Doctor of Philosophy

dissertationThe logit model is a powerful tool for analyzing the voluminous discrete-purchase scanner data and extracting significant economic information from the data. However, the traditional logit model is the only model that forces the product market demand function to be independent of prices, which is a serious drawback for economic applications. A close examination of the basis and foundation of the logit model reveals that the market demand function is not part of the basic logit model. The basic model's only output is the brand share. This work develops the approach and methodology for using other product market demand founctions with the basic logit model. Several alternative market demand functions are presented and analyzed including one that satisfies the budget constraint. Accordingly, it is called the consistent logit model. The expanded logit approach has the benefit that can it be applied as easily as the traditional logit model, and much more easily than the nested logit model. Another significaant contribution made by this work is the edtensive compilation of empirically-determined price elasticities from the literature, grouped according to model. This compilation can serve as a handy reference for economists and others needing price elasticity information