Qualitative matching of alcohol outlets using virtual geographic imaging technology

In matched-pair analyses, study groups are matched with comparison groups on certain criteria to control for extraneous factors that may confound results (e.g., socioeconomic status).  In some cases, available archival data sources are not sufficient to make a good match. To address this problem, we developed a new methodology using an innovative data source, virtual geographic imaging (VGI) technology, to identify appropriate comparisons. Our broader study evaluates effects of local malt liquor policies on urban crime. Beginning with three pilot cities, we will compare crime rates around alcohol outlets that are subject to local restrictions on high-alcohol malt liquor sales, with crime rates around similar outlets that are not subject to the restrictions. To identify comparison outlets, we first matched on neighborhood demographics and alcohol outlet densities. We then conducted VGI observations on the five closest matches to inform our final selection. Our VGI instrument measures store type (e.g., grocery versus liquor), opportunity to loiter (e.g., adjacent parking lot), land use (e.g., commercial versus residential), infrastructure condition (e.g., boarded windows) and social disorder (e.g., adjacent adult bookstore). We confirmed the reliability of our instrument by comparing VGI results with results from in-person observations in all three cities. We assessed inter-rater reliability in a sample of 25 outlets and found good concordance. We concluded that VGI observations can be a reliable and useful tool for matching target and comparison groups when archival data sources are inadequate and in-person observations are not feasible. We will describe VGI measurement methodology and potential applications.