Disclosure risk for microdata stemming from official statistics |
| |
| Authors: | U. Blien H. Wirth M. Muller |
| |
| Affiliation: | Institute für Arbeitsmarkt- und Berufsforschung der Bundes-anstalt für Arbeif (IAB - institute for Employment Research of the Federal Labour Office) Regensburger Str. 704 P8500 Nürnberg 30 Federal Republic of Germany;Zentrum für Umfragen, Methoden und Anatysen (ZUMA - Center for Survey, Methodology and Analysis) 82, 1 06800 Mannheim Federal Republic of Germany;University of Mannheim Fakultät für Soziologie (Lehrstuhl Prof. Walter Muller) A5 D-6800 Mannheim Federal Republic of Germany |
| |
| Abstract: | In the research project on data anonymity, the possibilities and difficulties of restoring the identity of respondents, whose data have been anonymized, were tested in realistic simulations. In this paper the results of an application of a) a matching procedure and of b) a method based on discriminate analysis are reported. In the experiments carried out, empirical data of a handbook about German scientists and scholars and the German microcensus were used. A check of the results by an independent data-trustee demonstrated that a real intruder has more difficulties achieving an identification than is frequently assumed. |
| |
| Keywords: | data confidentiality data anonymity privacy protection. official statistics discriminate analysis sample surveys |
|
