Estimation versus falsification approaches in sport and exercise science

Michael Wilkinson, Edward M. Winter

Research output: Contribution to journalArticlepeer-review

5 Citations (Scopus)
10 Downloads (Pure)

Abstract

There has been a recent resurgence in debate about methods for statistical inference in science. The debate addresses statistical concepts and their impact on the value and meaning of analyses’ outcomes. In contrast, philosophical underpinnings of approaches and the extent to which analytical tools match philosophical goals of the scientific method have received less attention. This short piece considers application of the scientific method to “what-is-the-influence-of x-on-y” type questions characteristic of sport and exercise science. We consider applications and interpretations of estimation versus falsification based statistical approaches and their value in addressing how much x influences y, and in measurement error and method agreement settings. We compare estimation using magnitude based inference (MBI) with falsification using null hypothesis significance testing (NHST), and highlight the limited value both of falsification and NHST to address problems in sport and exercise science. We recommend adopting an estimation approach, expressing the uncertainty of effects of x on y, and their practical/clinical value against pre-determined effect magnitudes using MBI.
Original languageEnglish
Pages (from-to)3-4
JournalJournal of Sports Sciences
Volume37
Issue number1
Early online date22 May 2018
DOIs
Publication statusPublished - 2019

Fingerprint Dive into the research topics of 'Estimation versus falsification approaches in sport and exercise science'. Together they form a unique fingerprint.

Cite this