On assessing contrasts between groups of animals