Dear Dario,
The naming of the options for "main effect" and "interaction" in the
batch interface for flexible factorial design is probably the source of
confusion. When you specify one "interaction" in the interface, SPM will
create a design that will allow you to test for main effects and
interaction of the two factors you entered. There is therefore no need
to also specify the "main effects" of these two factors - it would just
be redundant.
Best regards,
Guillaume.
On 23/01/17 13:19, Dario Zaremba wrote:
> Dear Guillaume, dear Christian,
>
> thank you for your advice regarding the design. I see the example in the CAT12 manual for longitudinal data in two groups, but I was wondering why you exclude the main effects of group and time in your model and include the interaction of group x time only. Following the tutorial of Jan Gläscher and Darren Gitelman for flexible factorial designs (originally for SPM5), I thought that the inclusion of main effects should increase sensitivity for an interaction effect (find attached, page 11). Could you please comment on that?
>
> Best regards,
> Dario
>
--
Guillaume Flandin, PhD
Wellcome Trust Centre for Neuroimaging
University College London
12 Queen Square
London WC1N 3BG
|