skip navigation

PUBLICATIONS

Register for Latest Research

Stay Informed
Register with NCJRS to receive NCJRS's biweekly e-newsletter JUSTINFO and additional periodic emails from NCJRS and the NCJRS federal sponsors that highlight the latest research published or sponsored by the Office of Justice Programs.

NCJRS Abstract

The document referenced below is part of the NCJRS Virtual Library collection. To conduct further searches of the collection, visit the Virtual Library. See the Obtain Documents page for direction on how to access resources online, via mail, through interlibrary loans, or in a local library.

 

NCJ Number: 51626 Find in a Library
Title: DESCRIPTIVE DISTORTIONS IN COVARIANCE BASED STATISTICS
Journal: SOCIAL SCIENCE RESEARCH  Volume:5  Dated:(1976)  Pages:107-126
Author(s): R A BERK; M HENNESSY; R MCCLEARY
Corporate Author: Academic Press, Inc
Promotions Manager
United States of America
Date Published: 1976
Page Count: 20
Sponsoring Agency: Academic Press, Inc
San Diego, CA 92101-4495
Type: Statistics
Format: Article
Language: English
Country: United States of America
Annotation: THE PEARSON PRODUCT MOMENT CORRELATION AND THE UNSTANDARDIZED REGRESSION COEFFICIENT ARE EXAMINED TO DETERMINE THEIR IMPACT ON CONSTRUCT VALIDITY IN EMPIRICAL RESEARCH.
Abstract: CONSTRUCT VALIDITY IN EMPIRICAL RESEARCH IS DEFINED AS THE 'FIT BETWEEN EMPIRICAL MEASURES AND THE THEORETICAL CONCEPTS REPRESENTED.' THIS PRESENTATION EXAMINES THE QUALITY OF CONCEPTUALIZATIONS, MEASUREMENTS, AND THEIR CORRESPONDENCE IN THE USE OF SUMMARY STATISTICS. THE DESCRIPTIVE USE OF TWO COVARIANCE-BASED MEASURES IS EXAMINED: THE PEARSON PRODUCT MOMENT CORRELATION AND THE UNSTANDARDIZED REGRESSION COEFFICIENT. USING ILLUSTRATIVE DATA, IT IS ARGUED THAT THE ALGEBRA NECESSARILY INVOLVED IN CORRELATIONAL MEASURES DISPORPORTIONATELY WEIGHTS SCORES FARTHER FROM THE MEAN, DESPITE IMPLICIT ASSUMPTIONS THAT ALL CASES IN THE SAMPLE OR POPULATION ARE EQUALLY IMPORTANT. SINCE REGRESSION COEFFICIENTS AND CORRELATION COEFFICIENTS HAVE THE SAME NUMERATOR, THE REGRESSION COEFFICIENT IS EXAMINED TO SEE IF SIMILAR PROBLEMS EXIST. FROM THIS INVESTIGATION, IT IS CONCLUDED THAT SLOPE MEASURES MAY TYPICALLY HAVE HIGHER CONSTRUCT VALIDITY THAN CORRELATIONAL STATISTICS. IF COVARIANCE CAN BE USEFULLY INTERPRETED AS MOMENTS OF FORCE, IT IS ARGUED THAT AN ALGEBRA WHICH BUILDS ON MORE DEVIANT CASES IS APPROPRIATE. USING THE REGRESSION COEFFICIENT AS A CONSTANT OF PROPORTIONALITY BETWEEN X AND Y LOGICALLY REQUIRES THE MULTIPLICATION PROCESS PRODUCING UNEQUAL WEIGHTING. IT IS CONCLUDED, THEREFORE, THAT FOR REGRESSION COEFFICIENTS THERE IS A GOOD FIT BETWEEN THE ALGEBRA AND THE AGGREGATE PROCESSES DESCRIBED. ILLUSTRATIVE DATA ARE INCLUDED. (RCB)
Index Term(s): Evaluation; Evaluation techniques; Statistical analysis
To cite this abstract, use the following link:
http://www.ncjrs.gov/App/publications/abstract.aspx?ID=51626

*A link to the full-text document is provided whenever possible. For documents not available online, a link to the publisher's website is provided. Tell us how you use the NCJRS Library and Abstracts Database - send us your feedback.