NCJ Number
124041
Journal
Criminology Volume: 28 Issue: 2 Dated: (February 1990) Pages: 153-182
Date Published
1990
Length
30 pages
Annotation
This study examines the reliability of four major sources of cross-national crime data collected by INTERPOL, the Comparative Crime Data File (CCDF), the United Nations, and the World Health Organization (WHO), by employing a framework that relates error to uses of the data.
Abstract
The three dimensions of the framework distinguish between descriptive and analytic uses of the data, classify errors using cross-sectional and longitudinal studies, and distinguishes between sampling and measurement errors. After defining the group of nations through a two-step process, data for violent crime and theft were analyzed by making six comparisons among the available data sets for two time periods, 1960-1972 and 1975-1980 and between descriptive and analytic uses of the data. Comparisons of crime rates for individual nations across data sets result in different findings that can be substantial and misleading for a particular nation. Comparisons of ranking of nations' crime rates within data sets will produce differences across data sets and therefore cause error. However, the data sets produce consistent results when used to assess the direction of change in crime over time and across nations. Finally, regardless of the data set employed, summary statistics used to describe the entire sample will yield similar results. The unreliability of existing data for cross-sectional analyses indicates the need for data system improvement. Currently, small-scale victimization surveys offer the best chance for uniform and comparable data for descriptive purposes. 6 tables, 22 notes, 41 references.