NCJ Number
81110
Date Published
1980
Length
54 pages
Annotation
This paper introduces users of Illinois statewide criminal justice data to aggregation concepts and problems. It applies accumlated knowledge in the literature dealing with aggregation problems to specific examples of problems faced by users of Illinois criminal justice data.
Abstract
The paper explains what it means to aggregate data in certain ways. Two common forms of data aggregation -- forced and chosen aggregation -- are explained using an imaginary data set. The paper also discusses the aggregation problems posed by the three major sources of statewide criminal justice data in Illinois: the Illinois Uniform Crime Reports, the Annual Report of the Administrative Office of the Illinois Courts, and the Inmate Master records of the Illinois Department of Corrections. The aggregation problems that face users of these data involve (1) aggregate levels of analysis that are available at only one level and (2) the availability of only one categorization scheme (one group of variables). Some of the issues in the aggregation literature that are reviewed include standardized and unstandardized measures of the relationships between variables; model specification, including the concepts of unit of analysis and statistical bias; and grouping processes. The report then presents four separate analyses using data obtained from the sources covered in the paper to demonstrate how aggregation of Illinois criminal justice data can affect research results and to show how the concepts emphasized in the literature can be used to understand the effects of aggregation on analyses. The report concludes that the key to understanding aggregation problems is to determine the effect that grouping processes have on models and measures of relationships between variables. The task for the researcher is to obtain as accurate an understanding as possible about grouping processes and make adjustments in analysis to correct for aggregation effects. Steps to be taken to confront aggregation problems are suggested. Endnotes, about 35 references, maps, and a list of offense categories are appended. Tables and graphs are included.