I am trying to do a 3dANOVA on some seed connectivity maps. I would like to do the analyses within a specific mask. My understanding is that when one does a 3dANOVA within a mask, the F and t-statistics should change for each voxel because there are fewer voxels tested. I tried to apply the mask by adding – mask to the 3dANOVA command. However, the F and t-statistics for each voxel still exactly matched the whole brain analysis. I know that the mask did work in other respects in that areas outside the mask did go to zero. What then is the correct way to do this analysis?