Prediction-Powered Data Fusion for Treatment Effect Estimation

2026-10-08

AI Summary

This study addresses the challenge of estimating average treatment effects (ATE) and conditional average treatment effects (CATE), where small-sample randomized controlled trials (RCTs) yield high estimation variance while observational studies suffer from confounding bias. To overcome these limitations, this work proposes a data fusion framework that avoids imposing strong assumptions on observational data. Leveraging a prediction-driven strategy, novel estimators and learning algorithms—namely AIPW-Fusion, DR-Fusion, and R-Fusion—are developed to preserve the unbiasedness of RCTs while harnessing large-scale observational data to enhance statistical power. Experimental results demonstrate that the proposed approach significantly reduces estimation variance and improves both the accuracy and robustness of CATE estimation. Furthermore, it yields precise ATE and CATE estimates accompanied by closed-form confidence intervals.

Abstract

Randomized controlled trials (RCTs) identify treatment effects without confounding but are often small, whereas observational studies (OBS) are large but may be confounded. Many estimators combining a small RCT with a large OBS have been developed for the average treatment effect (ATE) and the conditional ATE (CATE). However, existing ATE estimators either make assumptions on the OBS or do not borrow enough power from them. The CATE has been studied less than the ATE. Existing CATE methods either assume the OBS are unconfounded, rely on a model of the confounding function, or accept bias in exchange for lower variance. We therefore propose a framework that, without special assumptions on the OBS, fuses the OBS and the RCT by preserving the unbiasedness of RCT-based estimation while borrowing power from the large OBS to boost precision. Applying this principle, we build an ATE estimator, AIPW-Fusion, with closed-form weights and confidence intervals, and two CATE learners, DR-Fusion and R-Fusion. Experiments corroborate our findings.

Problem

• Treatment Effect Estimation
• Data Fusion
• Randomized Controlled Trials
• Observational Studies
• Average Treatment Effect

Innovation

• Treatment Effect Estimation
• Data Fusion
• Randomized Controlled Trials
• Observational Studies
• CATE