VERIFICATION OF THE GLUECK PREDICTION TABLE BY MATHEMATICAL STATISTICS FOLLOWING A COMPUTERIZED PROCEDURE OF DISCRIMINANT FUNCTION ANALYSIS

• NCJ Number: 5296
• Journal: JOURNAL OF CRIMINAL LAW, CRIMINOLOGY AND POLICE SCIENCE Volume: 6 Issue: 2 Dated: (JUNE 1970) Pages: 229-234
• Author(s): R A LABRIE
• Date Published: 1970
• Length: 6 pages
• Corporate Author: Williams and Wilkins Co
• Publication Format: Article
• Publication Type: Statistics

Annotation

EVALUATION OF A STATISTICAL METHOD FOR PREDICTING JUVENILE DELINQUENCY, DEVELOPED BY SHELDON AND ELEANOR GLUECK.

Abstract

A SERIES OF ANALYTIC MULTIVARIATE ANALYSES OF DATA PUBLISHED BY THE GLUECKS IN UNRAVELING JUVENILE DELINQUENCY DIRECTED TO ANSWERING SEVERAL CRITICISMS OF THE GLUECK PREDICTION DEVICES ARE PRESENTED. THE FIRST ANALYSIS INDICATES THAT THE MISSING OBSERVATIONS IN THE GLUECK DATA ARE TRULY RANDOM AND DO NOT INFLUENCE THE FINDINGS. SEVERAL REASONS ARE PROPOSED FOR THE EFFICACY OF USING EQUAL NUMBERS OF DELINQUENTS AND NONDELINQUENTS IN THE BASIC RESEARCH. USING THE DATA SAMPLES PRODUCED FROM APPLYING FOUR CLASSICAL METHODS OF HANDLING RANDOMLY MISSING OBSERVATIONS, A STEPWISE DISCRIMINANT FUNCTION ANALYSIS ESTABLISHED THE BEST ANALYTICALLY DERIVED SET OF PREDICTORS OF DELINQUENCY. THE PREDICTIVE ABILITY OF THE ANALYTICALLY DERIVED SCHEDULE WAS FOUND TO BE ALMOST EQUAL TO THAT OF THE BEST PREDICTIVE DEVICE OF THE GLUECKS. THE PAPER CONCLUDES THAT THE PREDICTIVE ABILITY AND THE ITEMS SELECTED BY THE GLUECKS FOR THEIR PREDICTIVE DEVICES ARE THOSE ARRIVED AT BY STRICT ANALYTIC MULTIVARIATE TECHNIQUES AND REPRESENT FINDINGS THAT CANNOT BE FAULTED BY CRITICISMS OF THE GLUECK METHODOLOGY. (AUTHOR ABSTRACT)