ABSTRACT
Out of reach for most of the 20th century, microdata have started to play a key role in research. From now on, this will intensify. The chapter explains why microdata weren’t prominent for so long, and how macro indicators came to dominate the statistical landscape of the past. Then, it articulates the changes underway and offers plenty of examples of how microdata have started to be used in research, notwithstanding confidentiality and privacy issues. Microdata can generate powerful insights by bringing socio-economic research as close to randomized experiments as possible, and specific new applications to that effect are discussed. The same is true for research around margins, which had not been possible up to now. The discussion extends to the role of microdata in statistical registers, which are conceptualized as ‘funnels’ capable of absorbing all kinds of data from multiple sources and growing in breadth, in depth, and over time becoming valuable information sources. Fact-checking tips discuss that microdata are absolutely essential in explaining differences in macro movements, and the use of alternative data that can substitute for microdata, such as synthetic data and data generated through machine learning.
