Online Program Home
My Program

Abstract Details

Activity Number: 458 - Differential Privacy Research and Applications at the U.S. Census Bureau
Type: Topic Contributed
Date/Time: Wednesday, July 31, 2019 : 8:30 AM to 10:20 AM
Sponsor: Government Statistics Section
Abstract #304448 Presentation
Title: Census Barriers Attitudes and Motivators Study: a Case Study in Differential Privacy at the U.S. Census Bureau
Author(s): Caleb Floyd* and Rolondo Rodríguez
Companies: U.S. Census Bureau and U.S. Census Bureau
Keywords: Differential Privacy; Multinomial Randomized Response Mechanism; Laplace Mechanism

The U.S. Census Bureau is modernizing its disclosure avoidance methodology, and as such, many data products in the future will use differential privacy to protect respondents’ confidentiality. We use such algorithms to protect privacy for the Census Barriers Attitudes and Motivators Study (CBAMS), which is a nationwide survey that covers topics related to census participation. In this paper, we create a public dataset by employing two differentially private methods – the Multinomial Randomized Response Mechanism on categorical variables and the Laplace Mechanism on continuous variables. We evaluate the privacy loss under local differential privacy and then create an accuracy measure to assess the accuracy/privacy tradeoff. Lastly, we discuss some of the successes and trials of implementing differential privacy to protect confidential survey data, and we share some of the lessons we’ve learned.

Authors who are presenting talks have a * after their name.

Back to the full JSM 2019 program