NIJ Innovations in Measuring Community Perceptions Challenge Winner: Duddon Research

This entry proposes measuring fear of crime by using a proxy measure comprising certain neighborhood and housing characteristics that local governments could compile from administrative records and publicly available sources. To test the proposed approach, the entry operationalizes the fear of crime construct by using residents’ perceptions of crime in their community, as reported in responses to questions from the U.S. Census Bureau’s biennial American Housing Survey. The entry then uses penalized regression models to determine which housing-related characteristics best predict whether residents agree that their neighborhoods have high levels of serious crime. Results using all available survey years indicate that nine distinct neighborhood features can be used to accurately predict residents’ perceptions of crime. Because local governments can readily measure these characteristics, the proxy measure could be used at frequent intervals and at different levels of geography. The entry illustrates how an existing national dataset may be used in a novel way to measure community-level fear of crime.

The lone team member for this submission was Shana M. Judge.

This entry was one of two first prize winners in the category 2, fear of crime, construct competition.