Differentially Private Causal Inference Under Hierarchical Design

In causal inference, matching is a crucial step as it allows for more reliable analyses by comparing properly matched pairs. However, real-world scenarios often present challenges for effective matching due to data fragmentation, as data are often collected and hosted by different local institutions. Additionally, the aggregation of separately stored datasets must prioritize privacy protection. To address these challenges, we propose a new hierarchical framework that implements differential privacy on raw data to protect sensitive information while maintaining data utility. We also design a data access control system with three different user access levels based on their roles, ensuring secure and controlled access to the aggregated datasets. We show the flexibility and practicality of our framework through simulation studies and analyses of a dataset from the 2017 Atlantic Causal Inference Conference Data Challenge. This paper contributes to the development of statistical methodologies in matching and privacy-preserving data analysis, offering a practical solution for data integration and privacy protection in causal inference studies.