SUMMARY Knowledge of one or more of the parameters in a maximum-likelihood problem sometimes results in a very considerable simplification, the analysis becoming almost trivial. A method of obtaining maximum-likelihood estimates and their asymptotic covariance matrix is given which makes it possible to capitalize on simplifications of this sort when they arise. Application is mainly in the field of regression, but other examples are mentioned. A numerical example of exponential regression is given.
Frank Richards (1961) studied this question.