Legal and technical means for protecting data in the production of statistics

Abstract

The increasing use of administrative records in statistics production creates special protection problems. Furthermore, the producers of statistics have to take into account psychological attitudes of the public.

The functions and applications of the methods used or known to protect statistical information from unintended disclosure, can be discussed under various aspects. Among them are legal aspects and techniques, the types of technical method used, and their relevance to the privacy discussion of the 60s and 70s.

The legal development should be observed carefully and efforts be made to ensure as high a level of protection as possible. As a rule, statistical records should be used exclusively for statistical purposes, and exceptions to this rule should be minimized. Legislation to protect privacy sometimes did not pay sufficient attention to the specific conditions of the statistical production process, in particular record linkage.

From the point of view of technical methods used, the protection of the statistical production apparatus and the transmission of data do not fundamentally differ from corresponding problems in other fields. However, the methods to prevent disclosures in tables constitute a specific statistical field. All such methods have to strike a balance between the risk of disclosure and the loss of information which they entail. The technical protection methods also include de-identification and randomized response.

The intensity with which the protection of statistics has been publicly discussed in several countries does not stem from actual cases of misuse. Nevertheless, the increased pressure for protection is a reality to be reckoned with, and the appropriate protection level has to be adapted to the subjective risk evaluation of the surrounding world as well. The producer of statistics has to guard, though, against exaggerated data protection that would unduly impinge on the information value of the statistical output. Legal data protection should be given increased attention as a high level of legal protection can replace costly technical protection.