He is most well known for the Rubin causal model, a set of methods designed for causal inference with observational data, and his methods for dealing with missing data.
