Modeling Qualitative Outcomes by Supplementing Participant Data with General Population Data: A New and More Versatile Approach

Although one often has detailed information about participants in a program, the lack of comparable information on non-participants precludes standard qualitative choice estimation. This challenge can be overcome by incorporating a supplementary sample of covariate values from the general population. New estimators are introduced that exploit the parameter restrictions implied by the relationship between the marginal and conditional response probabilities in the supplementary sample. An important advantage of these estimators over the existing alternatives is that they can be applied to exogenously stratified samples even when the underlying stratification criteria are unknown. The ability of these new estimators to readily incorporate sample weights make them applicable to a much wider range of data sources. The new estimators are also easily generalized to address polychotomous outcomes.