1. A correlation coefficient can be generated by any two interval or ratio variables without distinguishing between an independent variable and a dependent variable; because rXY = rYX, there is no need for such a distinction. Regression, on the other hand, requires explicitly identifying both the independent variable and the dependent variable in the analysis. The purpose of regression is to predict the values of Y using the values of X. Therefore, Y is the main variable and X is the supporting variable; they are not interchangeable.