Dear Peter,
Thank you for your response.
I have increased my sample to DCM analysis. I have 6 groups with 68 subjects (15, 10, 13, 11, 10 and 9 respectively). In order to find differences between the groups A and B (groupA > groupB), I drew my design matrix with fifteen 1s, ten -1s and forty-three 0s. I have found the following results (attached files):
Group_mean_P_095: Strong evidence to mean (68 subjetcs)
GroupA_GroupB_P_no_thresh: groupA>groupB (no threshold)
GroupA_GroupB_P_095: there is no differences when I set strong evidence. And this happened for all comparison. Do you know if it means that the variabillity is too high? Is there any option to find differences among them, please?
Thanks a lot.
Helka
|