Measurement problems in official and survey crime rates

Abstract This paper analyzes sources of error in the two major methods we use to measure crime in America—official police statistics and victimization surveys. The two produce quite different pictures of the volume and distribution of crime, but it is not clear that this is because victim-based statistics are “accurate.” Each measurement procedure has its characteristic errors, some of which it shares with the other. Comparisons of official and survey data on crime are helpful in revealing the dimensions of these error terms, and they point out the analyses which must be conducted if we are to specify their exact proportions.