Abstract

The Institute for Employment Research (IAB) provided the Institute for the Study of Labor (IZA) with comprehensive administrative data of the IAB and the German Federal Employment Agency valuable to research projects focussing on an evaluation of labor market policies. In addition, IZA conducted a supplementary survey. By linking these administrative and survey data we can create a dataset with manifold information on individuals that is quite groundbreaking in empirical evaluation research. However, the linked dataset entails social data and is subject to the strict regulations of §75 SGB X (German Social Code, book 10), which so far limits data access to researches at IZA. Considering the data’s high potential value to scientific evaluation research, a joint project of IZA and IAB aims at generating anonymised data that will be made available to researchers as a Scientific Use File (SUF).

The main challenge of the project grounds in the variety and duplicity of variables in the linked dataset (the latter meaning that survey and administrative data each contain some more or less identical variables). For that reason, finding a procedure to anonymise the data is nontrivial. Our research project is considered explorative and its results could have pioneering implications for the provision of linked datasets in scientific research.