Reflections on Learning from Observational Data

The social sciences study various aspects of human behaviour – social, economic and political based on observational data. Observational data are inaccurate subject to simultaneity, seasonality, structural breaks, random variation too many interlocking variables masking the underlying causal patterns. During past two decades or so, use experimental (RCTs) has become widely popular across sciences, creating a tension between supporters critics new old methodologies. In this paper, we first review these methodologies, both experimental, focusing how economists other scientists try learn about relationships from correlations contained in We then reflect whether methodologies should be way forward purely statistical broader policy development perspectives.