Connected Communities: Social Capital and Violence Risk in the Context of Big Data and Mobility Flows

• In Event: Morbidity and Mortality: Open Call Paper Session

Abstract

This study aims to advance social disorganization and routine activities theories, typically focused on dynamics within a local community, with new insights from ecological network and social capital theories by examining the extent to which people's spatial-network exposures to places where they engage in different routine activities outside of their home areas can significantly impact lethal violence in their local communities. We expand on a growing body of evidence on the role of extra-local exposure and we test hypotheses longitudinally and using a virtually complete population mobility data for the entire country for over a decade. The study combines a variety of big, complex, and open data sets, including population-based yearly data for counties across the country from the Census-Bureau’s Longitudinal Employer Household Dynamics, with more standard demographic data. Additional analyses compare results with county models that include GPS data from cellphones and other digital devices.

Multilevel panel models showed that, above and beyond spatial spillovers, mobility network spillovers strongly predicted later levels of local homicide, measured yearly for over a decade. The findings are consistent with expectations that population mobility flows play a role in creating networks of information and resource flows that explain the transmission of higher or lower lethal violence risk from one place to another over short and large distances alike. The findings advance on William Wilson's classic insights on spatial and institutional isolation and expand them by showing evidence of a) amplified violence as a result of network connectivity to other violent places, as well b) the potential of connectivity to low-violence places to bring an influx of new social capital that can buffer homicide risk locally. These findings hold across the country but are particularly salient for previously vulnerable communities that were previously thought to be spatially isolated.

Authors

• Corina Graif, Pennsylvania State University

• Christopher H. Seto, Purdue University