320 – Statistical Methods for Complex Survey Data
Confidentiality Approaches for Real-Time Systems Generating Aggregated Results
Jianzhu Li
Westat
Tom Krenzke
Westat
In recent years many government agencies and statistical institutions have endeavored to release statistical data through online real-time system - a high-technology product which is designed to convey information in a timely and flexible manner. In an online real-time system, the data users submit queries in required format and expect to receive tailored statistical results (e.g., tables with aggregated statistics) quickly. While a real-time system provides the convenience and flexibility of conducting statistical analysis, it faces the same challenge of protecting data confidentiality as traditional data dissemination approaches - the data products have to be screened and/or treated to ensure low risk of statistical disclosure. There are two types of confidentiality protection approaches for real-time systems: static and dynamic. The former implements the statistical disclosure control (SDC) treatments outside the system while the latter implements the treatments on the fly. In this paper we will review a few confidentiality approaches which are feasible for real-time systems, and discuss the use of risk and utility measures to justify the dissemination of data products.