This paper by Lung-Fei Lee discusses generalized econometric models with selectivity, focusing on models where the dependent variable is either censored or has qualitative characteristics. The author reviews existing literature on models with censored and qualitative dependent variables, noting that while normal distributions are commonly used, they may not always be appropriate. Lee proposes methods to construct models with specified marginal distributions, particularly for cases where the disturbances are not normally distributed. Two-stage estimation methods are also derived for these models, extending the traditional two-stage estimation procedures to handle more complex scenarios. The paper includes detailed derivations of likelihood functions and asymptotic covariance matrices for the proposed models, making them computationally tractable and theoretically sound. Additionally, the paper introduces multiple-choice models with mixed continuous and discrete dependent variables, providing a flexible framework for analyzing such data. The author concludes by highlighting the practical and theoretical advantages of the proposed models and estimation methods.This paper by Lung-Fei Lee discusses generalized econometric models with selectivity, focusing on models where the dependent variable is either censored or has qualitative characteristics. The author reviews existing literature on models with censored and qualitative dependent variables, noting that while normal distributions are commonly used, they may not always be appropriate. Lee proposes methods to construct models with specified marginal distributions, particularly for cases where the disturbances are not normally distributed. Two-stage estimation methods are also derived for these models, extending the traditional two-stage estimation procedures to handle more complex scenarios. The paper includes detailed derivations of likelihood functions and asymptotic covariance matrices for the proposed models, making them computationally tractable and theoretically sound. Additionally, the paper introduces multiple-choice models with mixed continuous and discrete dependent variables, providing a flexible framework for analyzing such data. The author concludes by highlighting the practical and theoretical advantages of the proposed models and estimation methods.