the calculation of effect size (r2) for a repeated-measures anova does not take into account variability that can be attributed to differences in subjects.