Neurons vs. Silicon: Regression Analysis and P values

Abstract

Regression analysis is everywhere in medical statistical inference, and so are its p values. Its omnipresence comes as no real surprise. At their core, regression analysis and health care research are about relationship examination (e.g., does a new therapy for hypertension improve clinical outcomes?) In medicine we want to understand what either causes or cures diseases and we learn about these causes through exploring relationships. Regression analysis is the statistical tool that allows us a quantitative view of these relationships. This chapter’s purpose is not to teach regression analysis’ but, after a brief review, to develop a disciplined approach to p value incorporation and interpretation in its use.