Why I Prefer Logits to Probits

Abstract

The author restates his long-established preference for the use of the logistic function over the integrated Gaussian function in bioassays with all-or-none responses. Many proponents of the probit frequently justify its use on the basis of a probably non-existent distr. of individual to tolerances, whereas logit theory can be related to many different physico-chemical phenomena. Six examples of this are given. Some discussion of fitting procedures and best estimates is included.