Survey of Economists’ Perspectives on Differentially Private Methods

• Activity Number: 219 - Expanding Access to Administrative and Survey Data for Public Policy Decision-Making
• Type: Invited
• Date/Time: Tuesday, August 9, 2022 : 8:30 AM to 10:20 AM
• Sponsor: Government Statistics Section
• Author(s): Joshua Snoke* and Aaron R. Williams and Claire McKay Bowen and Andrés Barrientos
• Companies: RAND Corporation and Urban Institute and Urban Institute and
• Keywords: Disclosure control; Differential Privacy; Survey data; Tax data; Validation server; Privacy

Abstract

Economists frequently use federal data, for example federal tax data, for methodological and applied research, but much of these data are hard to access due to privacy restrictions. Tools exist using the differential privacy (DP) framework to provide researchers with access to noisy versions of data or statistics, but little is known about the broader perspective towards these methods among economists. In an effort to understand economists’ knowledge of DP and attitudes towards it, we conducted a convenience sample survey of members of the American Economic Association and received over 800 responses. Our questions identified baseline knowledge about DP, attitudes towards the framework, types of statistical methods that are most useful to economists, and how the injection of noise under DP would affect the value of the queries to the user. This talk will review the results and discuss possible next steps to create better privacy-preserving tools for the economics community.