NCJ Number
137282
Journal
Journal of Quantitative Criminology Volume: 8 Issue: 2 Dated: (June 1992) Pages: 155-173
Date Published
1992
Length
19 pages
Annotation
This paper examines the common practice in cross-national research on homicide of using crime estimates for a multiyear period based on a smaller number of years than theoretically desired because of erratic data reporting for selected nations.
Abstract
Correlations between hypothetical baseline rates and various alternative estimates are examined to simulate the potential consequences of using temporally incomplete data rather than data for the full multiyear period of interest. The results indicate that this common practice is usually acceptable for certain national samples with regular and consistent data collection on homicides, but it is more problematic with nations that have irregular data collection on this offense. The author advises that any cross-national research that uses estimates of homicide based on temporally incomplete data should include as a variable an explicit indicator of the general level of data reporting. The researcher can then assess the extent to which incomplete estimates are being used for nations that are especially low in offense reporting relative to others in the sample. Various strategies might then be pursued to accommodate potential biases. Analyses might be performed both with and without problematic cases, and the impact on substantive conclusions can then be assessed. Additionally, some type of weighting procedure might be used to assign smaller weights to cases that require homicide estimates based on incomplete data and that exhibit generally low levels of reporting. 7 tables and 37 references