It is, when you prior is that the populations should be different with a difference in the opposite direction. But even than the magnitude of the observed effe

It is, when you prior is that the populations should be different with a difference in the opposite direction. But even than the magnitude of the observed effect is so noisy ( large type m error)