German Microdata Lab

Anonymization and Data Protection

Although access to official microdata was dealt with relatively liberally until the late 1970s, a fact reflected in the intensive use of these data by various research projects (e.g. DFG Sonderforschungsbereich 3 or a project entitled Vergleichende Analyse der Sozialstruktur mit Massendaten [The Comparative Analysis of Social Structure using Large-Scale Data sets] aka VASMA) and in informative research results, exceedingly restrictive access regulations were introduced in the context of the debate on data protection that commenced in the early 1980s. It was not until the so-called ‘census judgment’ by the Federal Constitutional Court, which explicitly recognized the research community’s need to use official microdata, that data access regulations – still restrictively defined in the 1980 Federal Statistics Law – were somewhat relaxed. The 1987 Federal Statistics Law included a so-called science clause, according to which official microdata were authorized for transfer to the research community, provided that the re-identification of those surveyed could only be accomplished at disproportionately excessive costs (concept of factual anonymity). However, it was not possible directly to adopt this statutory ruling in practice, since at that time, no knowledge was available about central aspects of factual anonymity. In this context, the microdata department carried out a joint research project with the Federal Statistical Office and the Chair for Applied Social Research (University of Mannheim; project manager: Prof. Dr. Walter Müller). One of the key outcomes of this project was recognizing that when using real data re-identification of individual data sets is not nearly as simple as frequently assumed. In this instance, a quasi-‘natural’ protective factor was, above all, represented by information incompatibilities originating from different data sources. This made it possible to develop transfer recommendations offering the maximum possible protection against re-identification risks while at the same time maintaining maximum potential for analysis of these data. The results of the anonymization project at a national level meant a significant breakthrough for the transfer of anonymized microdata; at an international level, they gave a new impetus to anonymization research.

Wirth, H., 2003: Angriffsszenarien auf wirtschaftsstatistische Einzeldaten - ein Überblick, S. 11-24. In: Forum der Bundesstatistik, Band 42. Stuttgart: Metzler-Poeschel.

Wirth, H., 1992: Die faktische Anonymität von Mikrodaten: Ergebnisse und Konsequenzen eines Forschungsprojektes. ZUMA-Nachrichten 30: 7-42.

Müller, W./Blien, U./Knoche, P./Wirth, H. u.a., 1991: Die faktische Anonymität von Mikrodaten, Bd. 19 der Schriftenreihe Forum der Bundesstatistik. Statistisches Bundesamt (Hrsg.). Stuttgart: Metzler-Poeschel.   

Lüttinger, P./Wirth, H., 2004: Zur Weitergabe von älteren amtlichen Mikrodaten an die Wissenschaft. Allgemeines Statistisches Archiv, Heft 88/4: 473-486.

Müller, W., Blien, U. & Wirth, H., 1995: Identification Risks of Microdata. Evidence from experimental studies. Sociological Methods & Research, 24 (2): 131-157.

Bender, S., Hilzendegen, J. & Schimpl-Neimanns, B., 1995: Die IAB-Beschäftigtenstichprobe: Eine neue Datei für die Arbeitsmarktforschung. ZUMA-Nachrichten 36: 122-129.