ABSTRACT
Data which quantify the effect of exposure of individuals to substances such as new drugs, or to radiation, are often described as discrete, or quantal. Puri and Senturia proposed an elaborate mechanistic model for the way in which insects might attempt to shed insecticide, through a random sequence of losses of random amounts. This model was then used to fit data by supposing that the individual insect hazard rate at any time was a function of the amount of insecticide remaining by that time. Mechanistic models may provide much more, since the parameters of the model might correspond to definite aspects of the supposed mechanism. By simplifying the data, models may be thought of as smoothing out variation, and so fitted models may be used for interpolation. The model-fitting required in the analysis of quantal response data nearly always involves numerical procedures for fitting non-linear models, and so inevitably computers are involved.
