Hi Marzia, These 3 t-tests are testing each if the respective group have FA > 0. The F-test is testing if any group has FA > 0. These test's aren't much useful for FA, as we know FA positive everywhere. Instead, I guess what you really want is to compare groups (I'll call them just A, B and C). For this, redefine the contrasts as: C1: 1 -1 0 (A > B) C2: 1 0 -1 (A > C) C3: 0 1 -1 (B > C) C4: -1 1 0 (A < B) C5: -1 0 1 (A < C) C6: 0 -1 1 (B < C) Define an F-test using C1 and C2. If this F-test is significant, this means that there is a group difference among the 3 groups. You can then proceed looking into the t-tests to see which groups differ. All the best, Anderson On 21 February 2015 at 17:46, Marzia Antonella Scelsi < [log in to unmask]> wrote: > Of course I can, I attach the files I created with the GLM gui :) > >