A Score Test for Non-nested Hypotheses with Applications to Discrete Data Models

This paper suggests that a convenient score test against non- nested alternatives can be constructed from the linear combination of the likelihood functions of the competing models. It is shown that this procedure is essentially a test for the correct specification of the conditional distribution of the variable of interest. As in Models for discrete data it is often necessary to fully specify the conditional distribution of the variate of interest, the test proposed here is particularly attractive in this context. The usefulness of the proposed tests is illustrated with applications to discrete choice and count data models.