Florian Foos, London School of Economics
Andrea Junqueira, Texas A&M University
“Correlation is not causation”. You must have heard this warning many times. But, what then is causation, and how can we test causal hypotheses, and identify the effects of policies and programs? Starting from the ideal of a randomized experiment, the module introduces participants to the design-based approach to causal inference, based on the potential outcomes framework. Instead of using fancy modelling techniques to correct post-hoc for potential biases, the module encourages students to think about challenges to causal inference at the design stage of a study. Published work will be evaluated based on how it addresses three key assumptions underlying causal inference: independence, excludability, and non-interference.
After introducing students to the design-based approach to causal inference in general, the module will cover the design, conduct and analysis of randomized field experiments, in particular. The goal of the course is to provide participants with the methodological knowledge and the practical skills to design, analyse, and eventually conduct their own field experiments. The module is taught as a combination of lectures, seminars and applied computer labs.
There is relatively little assumed knowledge, and the aim is to build the statistical foundations from the ground up. The only pre-requisite is a course covering linear regression and hypothesis testing. For those with a deeper statistical background, there will be opportunities for exploration of more advanced topics. Students are free to use either R or Stata.