HI, I ran into trouble trying to do a power computation. In the presence of interaction, I am simulating clinical trials in multiple centers. Suppose I have 3 centers, each with 20 patients, 10 in active drug and 10 in placebo and I have efficacy data for the 60 patients, ie, initial and final response variable scores for all 60 subjects. How do I compute the power of the study? (May be using an f-test?) Regards Krishna S P