My question is related to the appropriate comparisons that can be done at the first vs. the second level analysis.
We would like to compare between two conditions A and B which were presented in separate runs (sessions). Condition A was only present in session 1 and condition B only present in session 2. My question is whether it would be correct to compare them at the first level, or only at the second level. and why.
My thought is that it would not be optimal to compare them at the first level, because the 'session regressor', which accounts for overall differences in the mean of the two sessions, would also eliminate the differences between the conditions of interest. Is this correct?
Thanks a lot!
University of Haifa