Alcohol Outlet Density and Violent Crime in Chicago, IL

Introduction: Over the past two decades, multiple studies have documented the relation between density of retail alcohol outlets and violent crime. Less well established is whether this association varies as a function of neighborhood characteristics. The purpose of this study is to assess the degree to which the relation between density of retail alcohol outlets and violent crime is moderated by neighborhood structural and social characteristics.

Methods: For 2011, data on incidents of violent crime and locations of retail alcohol outlets in Chicago were aggregated at the level of the census tract. Two categories of crime were considered: violent crimes (i.e., homicide, aggravated assault/battery, and criminal sexual assault) and shootings (i.e., aggravated battery with a firearm). Two types of retail alcohol outlets were aggregated separately: package stores and bars. In addition, residential population and concentrated disadvantage (a composite of rates of poverty, unemployment, renter-occupied households, and female headed households) were included as control variables. In a subset of 30 census tracts, measures of neighborhood social processes (e.g., informal social control, neighborhood cohesion, and neighborhood norms) were merged with the administrative data to test moderation by social processes.

Results: In separate analyses, violent crimes and shootings were regressed upon residential population, concentrated disadvantage, number of bars, number of package stores and the interactions between bars/package stores and concentrated disadvantage. In the model predicting all violent crime incidents, neither the main effect for number of bars nor the interaction between number of bars and concentrated disadvantage were significantly related to the number of violent crime incidents. For package stores, the main effect (β=1.47; t(790)=6.50, p<.001) and interaction effect (β=1.67; t(790)=7.26, p<.001) were significant such that more package stores were associated with higher incidence of violent crime in areas of moderate to high concentrated disadvantage and unrelated to incidence of violent crime in areas of low concentrated disadvantage. The same pattern of results was found in the model predicting shooting incidents. Overall, the models accounted for 55% of the variance in violent crimes and 40% of the variance in shootings. Additional analyses will explore the moderating effects of neighborhood social processes.

Conclusions: The significant moderation of the relation between density of package stores and violence by concentrated disadvantage suggests that the impact of retail alcohol outlets is not uniform across neighborhoods. Future research is needed to understand these relations and the potential intervention and policy implications.