Gang,
yes, you're right that the 3dANOVA3 would be preferable, I just haven't been able figure out how to specify the interaction to get direction of effects. I have the main effects set up as -acontr A1 = A2 (-1 1) and –bcontr B1 = B2 (-1 1), but am specifying the interaction using 'fab', which does not show direction. What arrangement of numbers in 3dANOVA3 would produce (A1 - A2) = (B1 - B2) with direction?
Thanks in advance.
Brian