Empirical evidences on protecting population uniqueness at idescat

Abstract

This paper describes the process of statistical disclosure analysis and control applied by the Statistical Institute of Catalonia (Idescat) to microdata samples from census/surveys with some population uniques. Since 1995, by means of models which allows calculation of the risk and data protection procedures, some empirical evidences have been achieved in order to check the performance of μ-ARGUS in a real situation of unique populations, with large files and re-identification keys. The analyzing way used preferably is the measuring of dimensions (to a maximum of six) and the recodification on the changes of information loss versus the disclosure risk variations in the dissemination of anonymized registers. These results should be systematically extented to both social and business (micro)data so as they could allow us to test the effectiveness of new μ-ARGUS features in the undergoing CASC project.