NCJ Number
206509
Journal
Homicide Studies Volume: 8 Issue: 3 Dated: August 2004 Pages: 267-310
Date Published
August 2004
Length
44 pages
Annotation
This study illustrates how missing data problems in comparative homicide research may be addressed by composite measures.
Abstract
Due to the fact that the number of countries in the world is small compared with other datasets, missing data present a major problem in cross-national homicide research. Indexes such as the International Homicide Index, which takes advantage of multiple data sources on homicide, not only maximize the number of countries represented by the indicator, but also produce a more parsimonious and robust measure of lethal violence levels across the nations. International homicide data, like national level data, come from either police statistics or from vital statistics. Theoretically, if police and mortality statistics tap the same phenomenon, there should be perfect agreement between mortality and police. Clearly, that is often not the case. Two reasons for the lack of perfect agreement between police data and health data are: differences due to definitions of cases and ambiguities and failure of procedures that occur when reported deaths are placed in the wrong category. At most recent count, there are 193 independent states in the world, but world-regional analysis of violence typically work with a much more limited number of countries than all 193. There is no doubt that the missing data continue to provide a major problem in cross-national violence research. Although it is important to choose the data that maximize completeness, no single data source is complete or even close to complete. Homicide is the most frequently examined crime in cross-national studies of crime. There are data that homicide data are among the most reliable and valid cross-national indicators. Like other crime indicators, international homicide statistics suffer from definitional ambiguities and classification problems. The best way to deal with these problems is through a strong validation process. It is not surprising that validation is, at present, either weak or nonexistent for international homicide statistics. The process of validation enables the expert group to become more confident in the validity of the results produced. Cross-national research has reached a new stage with increased opportunities for improving the use of available data. Tables, notes, references