While linear models have been central to the applied statistics toolkit since at least the 1970s, in practice linearity (on some scale) is often either unrealistic or fails to embody adequately the key research questions in an analysis. The development of generalised nonlinear models has taken place largely in rather specific application contexts; some examples are compartmental decay models in pharmacokinetics, gene-environment interaction models in crop science, association models in sociology, and multiplicative scaling models in psychology and political science.

Researchers in this area: Firth