A Renewed Understanding of Complementary Cell Suppression

• Activity Number: 69
• Type: Topic Contributed
• Date/Time: Sunday, July 29, 2012 : 4:00 PM to 5:50 PM
• Sponsor: Committee on Privacy and Confidentiality
• Author(s): Lawrence Cox*+
• Companies: Lawrence H. Cox Consulting
• Keywords: disclosure limitation ; analysis of tabular data

Abstract

Disclosure in tabular data is associated with risky (sensitive) cells: in count data these are small counts, and in magnitude data cells where a total or net value is dominated by a few contributors. Cell suppression was the first technique developed for disclosure limitation in tabular data. Its use dates to the 1940s when only primary suppression-suppressing only risky cells-was performed. Additive relationships between cells renders primary suppression ineffective, thus complementary cell suppression (CCS) was developed. CCS involves suppressing additional, nonsensitive cells to thwart narrow estimation of sensitive values. CCS is an NP-hard mathematical problem-polynomial time algorithms for its solution are unlikely to exist. Nevertheless, sophisticated algorithms and software for CCS have been developed over the past 40 years. Unfortunately, internationally and within the US, many organizations employ primary suppression (only) or ad hoc CCS Moreover, user-developed techniques to analyze or work around suppressed data are not well understood. We report on a recent NISS workshop aimed at examining and improving suppression-based SDL and analysis.