Modeling Urban Neighborhood Violence: The Systemic Model and Variable Effects of Social Structure

Abstract

The systemic model of crime illustrates how neighborhood social structure influences social networks needed for the regulation of crime. This study examines whether those structural influences are stable or variable across a distribution of neighborhoods on violent crime. Study data are derived from the National Neighborhood Crime Study, yielding crime and structural data on a total of 6,927 census tracts within 69 U.S. cities. Quantile regression is used to model structural and spatially lagged violence effects on neighborhood violence. Results demonstrate that the influence of structural disadvantage decreases as neighborhood violence increases. Concomitantly, nearby violence, residential instability, and vacant housing effects become more salient as local violence increases. Local racial composition conditions these effects as well. Study findings suggest that local public safety interventions should be developed relative to a neighborhood’s ecological location on violent crime.