Recidivism in Context: A Meta-Analysis of Neighborhood Concentrated Disadvantage and Repeat Offending

Background: Most justice-involved persons will one day recidivate. Research to date has identified a number of individual-level risk factors for recidivism. More recently, scholars have brought a person-in-environment framework to the study of recidivism, directing attention to macro-level risk factors. Of macro-level factors, most attention has been afforded to neighborhood concentrated disadvantage, a multidimensional factor representing the convergence of poverty, employment, other socioeconomic indicators, and often racial composition. While theoretically connected, empirical evidence on the relationship between concentrated disadvantage and recidivism is mixed. This paper meta-analytically synthesizes existing research on the concentrated disadvantage–recidivism relationship. In doing so, we answer the following questions: (1) to what degree, if at all, does concentrated disadvantage predict recidivism, after adjustments are made for individual-level risk factors? And, (2) does the effect of concentrated disadvantage on recidivism depend on sample or study characteristics?

Methods: We conducted a systematic review and meta-analysis of studies, following the guidelines set forth in Preferred Reporting Items for the Systematic Review and Meta-Analysis Protocols (Moher et al., 2015). Eligible studies (1) were conducted in the United States from 1983 to 2017; (2) used concentrated disadvantage as a predictor variable; (3) had recidivism as an outcome variable; (4) had a follow-up period of six months or greater; (5) used a quantitative design; and (6) listed an odds or log odds ratio for the effect of concentrated disadvantage. Included studies were reviewed, coded, and assessed for quality by two researchers. We used multilevel, random effects models to calculate the pooled effect of concentrated disadvantage on recidivism (k = 32), and tested sample and study descriptors as effect moderators to assess sources of potential variance in effects.

Results: We found the pooled effect of concentrated disadvantage on recidivism was non-significant, after studies adjusted for individual-level risk factors (log OR=0.03, p = 0.07, k = 32). However, we also found effect variation across studies was beyond that of sampling error (Cochran’s Q = 146.47, p < 0.01; I² = 82.63). The concentrated disadvantage—recidivism effect was significantly moderated by sample age (i.e. juvenile populations) and recidivism type (i.e. rearrest vs. reconvictions), but not sample gender (i.e., percent female) or race (i.e., percent White). That is, concentrated disadvantage has a statistically significant effect on recidivism for youth (log OR = 0.09, p = 0.02, k = 14) and rearrests (log OR = 0.08, p = 0.002; k = 17), but not adults and reconvictions.

Conclusion and Implications: Together, the results of this review indicate that concentrated disadvantage does not have a statistically significant effect on recidivism after individual-level risk factors are accounted for. However, we also find that the effect of concentrated disadvantage on recidivism is conditional on sample age and how recidivism is operationalized. Thus, a person-in-environment lens should not be dismissed in recidivism research. We conclude with a discussion of key lessons for future research in this area.