Neighborhood level predictors of rape: A novel spatial regression approach

Despite decades of research examining the relationship between space and crime, there is a paucity investigating space and rape specifically. This study fills this gap by exploring the spatial clustering and neighborhood-level predictors of rape measured at the census tract level via geographically weighted regression (GWR) in Cleveland, Ohio (U.S.) over two years. In a novel application of bias-corrected GWR, results reveal two high-risk areas: the downtown business district and the economically and racially marginalized east side. By exploring spatial predictors of rape in two ways (overall frequency and per 500 women), we examine how the space is primarily used—to work, visit, or reside. Key predictors include: percent White, median household income, total population, and percent vacant buildings. However, these predictors are not uniform across the city, with some having larger, inverse, or non-significant effects depending on the neighborhood. Study's methodological advances include applying bias corrections to estimates from popular spatial data and allowing predictors to vary by tract (GWR), highlighting that rape predictors function differently in different areas. Findings provide insights into high-risk areas, spatial predictors of rape, and how these vary by tract, offering guidance on modifying the built environment to help reduce or prevent rape.