SPM Savants,
I want to perform a second level RFX group comparison between two
groups, and group 1 has 10 subjects with 1 scan each, and group 2 has
5 subjects with 2 scans each.
Can I use use those extra scans in group 2?
It doesn't seem quite right to enter them as extra subjects, since
that violates the assumption that the scans be independent. One could
average the two scans in group 2 in a fixed-effects manner before
carrying them to the second level, but that violates the assumption
that the within-subjects variance be equal for subjects in both
groups.
Any insight appreciated!
Thanks!
Jim Lee
|