Alternativas de análisis estadístico en los diseños de medidas repetidas

Abstract

Approaches to the stastistical analysis of repeated measures designs. The analysis of variance is often used to assess treatment effects in repeated measures designs. However, this approach is sensitive to violations of the sphericity. This paper illustrates several statistical procedures for repeated measures designs and multisample repeated measures designs, including several adjusted F statistics, multivariate analysis of variance, General Approximate Test, Improved General Approximate Test, Welch-James procedure, the Bayes approach and mixed model analysis. These tests are discussed with respect to their robustness to violation of assumptions for main as well as interaction effects when designs are either balanced or unbalanced.