U.S. flag

An official website of the United States government, Department of Justice.

NCJRS Virtual Library

The Virtual Library houses over 235,000 criminal justice resources, including all known OJP works.
Click here to search the NCJRS Virtual Library

Sources of Error in Survey Data Used in Criminal Justice Evaluation - An Analysis of Survey Respondents' Reports of 'Fear of Crime'

NCJ Number
81970
Author(s)
W T Bielby; R A Berk
Date Published
1979
Length
39 pages
Annotation
Preliminary analysis examine the measurement characteristics and substantive determinants of 17 measures of 'fear of crime' obtained from the first two waves of a multiwave panel survey of about 400 adults in four large metropolitan areas.
Abstract
The study focused on 17 items that measured five dimensions of fear of crime: (1) limit activities because of crime, (2) perceived increase in crime, (3) fear for safety in areas of the metropolitan region, (4) fear for safety in neighborhood, and (5) perceived likelihood of victimization. Various questions on crime and criminal justice were administered in each wave, and data were also obtained on personal events (health, employment, etc.) which might impact the outcome measures of interest. Demographic characteristics of respondents (age, sex, income, education, etc.) were obtained during the first wave, as well as detailed information on previous contact with crime and the criminal justice system and respondents' perceptions of the incidence of crime in their neighborhoods. A multiple-indicator structural equation model was used to document both the measurement characteristics and the substantive sources of variation in the 17 items. The five components of fear of crime are clearly distinguishable from one another empirically. The short-term stability of each dimension is substantially larger than zero-order observed correlations suggest, even after taking into account intertemporal error correlation as large as .50. The items reported most reliably are those dealing most directly with the personal well-being of the respondent. The substantive equations showed that gender (females express greater fear of crime than males), ethnicity of neighborhood (blacks show greater fear of crime than whites), and city have the most pervasive effects on fear of crime. In contrast, measures of contact with crime and the criminal justice system had effects on fear of crime that were modest at best. The activities of the continuing research are described, and 29 references and 9 notes are provided. A detailed discussion of the model used is appended. (Author summary modified)