In this article, I describe theescountcommand, which implements theestimation of an endogenous switching model with count-data outcomes, wherea potential outcome differs across two alternate treatment statuses.escountal-lows for either a Poisson or a negative binomial regression model with lognormallatent heterogeneity. After estimating the parameters of the switching regressionmodel, one can estimate various treatment effects with the commandteescount.I also describe the commandlncount, which fits the Poisson or negative binomialregression model with lognormal latent heterogeneity
© 2001-2024 Fundación Dialnet · Todos los derechos reservados