ECMA 35550 1 (Autumn 2021) The Practicalities of Running Randomized Control Trials
This course will meet on Mondays and Wednesdays from 8-9:20am
Course Description: This graduate level course is designed for those who plan to run a randomized control trial. It provides practical advice about the trade-offs researchers face when selecting the type of randomization technique to use, the number of clusters and people per cluster to include, the content of survey instruments, which analytical techniques to use and much more. How do you choose the right minimum detectable effect size for estimating the sample size needed to run a high quality RCT? How do you quantify difficult to measure outcomes such as women’s empowerment or ensure people are providing truthful answers when you are asking questions on sensitive topics like sexual health? When should you tie your hands by pre-committing to your analysis plan in advance, and when is a pre-analysis plan not a good idea? This course will draw on lots of examples from RCTs around the world, most (though not all) from a development context. Alongside field tips, it will also cover the concepts and theory behind the tradeoffs researchers face running RCTs.
The course is designed for PhD students but given its practical nature is open to and accessible to masters students who plan to work on RCTs.
Prerequisites: Students will need to have done or be in the process of taking advanced econometrics at Masters or first year PhD levels. This will not be a technical econometrics class, but a firm understanding of econometric concepts such as selection bias and statistical inference are essential and will not be covered in this course.