Disclosure risk assessment in statistical data protection

Abstract

Statistical data protection, also known as statistical disclosure control, is about methods that try to prevent published statistical information (tables, individual information) from disclosing the contribution of specific respondents, who may be individuals or enterprises. In addition to keeping disclosure risk acceptably low, methods used for statistical data protection should not significantly damage the utility of the data being protected. This paper surveys different ways to assess the risk of disclosure in the protection of both individual data (called microdata) and tabular data. A noteworthy result also presented is that the most widely used rule for assessing disclosure risk in tabular data protection is flawed.