Data Privacy & Confidentiality

Using big data to inform policy can be transformative, with analyses that often represent the entire population or allow data disaggregation to study small groups. This requires the responsible use of big data, including privacy protections when accessing data and publishing results to minimize potential harms to individual or group interests. While traditional methods for privacy protection are still relevant, newer more powerful methods have emerged and continue to evolve. MDI researchers build tools and conduct research that stimulates discourse about privacy – clarifying what must be protected, how it can be protected, and why increased privacy protections are needed.

Racecar Project: This project seeks to develop unobservable and compromise-resistant obfuscation channels for censorship-resistant communication.

Privacy Preserving Technologies in Education: This project documented K-12 and higher education institutions’ barriers to adopting innovative data access techniques, known as privacy preserving technologies (PPTs).  Beginning in the spring of 2023, three pilot PPT projects will begin with U.S. state education agencies, to show the value of these advanced technologies and how these barriers to implementation can be overcome.

NCES project: In collaboration with the U.S. Department of Education, MDI conducted a demonstration pilot of a privacy technology that joins two sensitive datasets in an encrypted fashion, in order to compute statistics on federal student aid. This project established MDI’s role as a major player in the federal government agencies’ innovation of their data infrastructure systems.

Advancing Use of Federal Earnings Data: This project is developing a Statistical Query Service for the Internal Revenue Service (IRS) to process state agency requests for linked earnings data. This Statistical Query Service will allow state agencies across social services, education, health, and beyond to obtain privacy-protected statistics on their constituents’ tax filing statuses, which will inform their program evaluation and evidence-based policy making efforts.

Faculty

Dayanand Manoli (He/Him)

McCourt School of Public Policy, Associate Professor

Kobbi Nissim (He/Him)

Department of Computer Science, Professor and McDevitt Term Chair

Micah Sherr (He/Him)

Department of Computer Science, Callahan Family Professor

Lisa Singh

Director, Massive Data Institute
Sonneborn Chair | Chair and Professor, Department of Computer Science | Professor, McCourt School of Public Policy