Abstract | The National Educational Panel Study surveys a vast amount of information about individuals as well as organizational units, such as schools and universities. In addition to the voluminous questionnaire, the panel structure significantly increases data size with every new wave. Because of this data abundance, the de-anonymization and (re-)identification of singular units seem to be a major problem when disseminating data to third parties. While de-anonymizing institutions per se is undesired and goes against scientific ethics, it also aids in the task of tracking individual persons. Such an event, however, is believed to be a significant problem in terms of data privacy, federal law, and respondents’ agreement. This obviously concerns national statistics institutes and federal data centers, which hold vast amounts of data, often without any firsthand agreement of the affected citizens. For this reason, researchers have rather sparsely reflected on motives of possible attackers and instead focused on anonymization (and even perturbative methods) and access technologies. This may not hold for the social sciences; however, instead of remaining astounded by the anxiety fed by a hardliner’s understanding of data privacy, which leads to irrational procedures of data anonymization and seriously limits and harms scientific research, one should ask: Who would be interested in such an attempt ? Which realistic options really exist for such an attacker ? And what benefit may be gained ? By discussing these trivial but difficult-to-answer questions, it is possible find a realistic and moderate way to secure individual content while maintaining a solid scientific database. This article tries to shed light on which realistic disclosure risks the NEPS has to deal with and which are merely theoretical constructs. After a brief summary of definitions and common statements, possible assault scenarios are discussed and benchmarked. This process explicitly involves the comparison of expected gain with necessary expenses. The main part of this work explores the framework established and the NEPS’ Scientific Use Files (SUF) with regard to feasible re-identification and potential profit. The text concludes by presenting the developed anonymization methods applied. (Orig.). |