Diagnosing and Mitigating Multicollinearity in Moderated Multiple Regression

Abstract

<p>This paper primarily explores the challenges associated with Moderated Multiple Regression Models, particularly how a moderator variable (m) influences the direction or strength of the relationship between an independent variable (x) and a dependent variable (Y). A significant issue arises when there is a high correlation between the independent variable and the moderator, leading to severe multicollinearity. That complicates the accurate estimation of the independent variables’ effects on the dependent variable (Myers, 1990). We develop five moderated multiple regression models with purpose of mitigating the multicollinearity in the analysis. Our empirical findings indicate that three of them perform good tested by the variance inflation factor and condition index. We finally suggest a process of standardizing both independent variable and moderator and taking the cross multiplication by those two standardized variables before conducting moderated multiple regression analysis.</p> <p> </p>