An Empirical Comparison Of Privacy-Protection Strategies For Geocoded Microdata
Maike Steffen* 1, Jörg Drechsler1, Milan Quentel1
Abstract
In response to an increasing need for detailed geographic information in various research areas, data-collecting institutions are increasingly geocoding their microdata. However, access for external researchers is typically very limited due to confidentiality concerns. In the literature, a variety of confidentiality-protection strategies have been proposed with the aim of enabling the dissemination of geodata while balancing the trade-off between privacy protection and utility preservation.
The objective of this study is to empirically compare and evaluate different confidentiality-protection strategies such as spatial aggregation, geographic masking techniques, and the generation of synthetic geodata.
These strategies are applied to a subset of the Integrated Employment Biographies, a large administrative database gathered at the Institute for Employment Research. The resulting datasets are evaluated with regard to their utility using, for example, methods to assess the spatial patterns and regression results across the datasets. Lastly, different risk measures are employed to quantify the remaining risk of disclosure.
*: Speaker
1: Institut für Arbeitsmarkt- und Berufsforschung