摘要：We study some types of models that do not fall into this category: - Models with discrete dependent variables, called qualitative response models where endogenous variables can take only two values (binomial dichotomous or binary response models) or a limited number of values (divided into several categories, multinomial, or model choice multiple); - Partially observed models (or models with limited dependent variables) where the dependent variable analyzes are reduced to a single value after a certain limit. We consider here models censored sampling designs. These latter models are characterized by a truncation process that depends on a latent variable that differs from the one that describes the observed data.
