Statistics, truth, and error reduction in sport and exercise sciences

Abstract

Researchers in sport science have to maximise the chances that their results are true and honest. In this paper we described statistical methods for reducing unexplained error variance. These methods increase the probability that researchers correctly identify real effects, and they improve the power of statistical tests. The editor invited us to write the paper as a result of a session on 'truth in sport and exercise science' at the 2004 European College of Sport Science conference (Clermont-Ferrand, France). Cooper contributed to the ANOVA analyses, wrote the section on multiple comparisons and he also helped edited the manuscript.