RATIO VARIABLES AND PATH MODELS (FROM STRUCTURAL EQUATION MODELS IN THE SOCIAL SCIENCES, 1973, BY ARTHUR S GOLDBERGER AND OTIS D DUNCAN - SEE NCJ-52004)

• NCJ Number: 52014
• Author(s): K SCHUESSLER
• Date Published: 1973
• Length: 28 pages
• Sponsoring Agency: Russell Sage Foundation
• Corporate Author: Seminar Press, Inc
• Publication Format: Document
• Publication Type: Report (Study/Research)
• Language: English
• Country: United States of America
• Note: REVISION OF A PAPER PRESENTED AT A CONFERENCE COSPONSORED BY THE SOCIAL SCIENCE RESEARCH COUNCIL AND THE SOCIAL SYSTEMS RESEARCH INSTITUTE OF THE UNIVERSITY OF WISCONSIN, MADISON, NOV 12-16, 1970

Annotation

COMPLICATIONS IN PATH ANALYSIS RESULTING FROM THE PECULIARITIES OF RATIO VARIABLES ARE EXPLORED, AND SOCIOLOGICAL EXAMPLES OF PATH ANALYSIS AND RELATED CORRELATIONAL ANALYSES ARE CITED.

Abstract

A RATIO VARIABLE IS ONE VARIABLE DIVIDED BY ANOTHER, WHILE A VARIABLE DIVIDED INTO A CONSTANT IS AN INVERSION OF THAT VARIABLE. IN SOCIOLOGY, THE DEFLATION OF ONE VARIABLE BY ANOTHER IS USUALLY UNDERTAKEN TO FACILITATE COMPARISONS AMONG UNITS, WHEREAS A NONLINEAR TRANSFORMATION OF A GIVEN VARIABLE GENERALLY IS UNDERTAKEN TO MEET ASSUMPTIONS OF A STATISTICAL PROCEDURE. PATH ANALYSIS CULMINATES IN COEFFICIENTS THAT MEASURE CAUSAL EFFECTS. THE WORK OF PEARSON, TAYLOR, AND OTHERS ON RATIO VARIABLES IS REVIEWED IN AN ATTEMPT TO DISCERN SOME OF THE PROBLEMS ENCOUNTERED IN THE USE OF RATIO VARIABLES. THREE GUIDELINES TO GUARD AGAINST THESE PROBLEMS ARE STATED: (1) THE COMPOSITION OF VARIABLES WHOSE RELATIONS ARE IN QUESTION SHOULD BE SPECIFIED CLEARLY; (2) THE INDISCRIMINATE USE OF APPROXIMATE FORMULAS FOR THE CORRELATION OF RATIOS SHOULD BE AVOIDED; AND (3) TERMS IN A COMPOUND QUANTITY (E.G., RATIO OR PRODUCT) SHOULD BE ANALYZED STATISTICALLY FOR THEIR RESPECTIVE CONTRIBUTIONS TO VARIATIONS IN THAT QUANTITY. EXAMPLES OF PATH ANALYSIS INVOLVING RATIO VARIABLES ARE DESCRIBED, AND CONSIDERATION IS GIVEN TO THE DISTRIBUTION OF RATIO VARIABLES, LEVEL OF MEASUREMENT, AND MEASUREMENT ERROR. SUPPORTING EQUATIONS ARE GIVEN. (DEP)