Dealing with Collinearity in Polynomial Regression

Abstract

In this study, we consider three different transformations for reducing collinearity in polynomial regression. In order to compare these
transformations, an extensive simulation study is conducted. The transformations are also applied to empirical data. Based on the
simulation study and real-data applications, it can be concluded that the transformation which yields the values of the independent variable
in the range of [-1,1] is the most effective in reducing collinearity, and allows the fitting of high-degree polynomial to data sets.