THE CALCULATION OF THE DOSE-RESPONSE LINE IN QUANTAL ASSAYS WITH SPECIAL REFERENCE TO OESTROGEN ASSAYS BY THE ALLEN-DOISY TECHNIQUE

Abstract

A comparison has been made of the probit, logit, angular and rectangular transformations in the calculations of quantal dose response lines by the method of maximum likelihood. Data from Allen-Doisy oestrogen assays have been employed. Studies of the χ2 test for linearity and of the calculated median effective dose (m.e.d.) show no significant differences between the results obtained with any of the transformations. When group numbers are equal the angular transformation is the method of choice as the constant weighting factor reduces the amount of computation involved. The significance of these results, both in the calculation of the m.e.d. or relative potency and to the theory of drug action, has been discussed.