Dear experts,
We have a query about degrees of freedom for ANOVA models with >1
within-subjects factor.
We used a flexible factorial model for a data-set with 15 subject, 3
sessions and 6 scans within each session (270 perfusion images). We have
subject as a factor, as well as a main effect of session (no interaction
term) and a globals column too giving 19 columns in the DM. We are most
interested in the session effect, but do not want to average the 6
within-session images if possible.
When we run the ffm for this, defining an f-contrast the df = [2,252]. But
we are thinking that for the main effect of session this is way too high.
Should we be concerned and if so is there a solution that doesn’t involve
averaging images?
Many thanks,
Owen O'Daly
|