Null hypothesis significance testing and its alternatives: A revision of some criticisms and their relevance for the medical sciences

Abstract

Abstract
Arguably, null hypothesis significance testing constitutes the most widely-applied tool for the evaluation of scientific hypotheses and decision making, not only in medical and health sciences but also in biology, psychology, political sciences, and other social sciences. However, since its introduction by Sir Ronald Fisher in 1925, the method has been the center of a heated debate where various criticisms related to conceptual and interpretational problems have been counterattacked by advocates of the technique who argue that the method, if used correctly, does have its place in the statistician’s toolbox. In this article, I revise some of the most pertinent criticisms of this debate, illustrate their relevance for research in the field of medical and health sciences and discuss some recommendations that have been proposed as a remedy or an alternative for significance testing.