A Study on the Re-Identifiability of Dutch Citizens
Matthijs R. Koot, G.J. van t Noordende, Cees de Laat, A. Serjantov, C. Troncoso · UvA-DARE (University of Amsterdam) · 2010
Abstract. This paper analyses the re-identifiability of Dutch citizens by various demographics. Our analysis is based on registry office data of 2.7 million Dutch citizens, ∼16 % of the total population. We pro-vide overall statistics on re-identifiability for a range of quasi-identifiers, and present an in-depth analysis of quasi-identifiers found in two de-identified data sets. We found that 67.0 % of the sampled population is unambiguously identifiable by date of birth and four-digit postal code alone, and that 99.4 % is unambiguously identifiable if date of birth, full postal code and gender are known. Furthermore, two quasi-identifiers we examined from real-life data sets turn out to unambiguously iden-tify a small fraction of the sampled population. As far as we are aware, this is the first re-identifiability assessment of Dutch citizens that uses authoritative registry office data. Key words: re-identification, data anonymity 1