EMPIRICAL CRIMINOLOGY - INTERPRETING RESULTS DERIVED FROM AGGREGATE DATA

• NCJ Number: 61522
• Journal: Journal of Research in Crime and Delinquency Volume: 16 Issue: 2 Dated: (JULY 1979) Pages: 294-306
• Author(s): T ORSAGH
• Date Published: 1979
• Length: 13 pages
• Corporate Author: National Council on Crime and Delinquency
• Sale Source: UMI Dissertation Services; Institute for Scientific Information
• Publication Type: Statistics
• Language: English
• Country: United States of America

Annotation

AN ECONOMIST EXAMINES PROBLEMS OF SUBJECTIVITY IN CRIMINAL JUSTICE RESEARCH AND ARGUES THAT USING AGGREGATE DATA PRODUCES INFERIOR STATISTICAL DESIGN AND RESULTS.

Abstract

STATISTICAL RESULTS BASED ON AGGREGATE DATA ARE, IN PART, ARTIFACTS OF DISCRETIONARY RESEARCH ACTIVITY. EMPIRICAL CRIMINAL JUSTICE CONCEPTS, SUCH AS THE HYPOTHESIS THAT SANCTIONS ACT AS A DETERRENT ON THE HOMICIDE RATE, ARE SOMETIMES DIFFICULT TO TEST, COSTLY, INEXACT, OR TOO NUMEROUS FOR SUCCESSFUL COMPARATIVE STATISTICAL ANALYSIS. STATISTICAL RESULTS ARE ALSO SENSITIVE TO THE SELECTION OF CONTROL VARIABLES, THE STRUCTURE OF THE EXPLANATORY MODEL (SINGLE VERSUS MULTIPLE EQUATION SYSTEMS) AND THE MODE OF ESTIMATION (ORDINARY LEAST SQUARES VERSUS A SIMULTANEOUS EQUATION ESTIMATION TECHNIQUE). FURTHER, AGGREGATION CAN PRODUCE HETEROSKEDASTIC DISTURBANCES, HETEROSKEDASTICITY PRODUCES BIASED STATISTICAL RESULTS, AND THERE IS NO GENERAL PROCEDURE FOR ELIMINATING AGGREGATION BIAS. IMPROPER CORRECTION FOR BIAS CAN PRODUCE MISLEADING RESULTS. JUDGMENT IS NEEDED TO INTERPRET VARIABLE AND OFTEN CONFLICTING RESEARCH RESULTS, AND A FORMAL METASTISTICAL METHODOLOGY SHOULD BE DEVELOPED TO TAKE INTO ACCOUNT THE COMPOSITE EFFECT OF CHANCE VARIATION, STATISTICAL DESIGN, AND HUMAN DISCRETION. FOOTNOTES AND REFERENCES ARE PROVIDED. (AUTHOR ABSTRACT MODIFIED--AOP)