Hi,
This all looks correct to me.
As long as you include the -D option to demean the data then this is fine.
All the best,
Mark
> On 29 Sep 2017, at 10:40, Arman Bordbar <[log in to unmask]> wrote:
>
> Hi,
>
> I am running an experiment in which I should find the correlation between FA values and behavioral data. My question is about setting the design matrix. I have already searched for my answer in forum and actually I have found that but as long as my results are strange I would like to have your confirm to exclude the eventual error of design matrix setting. Any suggestion would be highly appreciated.
>
> I have 43 scores of creativity and 43 fa maps. I have used Glm gui to set my design.mat and design.con. I should have just one EV right? And I put as values the manually demeaned scores in the order of the images in all_FA_skeletonized file. As contrast I have put two values. One 1 for positive correlations and one -1 for negative correlations. In the randomise command I add -D option and --T2.
>
> Do you see any error with my designs???
>
> Thanks in advance
>
> Arman
|