n. in statistics, refers to the classic dilemma on how to compute and test for the equality (or the difference) between the means of two normally-distributed populations when one cannot make any assumptions, whatsoever, that their variances are equal. Named after W.V. Behrens and Ronald Fisher.