Long-term Causal Inference Under Persistent Confounding via Data Combination

The dataset used in the paper is a combination of experimental and observational data for long-term causal inference. It includes short-term outcomes and long-term outcomes, and is used to estimate the average long-term treatment effect.