Hi there,
> The study involves two seperate tasks: A and B. Some subjects take A after
> B, and others take B after A. We know for sure that subjects perform
> differently in these two tasks. Now, we want to do a paired t test with
> within-subject covariates from the order of tasks and also the behavior
> variables in A and B. In addition, we also want to consider gender and
> patient control covariates in our paired t test.
this looks like an AnCova .. so I would suggest either perform an AnCova
(adjusting the code in spm_spm_ui .. under AnCova change 'nC',[0,1] to
'nC',[0,Inf] ) or perform a multiple regression analysis (with cste)
and enter the data like 1111-111 for the task condition (or 1111000 and
00001111) then your covariate like 111-1-1-1111-1-1-1 for task AB and
BA and any other score, RT, .. this should do the job (given that a
multiple regression is exactly like the AnCova)
hope this helps (and works)
cyril
|