Does a difference make a difference? comparing cross‐national crime indicators *

TLDR: This article investigated the reliability of four widely used cross-national data sets by constructing an error framework that relates types of errors to uses of the data, and found that for studies seeking aggregate descriptions of world crime or analytic explanations of crossnational crime rates, differences in the data sets do not make a difference in the results.

Abstract: This study investigates the question of reliability among four widely used cross-national data sets by constructing an error framework that relates types of errors to uses of the data. The findings indicate that (1) for nation-by-nation point estimation, the four data sets differ by varying degrees, (2) for aggregate point estimation in cross-sectional descriptive and longitudinal descriptive studies, they are statistically similar, and (3) for analytic or explanatory cross-sectional purposes, they yield statistically and substantively similar results. In short, for studies seeking aggregate descriptions of world crime or analytic explanations of cross-national crime rates, differences in the data sets do not make a difference in the results.