Dear Cyril and list,
I have been following this thread on covariates-of-interest and
covariates-of-no-interest. I have a contrast that I would like to correlate
with a predictor of interest using multiple regression. However, the
contrast is a second-order contrast, in the sense that it is a contrast on
two prior contrasts and multiple regression does not accept pairs of scans
like the paired t-test.
How would you go about this problem for a multiple regression model?
Best regards,
Martin
|