Ive got a simple anova with 1 factor which is groupe, with 3 levels.
The variance is heterogenous, it is really different between groups.
In SAS, I would add a statement REPEATED / GROUP = groupe; to take care of the problem.
Is something similar available in SPSS? I'm working with the function UNIANOVA. I know that there are multiple comparisons tests available for heterogeneous variance, but it isn't what I want, it does not correct the type III test nor the residuals of the model.