Introductory courses in causal inference rely on assumptions strong assumptions.
This course shows what can be learned when these assumptions are relaxed, how some of them can be tested, and how to better match the causal effects that we learn in a study to the policy question of interest.
TOPICS COVERED:
Partial identification
Testing instrumental-variables assumptions
Policy-relevant treatment effects
Spillovers from one participant's treatment to another’s outcome
Software Requirements: R (with the tidyverse); RStudio recommended
Who is the Course for?
Master’s and doctoral students, academic researchers, and economists in government, central banks, and international organisations who are broadly familiar with the methods covered in Angrist and Pischke’s Mostly Harmless Econometrics, particularly instrumental variables. Participants should have a working knowledge of R. Refresher materials are available at treatment-effects.com/beyond.
Choose to follow the Econometrics Pathway to deepen your knowledge of Econometrics with courses in Treatment Effects (morning session) and Machine Learning (afternoon session). Or mix and match, choosing different morning or afternoon sessions from our other pathways: Applied Economics or Macroeconomics.