Behrens-Fisher problem


Also found in: Wikipedia.

Behrens-Fisher problem

[¦ber·ənz ¦fish·ər ‚präb·ləm]
(statistics)
The problem of calculating the probability of drawing two random samples whose means differ by some specified value (which may be zero) from normal populations, when one knows the difference of the means of these populations but not the ratio of their variances.