Dear all, is it possible to run a post hoc comparison using the glht function on random factors from a mixed model? For example, in the model lie(Yield~1+Meadow, random=1|Site/Area/Quadrate). thank you, Vasillis Papathanasiou, Marine SciencesPalaion Fokon 35-37Thessaloniki, 54454Greecetel:+306945874855 [[alternative HTML version deleted]]