U.S. flag

An official website of the United States government, Department of Justice.

NCJRS Virtual Library

The Virtual Library houses over 235,000 criminal justice resources, including all known OJP works.
Click here to search the NCJRS Virtual Library

Aggregation in Data Tables - Implications for Evaluating Criminal Justice Statistics (From Confirmatory and Exploratory Analysis of the Spatio and Temporal Properties of Crime Data, P 22-57, 1981, by Reginald G Golledge and Lawrence J Hubert - See NCJ-78709)

NCJ Number
78710
Author(s)
L J Hubert; R G Golledge; C M Costanzo; T Kenney
Date Published
1981
Length
36 pages
Annotation
Implications of aggregation in data tables for evaluating criminal justice statistics are dicussed.
Abstract
Data collected on a set of objectors (e.g., cities) over a set of attributes (e.g., time points) can be subjected to a variety of aggregation schemes. In the study's first example of two aggregation schemes, a rank transformation was used within each row and on the rank sums for each column. This process ensures that each object or row contributes equally, so that some degree of natural comparability exists between the summary statistics of the two aggregation schemes. To develop more explicit relationships in terms of formulas, however, the transformations used in most of the study were based on obtaining z-scores. Observations within rows were aligned for location and scale. This convention allowed precise connections to be developed between the two aggregation schemes in terms of summary indices and z-statistics. Matrix extensions offer flexibility in defining different relationships among the attributes, but the problem of defining a transformation on the aggregate data matrix makes it difficult to develop precise formulas for connecting the two aggregation schemes. The methods of data aggregation discussed represent both ongoing procedures used in geography for aggregating data and alternatives to those standard procedures. The two-group discussion is concerned with the concordance between two classes, even though the various summary indices were subject to modification by the degree of internal concordance. Tabular data and 15 reference listings are provided, and definitions are appended. (Author summary modified)