A Simple Multiple Linear Regression Test for Differential Effects of a Given Independent Variable on Several Dependent Measures

Abstract

Multiple linear regression may be used to determine whether an independent variable of interest has a differential effect on two or more dependent variables. The initial step involves the separate standardization of each dependent variable. The values of the standardized dependent variable are pooled and treated for purposes of the analysis as constituting a single dependent variable. A within subjects independent variable is created and the levels of the variable are used to denote the different dependent variables. The data are analyzed with a split-plot analysis of variance for which the independent variable of interest is the between groups factor and the independent variable which distinguishes the dependent variables is the within subjects factor. The test of the interaction of these two factors provides a statistical determination of whether the independent variable of interest has a differential effect on the two or more dependent variables.