Commonality Analysis of Multivariate Measures of Association

• In Event: Adaptations of GLM for Advanced Data Analysis

Abstract

This paper presents a statistical software solution to conduct commonality analysis on two multivariate measures of association recommended in the literature to support canonical correlation (i.e., P_YX^2, R_XY^2). Our contribution fills a void in the canonical correlation methodological literature and is consistent with LeBreton and Tondandel (2008) and Azen and Budescu (2006) who offered techniques for relative weights and dominance analysis on multivariate measures of association, respectively. Using a heuristic example, previously published, we demonstrate that the commonality analysis of P_YX^2 can be derived from the commonality analysis of the canonical variates but not for R_XY^2.

Authors

• Kim Nimon, The University of Texas - Tyler

• Linda Reichwein Zientek, Sam Houston State University