Causal Inference in the Social Sciences I: Randomised Controlled Trials and Matched Designs

This one-week course provides a thorough introduction to the most widely used methods of casual inference, also known as program evaluation. This is an applied methods course, with equal emphasis on methods and their applications.

The target audience is practitioners and would-be practitioners of policy evaluation who want to measure the effects of policy interventions and policy changes. We especially recommend the course to researchers in social sciences (PhD students and young scholars) interested in measuring the effect of interventions and policy changes.

The knowledge and skills developed in this course are essential to understand and effectively participate in policy evaluations. They are necessary for anyone who works with the results of policy evaluations: those who commission evaluations and want to judge feasibility and methodological soundness of competing proposals, and those who use the results of evaluations and want to know the validity and reliability of their results.

The ideal student has some statistical background and is not afraid of working with data but does not have a firm methodological background in program evaluation or econometrics.

The course provides up-to-date methodological knowledge, together with the corresponding intuition and related skills in handling software, interpreting results and presenting those results.

By the end of this course, you should be able to:

• understand the methodology of an evaluation study, including its fine details, and form a well-grounded judgment about its value
• judge whether the evaluation was carried out in a methodologically sound way given the circumstances, and form an educated opinion on how much credibility one should give to its conclusions
• design – or effectively participate in a team that designs – an evaluation study using the methods covered: randomised experiments, matching, regression analysis
• perform an impact evaluation analysis – or effectively participate in a team performing such an evaluation – using the above methods
• interpret the results of analyses of experimental, matching-based or regression-based policy evaluations
• use the methods covered in this course in other contexts.

During the course, I will:

• provide a thorough technical introduction to the methods but emphasise also intuition and practical aspects
• show the methods in action by going over published evaluations
• replicate some of the published evaluation results using original data, check their robustness and dig deeper if possible.

Methods covered

1. Randomised Controlled Trials – intent to treat effects, imperfect compliance, local average treatment effects, encouragement design, power calculations
2. Matching – exact matching, matching on the propensity score
3. Linear Regression – ensuring common support, functional form choices; regression versus matching

Day 1 – Introduction to impact evaluation
Why do we need to evaluate? Identifying the counterfactual. Getting the question right and formulating hypotheses. Causal inference versus association. The potential outcomes framework. An example: The Job Corps program in the US Heterogenous Treatment Effects. Average Treatment Effects.

Day 2 – Randomised Control Trials (RCT)
Intro to RCT. Why does randomisation help? Randomisation in Practice. Examples. (The NSW program in the US; PROGRESA in Mexico). Intent to treat effects. Imperfect compliance. Local average treatment effects

Day 3 – RCT & Matching
RCT regressions. Intro to matching: Unconfoundedness. The Common support. Exact matching.

Day 4 – Linear Regressions & Matching
Simple linear regression. Ensuring common support. Functional form choices. Regression versus matching. Matching on the propensity score. Examples in matching. Teacher training in Israel.

Day 5 – Matching
Examples in matching. (The New Deal for Lone Parents programme in the UK; debate based on the NSW program in the US). Student presentations.

Basic knowledge of statistics and regression analysis at an undergraduate level, and some knowledge of Stata or R.

Each course includes pre-course assignments, including readings and pre-recorded videos, as well as daily live lectures totalling at least two hours. The instructor will conduct live Q&A sessions and offer designated office hours for one-to-one consultations.

Please check your course format before registering.

Online courses

Live classes will be held daily for two hours on a video meeting platform, allowing you to interact with both the instructor and other participants in real-time. To avoid online fatigue, the course employs a pedagogy that includes small-group work, short and focused tasks, as well as troubleshooting exercises that utilise a variety of online applications to facilitate collaboration and engagement with the course content.

In-person courses

In-person courses will consist of daily three-hour classroom sessions, featuring a range of interactive in-class activities including short lectures, peer feedback, group exercises, and presentations.

Disclaimer

This course description may be subject to subsequent adaptations (e.g. taking into account new developments in the field, participant demands, group size, etc.). Registered participants will be informed at the time of change.

By registering for this course, you confirm that you possess the knowledge required to follow it. The instructor will not teach these prerequisite items. If in doubt, please contact us before registering.