the problem of comparing the means of two samples drawn from different populations when the variances of these populations are not assumed to be equal. In other words, the problem involves the difficulty of making statistical inferences in the absence of equal population variances. The Behrens–Fisher problem may be addressed in many ways, such as by using a t test to estimate a probability level (p value) or by using confidence intervals to estimate the range of uncertainty around the mean difference. [W.U. Behrens; Ronald Aylmer Fisher]