Residential Mobility and Delinquency Revisited: Causation or Selection?
Objectives To assess the role of selection in the observed association between residential mobility and delinquency among adolescents. Methods This study draws on a sample of adolescents from the National Longitudinal Study of Adolescent Health (Add Health). We first examine whether adjusting regres...
| Main Author: | |
|---|---|
| Contributors: | |
| Format: | Electronic Article |
| Language: | English |
| Published: |
2014
|
| In: |
Journal of quantitative criminology
Year: 2014, Volume: 30, Issue: 2, Pages: 187-214 |
| Online Access: |
Volltext (lizenzpflichtig) Volltext (lizenzpflichtig) |
| Journals Online & Print: | Drawer...
|
| Check availability: | HBZ Gateway |
| Keywords: |
| Summary: | Objectives To assess the role of selection in the observed association between residential mobility and delinquency among adolescents. Methods This study draws on a sample of adolescents from the National Longitudinal Study of Adolescent Health (Add Health). We first examine whether adjusting regression models for several well-established determinants of moving attenuates the association between mobility and delinquency. We then employ propensity score methods to estimate the effect of residential mobility on delinquency among a sub-sample of movers and non-movers who had similar likelihoods of moving. Results The association between mobility and delinquency is significant and positive in regression models, although it is somewhat attenuated by additional control variables that are rarely considered in prior work. However, the distribution of mobility determinants differs substantially across movers and non-movers, potentially biasing these estimates. After covariate balance is achieved using a propensity score approach, we observe no differences in delinquency between groups. Conclusions Results suggest that certain adolescents are more likely to move than others, explaining the observed association between mobility and delinquency. Future research should therefore be mindful of selection when trying to account for differential outcomes between mobile and non-mobile adolescents. |
|---|---|
| ISSN: | 1573-7799 |
| DOI: | 10.1007/s10940-013-9200-7 |