Research Research areasCausal inference

Causal inference

In this research area we deal with: Designs and methods for the estimation of causal effects in experimental and observational studies, with emphasis on the development of analysis methods, including Bayesian analysis, for experimental studies with complications (non-compliance, missing data) and non-regular observational studies (instrumental variables, matching, regression discontinuity designs). Applications include public policy assessment and clinical trial analysis.

 

Among the specific topics, we are dealing particularly with:

Causal inference in experimental and observational studies. Bayesian methods for the analysis of causal effects in: (1) experimental studies in the presence of post-treatment confounded intermediate variables, such as studies with post-treatment complications (non-compliance, missing data); (2) studies aimed at identifying causal mechanisms; (3) studies with treatments that vary over time. Analysis of observational studies with binary and continuous treatment variables with ignorability of the treatment assignment mechanism. Analysis of studies with non-regular treatment assignment rules (such as regression discontinuity designs). Causal inference methods in the presence of interference.

People: Baccini, Doretti, Mattei, Mealli, Menchetti, Rampichini, Sera