Emmanuel,
> I have a 2 session design with condition A and baseline
> (BASE) in session 1 and condition B and BASE in session 2.
> I have not used global scaling in the model.
>
> Is the best way to do an A vs. B, RFX comparison a paired
> t-test of A-BASE vs. B-BASE?
Yes. Equivalently, you could also do a one-sample t-test of
(A-BASE)-(B-BASE). The BASE's don't cancel out as they are estimated
seprately for each session.
-Tom
-- Thomas Nichols -------------------- Department of Biostatistics
http://www.sph.umich.edu/~nichols University of Michigan
[log in to unmask] 1420 Washington Heights
-------------------------------------- Ann Arbor, MI 48109-2029
|