BACKGROUND: The present research evaluates the possibility of spatial heterogeneity in the effects on neighborhood crime rates of both traditional demographic indicators-immigrant concentration, racial composition, socioeconomic disadvantage, and residential instability-and a contemporary aspect of housing transition-foreclosure-that has garnered significant attention in recent scholarship. OBJECTIVE: This research advances previous research by explicitly assessing the merits of the typical "global" or "one size fits all" approach that has been applied in most neighborhood studies of demographic context and neighborhood crime rates by juxtaposing it against an alternative strategy-geographically weighted regression (GWR)-that highlights the potentially significant "local" variability in model parameters. We assess the local variation of these relationships for census tracts within the city of Chicago. METHODS: This paper utilizes GWR to test for spatial heterogeneity in the effects of demographic context and other predictors on neighborhood crime rates. We map local parameter estimates and t-values generated from the GWR models to highlight some of the patterns of demographic context observed in our analysis. CONCLUSIONS: GWR results indicate significant variation across Chicago census tracts in the estimates of logged percent black, immigrant concentration, and foreclosure for both robbery and burglary rates. The observed effects of socioeconomic disadvantage on robbery rates and residential stability on burglary rates also are found to vary across local neighborhood clusters in Chicago. Visual inspection of these effects illuminates the importance of supplementing current approaches by "thinking locally" when developing theoretical explanations and empirical models of how demographic context shapes crime rates.
All Science Journal Classification (ASJC) codes