An ad hoc planning committee will organize a 1.5-day public workshop to discuss emerging practical issues concerning the protection of confidentiality of data subjects and their individually identifiable data while making information publicly available. Federal statistical agencies are typically legally obligated to protect the privacy of data subjects' identities and sensitive attributes. The rapid emergence of accessible digital data and powerful computational resources--while enormously beneficial for research and policy making--has made protecting privacy while ensuring analytic usefulness increasingly difficult. Most of the tried and true methods employed to obfuscate data were designed for a different era and can no longer can be counted on.

Researchers in computer science, economics, mathematics, and statistical science have been developing new methods for creating public use data products with provable guarantees of privacy protection. Among these, differential privacy has emerged as an important approach because of its ability to provide provably formal privacy protection. To date, however, this research has not been widely transferred into practice, despite its theoretical advantages over typical approaches to privacy protection. Workshop participants will have the opportunity to learn about the promises and limitations of using formal privacy methods for dissemination of data products.

The planning committee will define the specific topics to be addressed, develop the agenda, and select and invite speakers and other participants. After the workshop, a proceedings of the presentations and discussions will be prepared by a designated rapporteur in accordance with institutional guidelines.