Examining Measurements of Fear of Crime Across Surveys: Content, Cognitive Processes, and Psychometric Properties

Purpose: To enhance the understanding of different survey questions and the opportunities they provide to measure fear of crime as a latent trait, examining both crime-type and consequence items.Methods: Items from five surveys used in the Stockholm Region were coded into content areas. Cognitive interviews (n = 19) were analysed for response processes. Rasch analyses were conducted on the Swedish Crime Survey (n = 4,264) and the City of Stockholm Crime Survey (n = 7,333).Results: The surveys covered overlapping topics but used distinct phrasings. Response comprehension was generally adequate, yet some terms were interpreted differently. Respondents found it difficult to isolate fear of crime from other motives in the consequence items. There was no clear advantage to either frequency- or consequence-framed items, whereas intensity stems imposed a lower response burden. Psychometrically, unidimensional solutions were hard to achieve for either crime type or consequence item sets. The evidence suggests heterogeneity across population groups.Conclusions: Current practice involves both strengths and weaknesses. This study suggests paths toward enhanced measurement of the fear of crime. Measuring fear of crime as a single, invariant latent trait is far from straightforward. The heterogeneity of responses across population groups indicates that different groups evaluate their fear of crime types and consequences differently, challenging the assumption of unidimensionality. However, the cognitive interviews suggest ways to clarify item wording and reduce respondent burden.