Dear Miranda,
as mentioned in another mail from today the flexible factorial model (and also the paired two-sample T-test) don't handle these covariates properly:
https://www.jiscmail.ac.uk/cgi-bin/wa-jisc.exe?A2=SPM;c54b28f0.2209
If you only have two time points you could use the difference images and a simple multiple regression with your covariate.
Best,
Christian
On Wed, 10 Aug 2022 18:25:47 +0100, Miranda Ramirez <[log in to unmask]> wrote:
>Hello all
>
>I have a within subjects design, and I want to examine a covariate of interest. Each subject had two maps generated (wins > control and nowins > control). These are my two con maps going into the second level as a paired t-test (wins>control > nowins>control). I want to see how learning rate (covariate of interest) scales with this comparison, but I'm confused on how to enter the covariate and what the contrast vector should be.
>
>Should there be two vectors for the covariate values, one for the con_0001 and one for con_0002, or should they be in the same vector?
>
>I only have one learning rate per subject, so should this learning rate value be duplicated (0.5 0.5 per subject) or with zeros separating (0.5 0 per subject)?
>
>What should the contrast vector look like? My guess with how my design is currently set up, it would be .5 -.5 .5 -.5.
>
>
>Thanks!
|