Administrative Data and Replicable Science: Does It Fit?

• Activity Number: 234
• Type: Topic Contributed
• Date/Time: Monday, August 5, 2013 : 2:00 PM to 3:50 PM
• Sponsor: Business and Economic Statistics Section
• Author(s): Stefan Bender*+ and David Schiller
• Companies: IAB (Institute for Employment Research) and IAB
• Keywords: administrative data ; research data center ; replication ; confidential data

Abstract

The generating process of administrative data is in most cases quite different to the generation of surveys data. Admin data are not generated for researchers needs; it is usually created in the workflow of admin processes. Most of the data are quite huge and there is little knowledge about the data quality. In addition there is a lack of knowledge on the researchers side, how to handle and analyze admin data. Due to a high disclosure risk access to admin data is restricted and in most of the cases only possible through research data centers (RDCs). According to that RDCs have to deal with this situation by finding (new) solutions for generating an environment, which supports researchers. Thereby the environment must support the whole data life cycle from data generating to the archiving of data, programs and results. To enable replication studies this environment has to include several dimensions like identifiers for the data, archiving solutions and tools for replication including data documentation, program and knowledge sharing and so on. We will show new developments in a German RDC, which try to fulfill researchers' and replication standard needs.