2026/07/30 by Stephen Douglas
paper · doi:10.1177/00111287261473209
Crime is often concentrated at a small number of places, yet substantial variation exists within and across spatial units. This study examines how land use relates to crime while accounting for parcel-, street-segment-, and neighborhood-level variation. Using 5 years of geocoded crime data from a small U.S. city, parcels are nested within street segments and census block groups and analyzed using multilevel negative binomial models. Results show crime is highly concentrated but unevenly distributed within land use categories, with substantial variation at the parcel and street-segment levels. Supplemental analyses indicate modest spatial clustering at the parcel level. These findings highlight the importance of micro-level variation in understanding crime at place.