Dear everyone, I have a 2x3 design (medication x stimulus type) but very few subjects (6) and I would like to perform intra-subject stats using r's aov (I've already run the group analysis). Is there a conceptual problem with this, as long as I don't interpret the results as representative of the population with the disease I'm studying? Could anyone point me to any discussion of this practice, as to its merits for example? Many thanks, U. Pasos [[alternative HTML version deleted]]