Hamilton Institute Seminar

Wednesday, March 4, 2020 - 13:00 to 14:00
Hamilton Institute Seminar Room (317), 3rd Floor Eolas Building

Speaker: ​Professor Natalie Shlomo, University of Manchester, UK

Title: "Statistical Disclosure and Differential Privacy"

Abstract: For decades, statistical agencies have been disseminating statistical data in the form of microdata from social surveys and tabular data from censuses, surveys and registers. There are   many publications detailing disclosure risk scenarios, types of disclosure risks, statistical disclosure control (SDC) methods and the quantification of disclosure risk and data utility. However, these traditional forms of statistical data and their confidentiality protection rely heavily on assumptions that may no longer be relevant. In recent years, we have seen the digitalization of all aspects of our society leading to new and linked data sources offering unprecedented opportunities for research and evidence-based policies. These developments have put pressure on statistical agencies to provide broader and more open access to their data. On the other hand, with detailed personal information easily accessible from the internet, traditional SDC methods may no longer be sufficient and this has led to the opposite effect of statistical agencies restricting and licensing data as an SDC method instead of broadening access.   To meet the demands and challenges for disseminating more open and accessible data, particularly via synthetic data and web-based platforms where outputs are generated and protected on-the-fly without the need for human intervention, statistical agencies have been investigating more rigorous data protection mechanisms with stricter privacy guarantees to incorporate into their SDC toolkit.   One such mechanism is Differential Privacy (Dwork, et al. 2006), a mathematically principled method of measuring how secure a protection mechanism is with respect to personal data disclosures. We present some future dissemination strategies under consideration by statistical agencies and the potential for Differential Privacy to protect the confidentiality of data subjects with well-defined privacy guarantees. 
This work is joint with Professor Yossi Rinott, Dr. Christine O’Keefe and Professor Chris Skinner.