Dr Mylene Lagarde COW.3.02

This course is compulsory on the Executive MSc in Evaluation of Health Care Interventions and Outcomes, in collaboration with NICE. This course is not available as an outside option.

Randomized trials have long been used in the clinical world to test the efficacy of drugs and other medical treatments. Recently, social scientists have started using the same approach, using random assignment to allocate resources or implement a policy intervention differently to different groups, in order to determine the causal effects of the policy of interest. The popularity of randomized evaluations has grown especially, but not exclusively, among researchers and policymakers in low- and middle-income settings.

Conducting a successful randomized evaluation involves many inter-related steps and a good understanding of a few statistical concepts. Randomized evaluations also usually require to design and organise the data collection of relevant and useful information, which involves a number of steps to avoid critical pitfalls.  It is therefore essential to understand these different steps to design and implement randomised evaluations adequately, or to be able to critically analyse them.

This course proposes a hands-on and intuitive approach to designing and conducting a randomised evaluation. In the first half of the course, we will discuss reasons for undertaking randomised evaluations; how to design the randomised experiment to ensure it answers the question(s) of interest (including issues of statistical power and sample size calculation); how to deal with threats to randomisation (e.g. attrition, spill-overs). In the second half of the course, we will discuss practical issues raised by primary data collection, including how to best measure outcomes of interest; how to design good tools and how to conduct and manage fieldwork. We will also discuss the ethics of randomised policy evaluations.

The course will be a mix of lectures and small-group discussions and exercises in seminars.

Seminars will be designed to encourage students to critically engage with the topics and apply the technical skills taught. Each seminar will be closely aligned with the lecture content to give students the opportunity to apply the new knowledge. Seminars will be built around a group project defined at the beginning of the week: students in a group will gradually build the design of their randomised evaluation and plan the different components, working towards a final product which will form the basis of their evaluation. This will ensure that students maintain their engagement throughout the course, but also apply their skills immediately.

The course will be delivered in no less than 24 hours and consists lectures and seminars.

Students will be expected to produce 1 piece of coursework in the ST.

Students will be asked to submit a short 1,000 word draft protocol after the end of the course. The draft protocol will serve as an outline of the final research protocol, which will account for 100% of the final mark. 

While some aspects of their work may still be work in progress (e.g. using bullet points), students will be expected to write up one section of their protocol in a more detailed way. The outlines will be graded and feedback given to students. Students will be able to use this feedback in their writing of the summative work.

• Glennerster, R., & Takavarasha, K. (2013). Running Randomized Evaluations (STU - Student edition ed.): Princeton University Press.

• Gerber, A. S., & Green, D. P. (2012). Field Experiments: Design, Analysis and Interpretation. New York, NY: Norton.

• Ustun, T. B., Chatterji, S., Mechbal, A., & Murray, C. J. L. (2005). Quality assurance in surveys: standards, guidelines and procedures. In W. H. S. W. Collaborators (Ed.), Household Sample Surveys in Developing and Transition Countries.

• Glennerster, R. (2017). The Practicalities of Running Randomized Evaluations: Partnerships, Measurement, Ethics, and Transparency. In E. Duflo & A. Banerjee (Eds.), Handbook of Field Experiments: North Holland.

• Dupas, P., & Miguel, E. (2017). Impacts and Determinants of Health Levels in Low-Income Countries. In E. Duflo & A. Banerjee (Eds.), Handbook of Field Experiments: North Holland.

Research project (100%) in the post-spring term.

1 x Research protocol (3,000 words max) to plan the randomised evaluation of a particular health programme – 100% of grade. 

Students will be asked to pick one programme/policy from a proposed list. They are also given the option to choose their own intervention (after discussion with their seminar leader) if they would prefer.