Methodological challenges of register‐based research

Abstract

Administrative data have become more and more important for both official statistics and academic research. However, the, more administrative registers are used, the more methodological issues arise (Bakker, 2009; Daas, Kuijvenhoven and Zeelenberg, 2010a). This was apparent at the 2011 congress of the International Statistical Institute, when more than ten sessions focused explicitly on the quality, linking and analyses of administrative data and many others discussed studies based on administrative data. Moreover, the increasing use of administrative data is one of the core transitions in research projects funded in Europe. In both the European Statistical System networks (ESSnets), launched by Eurostat, and the Framework Programmes, funded by the European Commission, the primary focus is no longer solely on data collected by questionnaires. Both Eurostat and the European Commission promote the development of new theory and applications in the area of the use of administrative data. Examples of recent initiatives are the ESSnet on ‘Data Integration’, the ESSnet ‘On the Use of Administrative and Accountants Data’ and the seventh framework Programme BLUE Enterprise and Trade Statistics. Each of these projects has a strong component that focuses on key methodological issues in the use of administrative data and registers for research purposes. However, the theory of register-based research is scarcely out of the egg. After the first mention of this problem in the mid-1990s (Hoffmann, 1995; Statistics Denmark, 1995), the rudiments of such a theory were merely sketched in the first decade of this century (UNECE, 2007; Wallgren and Wallgren, 2007). It will take a serious period of time to develop it further and take it to an acceptable level; just as it took time to develop and accept the use of probabilistic ideas in to the