Dear Joana,
On Tue, 6 Jul 2021 16:31:43 +0000, Joana Crisóstomo <[log in to unmask]> wrote:
>Dear SPMers,
>
>I am running a TFCE analysis using the design matrix below. This is a longitudinal analysis with 2 groups, 2 time points, and a covariate (age).
Please read the many discussions on the SPM mail list about the meaningless use of such covariates for longitudinal designs. Depending on the age differences between your scans you will diminish your longitudinal effects in the worst case. In the best case your covariate will be ignore because it will almost not change between time points.
>
>[cid:image003.jpg@01D7728C.C848C880]
>
>
>I select Smith permutation method, as suggested when any nuisance variables exist. However, in Matlab command window, it seems the Draper-Stoneman method is used instead. Can you clarify which method is used?
Probably the covariate is not correctly recognized in longitudinal designs. Please remove it because of the reasons above.
Best,
Christian
>
>Thank you for your help.
>Best,
>Joana Cris�stomo
>
>
>
>
>
|