Que signifie la significativité statistique ? L'apport de la taille d'effet et de la puissance statistique

Abstract

Power analysis and effect size are two statistical concepts, which are not widely used in sport sciences in spite of their relevance. This article shows, through two distinct studies, the numerous advantages of using these concepts. Using power analysis considerations, it becomes clear in the first study that one could not reach statistical significance due to the small sizes of the two samples. In this case, it seems better to rely on effect size to conclude because the effect size shows the actual importance of a phenomenon, emphasizing its scientific significance beyond the usual statistical (non) significance. on the contrary, the results of the second study are statistically significant because of the huge size of the sample but its explanatory power, as estimated by the observed effect size, is rather limited. These results have thus to be pondered. Effect size and power analysis are also very useful in experimental planning in order to determine adequate sizes for the samples to collect. Finally, the concept of effect size can be used to compare several studies in a meta-analysis.© De Boeck Université.