Privacy Preserving Record Linkage in the Context of a National Statistical Institute

Abstract

Linking databases containing records of the same person is an increasingly used technique within national statistical institutes (NSIs). No unique personal identification number is available for linkage in many countries (such as New Zealand, Australia, the UK, and Germany). In such cases, the linkage is most often based on quasi-identifiers such as surname, first name, address, and place of birth. Since knowledge of this kind of identifier is widely seen as private information, different methods for Privacy-Preserving Record Linkage (PPRL) have been developed. The report reviews existing methods for PPRL, considers privacy vs linkage quality of record linkage within an NSI, discusses the kind and probability of privacy attacks and gives recommendations for linkage practice in official statistics.

Keywords: Official Statistics, Record-Linkage, PPRL, privacy, five safes