Interviewing administrative records. A conceptual map for the use of big data for economic research

Abstract

Businesses, academia and official statistics are turning more and more to novel data sources besides traditional sampling surveys, but the debate about their defining features and the challenges they pose for research is still open. In this paper I propose a conceptual map of what data are in the field of empirical economic research to clarify what are the conditions and possible strategies to fully grasp their opportunities, particularly in the case of big data of administrative origin. The conceptual map is inserted into a recent literature addressing the clarification of the very notion of data, and exemplified using three well-know cases of failures and best practices in the use of large data samples. The conceptual map is then used to discuss the case of labour market research based on social security data.