Reliability and stability of the standard fear of crime indicator in a national panel over 14 years

Abstract

"Reducing citizens' fear of crime is a salient policy topic in Western societies. For the measurement of fear of crime, variations of a single item measure (Is there any area near where you live that is, within a mile where you would be afraid to walk alone at night?) have been used in hundreds of surveys for half a century. Despite a long and on-going discussion on the doubtful usefulness of this standard indicator, no estimate for the reliability of this item is available. Using panel data from the British Household Panel Survey in combination with the UK Household Longitudinal Survey, reliability and stability of the standard fear of crime indicator are estimated using quasi-Markov simplex models for the rst time. The model shows estimated reliabilities of about .67 for the standard indicator. This result is compared with results for other single item measures using the same data set. Finally, comparisons with reliability recommendations and recent reviews of survey measures are reported." (author's abstract)