Hi, Just look at the 1-factor 4-level case on the wiki (http://fsl.fmrib.ox.ac.uk/fsl/fslwiki/GLM#ANOVA:_1-factor_4-levels__.281-way_between-subjects_ANOVA.29) and adapt it to 3 levels. All the best, Mark On 7 Aug 2014, at 19:59, Alena Piovar <[log in to unmask]<mailto:[log in to unmask]>> wrote: Hi Eugene, thanks for answer, but how to set up contrast if I want to do ANOVA then? Best Alena 2014-08-07 13:03 GMT-05:00 Eugene Duff <[log in to unmask]<mailto:[log in to unmask]>>: Hi Alena - These are fine if all you want are the individual t-tests. Eugene On 7 August 2014 18:24, Alena Piovar <[log in to unmask]<mailto:[log in to unmask]>> wrote: Hi, I have basic question about how to set up three group difference matrix and contrast correctly. Should I use the F-test as well? is the use group EV1 EV2 EV3 input1 1 1 0 0 input2 1 1 0 0 input3 2 0 1 0 input4 2 0 1 0 input5 3 0 0 1 input6 3 0 0 1 contrast EV1 EV2 EV3 C1 group A>B 1 -1 0 C2 group A<B -1 1 0 C3 group A>C 1 0 -1 C4 group A<C -1 0 1 C5 group B>C 0 1 -1 C6 group B<C 0 -1 1 is the contrast set up correct? Looking forward to you answer Best Alena