 |
 |
Hi Isik, For some reason I keep receiving a message that the attached Excel file is corrupted. At the end it opens but it isn't clear whether there was data or formatting lost. Would you mind in sending in .ods format? (use OpenOffice or LibreOffice, which are free; if you use FSL in Linux, your machine should already have either of these installed) Thanks. Cheers, Anderson On 15 February 2017 at 02:47, Isik <[log in to unmask]> wrote: > Dear PALM experts, > > I am trying to build a model to test correlation differences accross group > and condition between a continuous variable and my measurements > (connectivity). > > I have 2 groups, 3 within conditions and a continuous variable (age). > > For within measures, I am interested in age x group x condition, and age x > condition contrasts. > And for between the groups contrasts, I would like to see the slope > differences (age, age x group, and age x group for each condition). > > I prepared the design matrix (I am attaching here), and I would like to > make sure that I generated the correct design to test what I need. Does it > sound ok? > > My questions are > > 1. Is it important for this design to demean the continuous variable (I > think not, considering that I am only testing for slopes) > 2. In between group design, should I consider the mean for each condition > instead of the first 2 EVs? > 3. I am not exactly sure if the EB is correct -- I am doing block > permutations and within permutations. The variation also should be > comparable within the groups since they are connectivity values. > > I am running the palm code as > > -d design.mat > -T > -eb EB.grp > -ee > -ise > -whole > -within > -cmcx > -cmcp > -syncperms > -zstat > > thank you > isik >