ABSTRACT
Multinomial data multinomial response is usually ordinal, providing a more detailed picture of gradual change than the one offered by a binary response. This chapter describes a different, pragmatic and practical approach towards analysis of multinomial dose-response data, exploiting that such data may be analyzed through a series of models fitted to suitably derive binomial dose-response data. One key advantage of fitting multiple models to binomial dose-response data is that it becomes easier to interpret the results. Another advantage is that over-dispersion, which can also occur for multinomial data, may be dealt with the same way as for binomial data. Arbovirus Morgan discusses a trichotomous multinomial data example where chicken embryos were exposed to increasing doses of an arbovirus in order to investigate the potency of the virus.
