Linking Places of Worship to Crime

The U.S. Census Bureau provides data on various geographic/spatial units. We draw on block groups specifically to link places of worship to crime in place, primarily because block groups have been designed to be homogenous on a range of sociodemographic characteristics, including income and poverty, educational attainment, household structure, age, and length of residence [37 (link), 38 ]. Thus, our selection of block groups as our units of analysis is consistent with previous empirical work that has examined “neighborhood effects” on a range of outcomes, such as ethnic and racial segregation [e.g. 39 (link)], social networks [e.g. 40 (link)], walkability and health [41 (link)], gentrification [e.g. 42 (link)], and crime [43 (link)], to name a few.
The present study involves secondary data analysis of block groups from publicly existing data, and therefore did not require institutional review board approval. For our analysis, we estimate crime models using a sample of 449 block groups (out of the 450 in DC); one block group has been dropped because it is missing necessary information from the U.S. Census American Community Survey (ACS). We cannot use constituent tract information as a substitute for missing block group information, because for this block group, the tract and block group boundaries are exactly the same. We suspect missing data for some variables is attributed to the fact that this area largely encompasses Georgetown University and its affiliated facilities.