On combining estimates of relative potency in bioassay

Abstract

Biological assays based on a quantitative response in particular (e.g. hormone assays) typically involve the estimation of (log) relative potency (= M) with respect to some standard preparation on the basis of one or more symmetrical determinations usually at equally spaced (log) dose intervals. It is the purpose of this paper to consider further statistical aspects of estimates of M, and especially confidence limits for M based on a combination of separate assays. These are given as examples of tests of linear hypotheses. Two applications are given on the combining of quantitative bio-assay data.