Against “p-hacking”: Contribution to the discussion on empirical sport science

Büsch D., Strauß B.

Abstract

In general, empirical results are analyzed using statistical methods to examine and discuss differences and interrelationships. Statistical methodological evaluation and the underlying reporting strategy can be described as a technical language that has to be learnt for successful communication between researchers, authors and reviewers; however, empirical science is a constantly changing environment which is why the statistical methods applied have to be refined to adhere to new empirical approaches. This line of thought will be discussed in this article and five essential recommendations will be presented: (a) directionality of hypotheses, (b) confidence intervals, (c) effect sizes, (d) confidence intervals of effect sizes and (e) practical meaningfulness.