COMPARING NONNESTED REGRESSION-MODELS

A method for comparing the fits of two non-nested models, based on a s uggestion of Davidson and MacKinnon (1981), is developed in the contex t of linear and nonlinear regression with normal errors. Each model is regarded as a special case of an artificial ''supermodel'' and is obt ained by restricting the value of a mixing parameter gamma to 0 or 1. To enable estimation and hypothesis testing for gamma, an approximate supermodel is used in which the fitted values from the individual mode ls appear in place of the original parametrization. In the case of nes ted linear models, the proposed test essentially reproduces the standa rd F test. The calculations required are for the most part straightfor ward (basically, linear regression through the origin). The test is ex tended to cover situations in which serious bias in the maximum likeli hood estimate of gamma occurs, simple approximate bounds for the bias being given. Two real datasets are used illustratively throughout.