Ciao Alessandro, I figured out that we can do a factorial analysis in two steps: First, make full model object with all main effects, interactions and biological replicate effects. Use matest (from maanova package) to test the interactions effects. Second, make a model object without interactions effects (only main and biological replicate effects). Use matest (from maanova package) to test the main effects. I have been used mixed model in both. I hope this help you, but, if you have another questions, please, ask me! Bests Marcelo 2008/2/20, Alessandro Fazio <fazioalessandro at="" hotmail.com="">: > > Ciao Marcelo! > > What do you mean step-by-step? I have used recently MAANOVA and I had a > similar problem in analysing interaction effect... > > Thanks > > -alessandro