Statistically assessing the causal impact of development policies has now become an extremely large industry. While there will always be five available identification strategies, doing things right in a policy-relevant manner is neither obvious, nor easy. The literature on impact evaluation is a subset of econometrics, sometimes with a vocabulary of its own. As such, econometric methods that you have learnt will figure prominently in what follows. There is no textbook for this course. The course will be arranged around a selection of readings that touch on various technical aspects of impact evaluation, as well as applications. Roughly one third of class time will be devoted to showing how to do the stuff mentioned in the readings using R, and you will be provided with the code shortly after the class, so that you can experiment on your own. An important component of this course is embedded in the problem sets, which mostly revolve around doing geeky stuff in R. There will be 3 problem sets, a midterm and a final.