The goal of this project is to improve replicability and reproducibility in social science by developing easy-to-use tools for researchers to share confidential research data in a privacy-protective manner, supported by rigorous computational, institutional, and legal foundations. We are motivated by the opportunities in social science created by massive new sources of data and developments in data analysis and sharing, and by the threat that privacy concerns pose to realizing the full potential of social science research. By leveraging ongoing multidisciplinary collaborations and theoretical advances in computation, statistics, law, and social science, the project aims to improve the replicability and reproducibility of data in empirical social science. Our goal is to develop and extend integrated privacy-preserving tools for enabling access to, use, and disclosure of social science data. The proposed project builds on a successful, ongoing multidisciplinary collaboration supported by an NSF Frontier grant: Privacy Tools for Sharing Research Data.
Objectives for this project include (1) analyzing the institutional and stakeholder incentives for managing research data privacy and the policy consequences of implementing new computational and legal privacy tools and concepts; (2) designing a blueprint for securing large-scale confidential archival data in the Dataverse repository; (3) exploring applications of our new computational and legal privacy tools to massive data and selected use cases, including online education data, human subjects research data, and economic data protected by NDAs; and (4) expanding research collaborations to engage with other differential privacy and privacy law experts, ongoing data privacy and dissemination efforts at MIT and Harvard, and several related Sloan Foundation projects.
This project is supported by a grant from the Sloan Foundation. For more information, please see the original proposed project description: