Genesin L., Bertoni M., Paggiaro A., Menardi G. (2025). Unobserved heterogeneity and identification of causal effects using mixture models, Conference Paper (SIS 2024)

Abstract
We explore the opportunity of identifying potential sources of latent heterogeneity by employing a finite mixture regression model aimed at evaluating the causal effect of a treatment. Within each latent class of heterogeneity, the model describes the relationship between the outcome and the treatment, controlling for a set of covariates, and manages the endogeneity of the treatment. The proposed methodology can be applied to data of arbitrary nature, also allowing for the identification of possible heterogeneous effects.