The problem of disseminating tabular data such that the amount of information provided satisfies the public need while protecting individually identifiable data is a problem in all governmental statistical agencies. The problem falls into the category of Statistical Disclosure Control and provides many difficult policy and technical challenges for these agencies. In order to achieve the double mission of dissemination and confidentiality protection, the agencies must balance conflicting objectives. Traditionally, agencies have relied on selective suppression of sensitive cells. Because of the difficulty of suppressing optimally and the problems that may result from publishing tables with omitted cell values, new ideas have been proposed based on selective adjustment of cell values. One such method is "Controlled Tabular Adjustment" by Cox and Dandekar (2002). In this paper we discuss the theoretical, computational, and practical issues of these two approaches to Statistical Disclosure Control.
Source: U.S. Census Bureau, Statistical Research Division