This study examines the spatial relationship between violent crime and residents' perceived neighborhood violence across Chicago's 77 community areas. It asks: what social and environmental factors explain discrepancies between objective crime rates and residents' fear of crime? Using incident-level data from the Chicago Police Department and survey-based perception measures from the Chicago Health Atlas, this study computes a crime perception gap for each community area, defined as the standardized difference between perceived violence and actual crime rates. A three-stage analytical approach is employed: factor analysis to reduce neighborhood health and environment predictors to underlying dimensions, k-means clustering to classify neighborhoods, and spatial autocorrelation analysis using Global and Local Moran's I to assess the geographic structure of crime, perception, and their divergence. Results show that both violent crime rates and perceived neighbor- hood violence show strong and consistent positive spatial autocorrelation across all spatial weight specifications. Iterated principal axis factoring with promax rotation yields three interpretable dimensions whose substantial inter-factor correlations (φ = 0.39–0.56) confirm that oblique rotation is appropriate for this data: Concentrated Disadvantage, Urban Infrastructure Quality, and Environmental Vulnerability. K-means clustering identifies k = 4 as the optimal neighborhood typology: a stable outer-residential cluster, a high-amenity dense urban cluster, a high-deprivation high-crime core, and a moderate-deprivation urban cluster. The perception gap varies meaningfully across these types, with the moderate- deprivation cluster showing the highest positive gap and the high-deprivation core showing a slight negative gap. The crime perception gap does not exhibit statistically significant spatial clustering, and regression models indicate that the three structural factors do not significantly predict variation in the perception gap. These findings contribute to urban criminology by demonstrating that residents' perceptions of violence broadly track the geographic distribution of crime across Chicago while highlighting specific neighborhood contexts, particularly communities where linguistic and informational barriers may limit access to official safety communications, in moderate-deprivation contexts, where meaningful mismatches occur and where broad structural conditions alone appear insufficient to explain the divergence, suggesting a role for more localize informational and contextual mechanisms.
Mia Sowder (Fri,) studied this question.