Network spillovers and neighborhood crime: A computational statistics analysis of employment-based networks of neighborhoods.

Research on communities and crime has predominantly focused on social conditions within an area or in its immediate proximity. However, a growing body of research shows that people often travel to areas away from home, contributing to connections between places. A few studies highlight the criminological implications of such connections, focusing on important but rare ties like co-offending or gang conflicts. The current study extends this idea by analyzing more common ties based on commuting across Chicago communities. It integrates standard criminological methods with machine learning and computational statistics approaches to investigate the extent to which neighborhood crime depends on the disadvantage of areas connected to it through commuting. The findings suggest that connected communities can influence each other from a distance and that connectivity to less disadvantaged work hubs may decrease local crime-with implications for advancing knowledge on the relational ecology of crime, social isolation, and ecological networks.