Dear FSL experts,
I'm wondering if someone can help with a question about the best way to demean trial-by-trial behavioural scores before entering them into my model.
My participants rated their "nervousness" on a 0-10 scale for each trial of a paradigm with multiple stimulus types. I want to add these ratings as weights in my 3-column EVs in the first-level analysis. Should they be demeaned on a participant basis (i.e., subtract each participant's *overall* mean nervousness rating from their rating for each trial) or should they be demeaned by participant and stimulus type (i.e., subtract each participant's mean nervousness *for that stimulus type* from their rating for each trial)? It may be relevant that the mean (and variance) of the nervousness ratings do vary by stimulus type.
Many thanks in advance,
Anneke
|