Measuring crime through victimization: Some methodological lessons from the ICVS

Abstract

The objective of this study is to demonstrate that there are design errors in the International Crime Victims Survey (ICVS) that lead to biases in the estimates of victimization rates. This will be verified, firstly, by comparing the crime rate derived from the ICVS microdata with that based on the reference population and, secondly, examining the consequences of two ICVS methodological decisions: the use of individual weights and the re-weighting procedure in accordance with household size. These sample design failures, which generate distorted crime rankings, call for the need to make corrections both in the subsequent waves of the ICVS and in the final design of a future European survey.