Statistical Considerations for the Univariate Analysis of Repeated-Measures Experiments

Abstract

The areas of research within the field of motor learning almost always permit a repeated measures design. The ability to describe the effects which occur over different periods of time is essential for the analysis of experiments which seek to describe the interaction between a given treatment and its effect on learning. Even though researchers in motor learning do not hesitate to set up repeated-measures designs, they often violate basic assumptions of statistical inference which are necessary for valid conclusions. It is therefore the purpose of this article to describe the basic statistical assumptions which underlie analysis of repeated measures and to describe the alternatives available when these assumptions are violated.