Objective Observational Study Design Using Big Data for Causal Inference
*Donald Rubin, Harvard University 

Keywords: Objective Observational Study

The principles of the objective design of observational studies do not depend on the size of the data sets being studied, but rather follow well-established principles that have evolved from the design of proper randomized experiments. Of course situations involving observational studies are typically more complex than those involving randomized experiments, and larger data sets create more opportunities for innovative and clever analyses as well as opportunities for inapposite data-dredging exercises. The focus of this presentation will be on practical advice to implement the former while avoiding the pitfalls of the latter.