Disclosure Limitation Techniques for Tabular Data
Joe Fred Gonzalez · 2009
In order for the U.S. Federal Statistical System to collect and release reliable and valid data, it must gain the cooperation and trust of its survey respondents (persons or establishments). To achieve this, federal statistical agencies must pledge confidentiality (under penalty of law) to protect respondents’ data prior to collection and public release. Survey data released to the public are usually in the form of microdata and tabular data. This paper will focus on techniques for disclosure avoidance of two-dimensional tabular data. Tabular disclosure limitation techniques that will be presented are: complementary cell suppression, minimum-distance controlled rounding, unbiased controlled rounding, controlled rounding subject to subtotals constraints, and controlled tabular adjustment (CTA). Before and after perturbation results will be compared.