Seminario RDMD with Incomplete Information. An Application to English Schools

Abstract

We extend the Research Design Meets Market Design (RDMD) framework to settings where information on some priority criteria is partially missing. Using a probabilistic approach and machine learning methods, we obtain the correct assignment cutoffs under mild assumptions. We then derive a closed-form measurement error correction that recovers the RDMD estimand. An application to English secondary schools demonstrates the empirical relevance of our methods.