This is a question about Random Effects and Grand Mean Scaling? the study had two groups with 12 subjects in each group --- 3 runs/subject --- 4 block types repeated across the 3 runs (counter = balanced) -- I created and estimated a fixed effect model for each individual = subject treating the three runs as separate sessions ( I did not = concatenate runs) -- My concern is this: since I did not create a fixed effect model incorporating all subjects = together I did not do any 'groupwise' or 'subjectwise' scaling ---- Can = I (or do I need to) compensate for this during the random effects = analysis (two sample t-test)? Is there way to do scaling across both = groups to normalize intensity levels? Thanks, Sigi =20