Combining V-test probabilities in orientation studies: a word of caution