The Cubic Logistic Model for Quantal Assay Data

SUMMARY Three‐ and four‐parameter models have recently been proposed for quantal assay data. These models are useful for judging whether the fit of simpler standard models, such as the logit, can be improved; better fits could result in better determination of extreme dose levels. However, a disadvantage of these new models is that they are often difficult to fit to data, and so are unlikely to be widely used. One of these models is well‐approximated by a much simpler model, for a wide variety of cases, and maximum‐likelihood estimates of parameters for this model can be obtained readily by the method‐of‐scoring.