Statistical Disclosure Control

Objectives

The two day course will introduce basic and advanced concepts of statistical disclosure control, privacy and confidentiality. The topics covered include the motivation of statistical disclosure control in terms of disclosure risk scenarios and types of disclosure risk; measuring disclosure risk for traditional outputs: microdata and tabular data; common methods of statistical disclosure control applications; the impact of statistical disclosure control methods on utility. In addition, we introduce differential privacy, a mathematical rigorous definition of a perturbation mechanism that was developed by computer scientists, which provides formal and quantifiable guarantees of confidentiality. Differential privacy is currently being explored by statisticians working within statistical agencies as we move towards more advanced and flexible modes of data dissemination.

Course outline

Introduction and motivation of statistical disclosure control for statistical outputs: disclosure risk scenarios, types of disclosure risk

• Measuring and quantifying disclosure risk for tabular data, microdata and other forms of disseminated outputs

• Disclosure control methods for statistical outputs

• Measuring and quantifying the impact and the effect of disclosure control methods on the quality of the data

• The differential privacy standard

• New forms of data dissemination and future challenges in statistical disclosure control

Prerequisites

Not indicated