Adapting Field Experiments to Improve Generalizabilit

• In Event: Causal Complexity: Effect Heterogeneity, Mediation, and Generalization

Abstract

We propose a unique multi-methods strategy for robust generalizations of RCT experimental effects. There are five crucial elements of the method: 1) a classic RCT conducted in 6 rural Districts of Ghana; 2) a digital stratification frame; 3) MrP generalization of initial RCT treatment effects; 4) ChatGPT generalization of initial RCT treatment effects; 5) online supplement to update, adaptively, generalized treatment effect

Author

• Raymond Duch, CESS Nuffield