Hi Hsekar, In FSL (and in imaging in general), the t-tests always have a direction. If we want the equivalent answer of the two-tailed, and if alpha=0.05, we look at both results at 0.05/2. With many components, we then do 0.05/(2*n). If we had from the very beginning two-tailed t-tests, we would simply do 0.05/n, but not 0.05/2, nor 0.05/(2*n). It works in the same way for ANOVA (or any F-test): we do 0.05/n, as these are already non-directional. Hope this clarifies. Cheers, Anderson -- Anderson M. Winkler FMRIB / Analysis Group [ Blog <http://brainder.org/> | Twitter <http://twitter.com/AndersonWinkler> ] On 10 August 2015 at 02:25, hsekar.b <[log in to unmask]> wrote: > Dear Anderson > Thank you for the mail, > I was interested what would be the denominator, if i would have an ANOVA > design involving 3 groups. > Is it 0.05/n * 3 (because of 3 groups) > As in the case of two tailed t test involving 2 groups p is < 0.05/n*2. > > > Thank you in advance > > HB >