ABSTRACT
We have seen in the various chapters of this book that regression models can provide a powerful tool to address certain questions in medical research, in particular, about the “pure” effect of one variable on another taking concurring effects of other covariates into account. If we can, in advance to an analysis, specify completely the regression model we need to address the question of interest (or several models to address several questions of interest), then regression models can demonstrate their power and beauty to their full extent.
