensity Score Approach towards Assessing Neighborhood Risk of Parole Revocation
Previous research pertaining to parole has primarily concentrated on individual level characteristics as predictors of parole success, holding offender residency as a constant. Recently, there has been a growing realization that we need to understand how neighborhood contexts impact successful reint...
| Main Author: | |
|---|---|
| Contributors: | ; |
| Format: | Electronic Article |
| Language: | English |
| Published: |
2015
|
| In: |
American journal of criminal justice
Year: 2015, Volume: 40, Issue: 2, Pages: 377-398 |
| Online Access: |
Volltext (lizenzpflichtig) Volltext (lizenzpflichtig) |
| Journals Online & Print: | Drawer...
|
| Check availability: | HBZ Gateway |
| Keywords: |
| Summary: | Previous research pertaining to parole has primarily concentrated on individual level characteristics as predictors of parole success, holding offender residency as a constant. Recently, there has been a growing realization that we need to understand how neighborhood contexts impact successful reintegration. It is difficult to evaluate the increased risk associated with neighborhood characteristics on parole revocation because residency after prison is not randomly assigned. Consequently, methodological adaptations must be made to account for selection into disadvantaged neighborhoods. To reduce overt bias, we used two propensity score strategies, matching and inverse probability of treatment weighting, to determine the impact of parolee residence on parole outcome. Our analytic strategies indicated a weaker effect for neighborhood context when compared to traditional models. In essence, we found that both techniques were effective at decreasing the threat to internal validity allowing for greater confidence in reporting an average treatment effect. The implications of our findings are discussed as well as the utility of these analyses in criminal justice research. |
|---|---|
| ISSN: | 1936-1351 |
| DOI: | 10.1007/s12103-014-9269-z |