This paper discusses the application of discrete choice models in welfare economics. It shows how conventional methods of applied welfare economics can be adapted to handle discrete choice situations, focusing on the computation of excess burden of taxation and evaluation of quality change. The results are applied to stochastic utility models, including probit and logit analysis. The paper emphasizes the importance of rigorous guidelines for applied work in discrete choice situations. The paper begins by discussing the traditional methods of applied welfare economics in continuous choice situations, including the computation of compensating variation and the relationship to the Harberger "excess burden" and Marshallian consumer's surplus. It then extends these methods to discrete choice situations, showing how the computations are performed in such cases. The paper also discusses the application of these results to quality changes in goods and services, and how they can be applied to stochastic utility models. The paper concludes by demonstrating that the conventional methods of applied welfare economics can be generalized to handle cases involving discrete choices. It emphasizes the importance of careful generalization, as some have suggested that conventional welfare change computations may have no relevance in discrete choice cases. The paper also highlights the differences in the relation between ordinary and compensated demand curves that arise from discrete choice, which affect the way empirical approximations should be carried out. The paper provides a rigorous framework for applying these methods in discrete choice situations, and shows how they can be used to evaluate the welfare impact of changes in quality and price.This paper discusses the application of discrete choice models in welfare economics. It shows how conventional methods of applied welfare economics can be adapted to handle discrete choice situations, focusing on the computation of excess burden of taxation and evaluation of quality change. The results are applied to stochastic utility models, including probit and logit analysis. The paper emphasizes the importance of rigorous guidelines for applied work in discrete choice situations. The paper begins by discussing the traditional methods of applied welfare economics in continuous choice situations, including the computation of compensating variation and the relationship to the Harberger "excess burden" and Marshallian consumer's surplus. It then extends these methods to discrete choice situations, showing how the computations are performed in such cases. The paper also discusses the application of these results to quality changes in goods and services, and how they can be applied to stochastic utility models. The paper concludes by demonstrating that the conventional methods of applied welfare economics can be generalized to handle cases involving discrete choices. It emphasizes the importance of careful generalization, as some have suggested that conventional welfare change computations may have no relevance in discrete choice cases. The paper also highlights the differences in the relation between ordinary and compensated demand curves that arise from discrete choice, which affect the way empirical approximations should be carried out. The paper provides a rigorous framework for applying these methods in discrete choice situations, and shows how they can be used to evaluate the welfare impact of changes in quality and price.