Usefulness of Correlation Analysis

Abstract

The measure of correlation coefficient (r or R) provides information on closeness of two variables. Irrespective of non-linear correlation, this paper mainly considers the linear correlation analysis, as it is most likely applied in social science studies. Explicitly, the purpose of carrying out correlation analysis is almost the same in quantitative analytical studies, thus becoming useful to explore the association between independent and dependent variables. This paper, as an extension, attempts additionally to explain the usefulness of linear correlation coefficient between two variables in the context of identifying the level of multicollinearity and mediating/moderating status of independent variables in a model. This paper also demonstrates how the level of multicollinearity can be explored by using correlation coefficient of two independent variables in a regression model.

Keywords: correlation, multicollinearity, Variance Inflation factor, VIF

JEL Classification: C12, C18, C20, C30