Organizer(s):

Ruobin Gong, Rutgers University

John Abowd, Cornell University & United States Census Bureau

Abstract

Advances in computing power and optimization algorithms have made traditional statistical disclosure limitation methods, especially those used for unweighted tabulations of person and household data, increasingly vulnerable to successful confidentiality-breaching attacks. Effective mitigation of this disclosure risk, particularly for a highly granular statistical product like the Decennial Census of Population and Housing, which routinely publishes far more statistics than it has data points, requires careful evaluation and application of quantifiable and tunable confidentiality protections. This talk provides a technical overview of the differential privacy framework for the 2020 Census Disclosure Avoidance System. I examine some of the implementation constraints, policy decisions, criticisms, and lessons learned from its development and implementation.

Speaker Bio:

John Abowd is an executive senior advisor for the Research and Methodology Directorate of the Census Bureau and is the former associate director for research and methodology, and chief scientist, a position he held from June 2016 to October 2022. The Research and Methodology Directorate leads critical work to modernize our operations and products.

He is currently leading the agency’s efforts to create a differentially private disclosure avoidance system for the 2020 Census and future data products. His long association with the Census Bureau began in 1998 when he joined the team that helped found the longitudinal employer-household dynamics program. In 2008, he led the team that created the world’s first application of a differentially private data protection system for the program’s OnTheMap job location tool. Abowd is also the Edmund Ezra Day Professor emeritus of economics, statistics, and data science at Cornell University. He is a fellow and past president of the Society of Labor Economists. He is also a fellow of the American Association for the Advancement of Science, American Statistical Association, and Econometric Society, as well as an elected member of the International Statistical Institute. He earned his M.A. and Ph.D. in economics from the University of Chicago and A.B. in economics from the University of Notre Dame.

Parking:

If you do not have a Rutgers parking permit and you plan to drive to the event, there will be free parking available, but you must register your vehicle to park. Once your vehicle has been registered, please park in Lots 59, 60A, or 60B on the day of the event. If you do not register your vehicle, or if you park in unauthorized lots, you may receive a citation. If you are Rutgers-affiliated and already have a Rutgers parking permit, you must park only in lots where you are authorized to park.

Lunch:

There will be lunch with the speaker beginning at noon at DIMACS in room 401 of the CoRE Building. For those who wish to join for lunch but are not affiliated affiliated with the Rutgers Statistics Department, please send email to Nicole Clark-Johnson on or before March 17 to let her know you will attend. (Email is not needed if you are part of the Statistics Department or do not plan to attend lunch.)