A Comparative Assessment of Methods for Protecting Confidentiality
Wilson, D. (2005, August). A Comparative Assessment of Methods for Protecting Confidentiality. Presented at Joint Statistical Meetings, Minneapolis, MN.
A variety of statistical methods have been developed over more than thirty years in order to address the dual issues of confidentiality and analytic utility faced by data producers. The early focus of statistical disclosure limitation techniques focused on tabular data while methods applicable to microdata were subsequently developed.
This paper discusses some of the statistical disclosure techniques that have been developed and compares the assumptions behind the techniques in terms of when the techniques are applicable, qualitatively assesses the ability to implement the techniques, explores the availability of software for implementing the techniques, and examines the ability of the methods to limit disclosure risk and information loss. Differences in the type of disclosure issues associated with microdata and with tabular data are discussed and methodologies relevant to each type of data are described. The process of identifying “risky” data and choosing appropriate statistical disclosure methods is outlined.