Since in multiple linear regression, the coefficient vector estimated by Ordinary Least Square (OLS) and Ridge Regression (RR) are only estimations of the complete effect that a set of explanatory variables has over a response variable, in this paper we propose a general standard process, composed of steps to perform the decomposition of a response variable that is needed to obtain, in separate form, the direct effect, correlation effect between any pair of variables, quadratic effect and so on, that a set of explanatory variable has an over response variable in multiple linear regression.