Appropriate Scores in Bio-Assays Using Death-Times and Survivor Symptoms

Publisher Website
Google Scholar
Abstract
Many bio-assays can be arranged in a (mxk) contingency table with k doses and in m categories of biological observation, ranked in order of increasing effect of treatment (e.g., survival times [forward arrow] , survival with symptoms[forward arrow] , and survival without symptoms). The author obtained a set of m scores that satisfies one criterion of an efficient bio-assay: The variance of the linear regression of the mean scores on log dose is the highest possible fraction of the total variance. A procedure is described for combining data from separate experiments of similar kind, to obtain a score system so that the common slope is the maximum possible fraction of the total variance adjusted for individual means. Significance tests for different score systems are described and the method is applied to an inter-institutional assay of tetanus toxoid comprising 96 experiments.