Hi Silvano,

Would you mind in sending your input data, design, full command line, and version used? You can put in some file sharing service and send the link off-list. Happy to have a look in a few days.

Thanks!

All the best,

Anderson


On 20 August 2017 at 07:24, Silvano Sele <[log in to unmask]> wrote:
Dear PALM experts

I run an analysis in PALM corrected over all 8 contrasts and was surprised by the large change of the p-values.  I have a 2 (between)*3(within design). I used a design for the 6 interactions (3*2 (onesided)) and a design for the between group contrast(also onesided both directions) according to a design i found in the mailinglist (using -ee -ise -within -whole).The minimum p-value i got in a between group contrast  (in logp) using TFCE changed from 2.98 (tfce_tstat_fwep) to 1.02(tfce_tstat_cfwep). The minimum voxelwise p-value only changed from 2.43 (vox_tstat_fwep) to 1.91 (vox_stat_cfwep) in this contrast. I was expecting at maximum a change in p-values times 8 and I'm now not sure how to interpret the tfce cfwe-p-values and if I should correct the main effect of group for these interactions.

Thank you very much for the help.

Best wishes
Silvano