Dear Michael,
Two options:
* at the first level (within-subject), form a contrast for the effect of
interest ([1 1 -1 -1], [1 -1 1 -1], [1 -1 -1 1] or [1 1 1 1]) and use a
one-sample t-test at the second level with these contrast images as input.
* use a flexible factorial design with three factors - a subject factor
and your two within-subject factors - and model main effect of 1 (ie
subject) and interaction of [2 3].
Best regards,
Guillaume.
On 24/08/17 21:02, Michael L. wrote:
> Hi,
>
> I have a simple 2x2 within-subjects design and I am wondering which design (flexible or full factorial) I should choose at second level. Since I'm interested in testing the main effects and interaction and only have one group of subjects I think I should choose full factorial, but I also read some posts that made me think that the flexible factorial option is correct. So which design would be more appropriate in my case?
>
> Best regards,
> Michael
>
--
Guillaume Flandin, PhD
Wellcome Trust Centre for Neuroimaging
University College London
12 Queen Square
London WC1N 3BG
|