Dustin Gilbreath: Selection bias, matched sampling, and improving second best options in impact evaluation
While randomized control trials remain the gold-standard in impact evaluation, in many cases they are unlikely to be carried out, whether feasible or not: try to convince the Minister of Infrastructure in your country to put irrigation systems in a random location. To deal with this a variety of silver standard impact evaluation techniques are regularly used. In this talk, Dustin will describe how to improve these at the data collection stage through matched sampling, which has significant advantages in reducing selection bias.
Dustin Gilbreath is the Deputy Research Director of CRRC-Georgia, a data-focused think tank in Tbilisi, Georgia. He is responsible for quasi-experimental and experimental research design; oversight of evaluation projects; and policy analysis. Dustin has worked on state capacity, public transport safety, election forensics and administration, polling and society, and think tanks. His writing has been featured in the Washington Post, among other international media outlets. He holds an MA from the Department of Political Science at CEU.