Constructing boundaries from buffers: a critical examination of the impact of drinking establishments on crime

Date
2022
Journal Title
Journal ISSN
Volume Title
Publisher
University of Delaware
Abstract
OBJECTIVES: Advancements in technology have increased the capability of researchers to investigate the relationship between crime and place (Andresen & Malleson, 2013; Hunt, 2019). One area of scholarship that has received considerable attention concerns the relationship between crime patterns and alcohol related establishments (Groff & Lockwood, 2014; Wheeler, 2019). However, the majority of research in this area have relied on pre-existing census boundaries that do not align with the objectives of investigations analyzing the impact of facilities on crime. The purpose of this dissertation is to examine how different methods for creating new geographic units through merging buffer zones and census block groups can help enhance our understanding of the relationship between drinking establishments and crime. ☐ METHODS: The primary data used for this project were Delaware police calls for service. Location quotient analyses revealed that property and violent crimes were more densely concentrated within 1600-foot buffer zones around drinking establishments relative to the crime density within the Wilmington Metropolitan area overall. Buffer zones and census block groups were merged using two approaches (i.e., the union and update methods) to determine if results from negative binomial models varied depending on the unit type. As an additional check on the robustness of results, models varied in terms of the operationalization of the dependent variable. ☐ RESULTS: Eight negative binomial models were examined to explore the impact of being part of a drinking establishment buffer zone on crime rates within the newly created geographic regions. Regardless of geographic unit type, operationalization of the dependent variable, and crime type analyzed, being part of a drinking establishment buffer zone was positively associated with crime rates. However, variation in models concerning other significant predictors demonstrate that the modifiable areal unit problem is an issue for this study. ☐ CONCLUSION: While the results for variables such as buffer zone status and the level of crime counts in neighboring units appeared to be robust predictors across methods, the differences in terms of significance, magnitude, and direction between the coefficients of other socially relevant variables reinforce that there should be caution when deciding on the unit of analysis and operationalization of dependent variables.
Description
Keywords
Delaware police calls, Delaware, Merging buffer zones, Drinking establishments/crime dyad
Citation