Polychotomous Quantal Response in Biological Assay

Publisher Website
Google Scholar
Abstract
A general procedure is described for analyzing quantal bioassay data in which the response is one of several ordered outcomes, such as alive, moribund, dead, or varying degrees of moribundity. An example involving housefly mortality and moribundity is described. The parameters in the underlying tolerance distributions are estimated by minimum chi-square techniques through a generalization of the procedure developed by Berkson for the case of a dichotomous response. This generalized minimum chi-square procedure yields estimates more readily than by the maximum likelihood techniques, yet the same asymptotic properties hold. It is also evident theoretically, as well as from the data, that the variance of the logarithm of the relative potency estimated by using a polychotomous response is less than that of the corresponding estimate obtained by pooling the numbers of moribund and dead flies in order to make the response dichotomous.