My research focuses on causal inference and impact evaluation. I use empirical data from a variety of contexts to answer questions about internal and external validity. Some questions which I am trying to answer include: under which conditions non-experimental methods (e.g. matching, difference-in-difference, regression discontinuity design) work well, and how can we successfully predict the effect of a program in a new area. To answer these questions I try to use many datasets and techniques from econometrics and machine learning.
- Estimating long-term treatment effects without long-term outcome data. 2020. GPI Working Paper No. 10-2020
Work in progress
- How biased are observational methods in practice? Accumulating evidence using randomised controlled trials with imperfect compliance.
Co-authors: Gharad Bryan, Sylvain Chabé-Ferret, Jonathan de Quidt, Greg Fischer, Jasmin Fliegner & Roland Rathelot
- Forecasting impacts: accuracy over time