D02: Causal Inference

This intensive five-day course provides a practical and conceptual foundation in causal inference, equipping participants to identify and estimate causal effects using real-world data. Beginning with regressions and experiments, students learn how to adjust for confounders and understand the logic of randomization. The course then progresses to matching methods for observational data, regression discontinuity designs that exploit treatment thresholds, instrumental variables for addressing endogeneity, and difference-in-differences approaches for analyzing policy impacts over time. Each day integrates theory, hands-on coding exercises in R, and applied case studies drawn from social policy, politics, and development research. Readings from Demystifying Causal Inference (Dayal, Murugesan) and Mastering ‘Metrics: The Path from Cause to Effect (Pischke, Angrist) reinforce learning and encourage critical thinking about research design. By the end, participants gain both the analytical tools and intuition needed to credibly assess causal relationships and apply empirical methods to complex policy and social science questions. Prior knowledge of statistics (particularly regression analysis) and R is highly recommended.

The course begins with a discussion of causality and identification, providing students with an intuitive understanding of cause-and-effect relationships. Then, each of the following four sessions delve into a specific method to improve causal inference, and the core assumptions of these techniques. In each case, we will apply the learned techniques to real-world applications, using actual case studies to illustrate the concepts under consideration. Hands-on exercises in R will help participants gain  skills, reinforcing concepts through data analysis.

By the end of the course, participants will be equipped to critically apply these learned concepts to the contexts of their own studies. Prior familiarity with basic statistics and R is recommended.

Learning Outcomes

• Understand the foundational methods of causal inference.
• Evaluate and interpret the techniques covered in this course.
• Apply causal inference techniques to real-world data and policy challenges.
• Implement these techniques in their own reasearch, using R.

Day 1: Regressions and Experiments 

The first day introduces the core ideas of causal inference, focusing on understanding how regression can adjust for confounders and the role of randomization in experiments. Participants learn to interpret causal graphs, estimate average treatment effects, and apply these methods to real-world social policy examples. Through practical exercises and guided readings, students build foundational skills for identifying causality using regression-based approaches.

Day 2: Matching Methods

Day two explores how observational data can simulate experimental conditions through matching techniques. Participants learn and apply methods such as exact, propensity score, and coarsened exact matching to estimate causal effects. Participants will also get hands-on practice creating balanced samples. The session emphasizes practical implementation and conceptual understanding, supported by hands-on exercises and readings.

Day 3: Regression Discontinuity Design (RDD) 

The third day focuses on Regression Discontinuity Design (RDD) as a quasi-experimental tool for identifying causal effects when treatment assignment is based on a cutoff. Participants examine key assumptions and potential pitfalls while applying RDD to real data. Through coding exercises and conceptual discussions, they learn how sharp treatment thresholds can generate credible causal estimates. Readings from 

Day 4: Instrumental Variables (IV)

Day four introduces Instrumental Variables, emphasizing how exogenous instruments help isolate causal relationships when treatment is endogenous. Participants explore the principles of exogeneity and exclusion restriction, applying these to study the historical impact of colonial institutions on development. Hands-on exercises reinforce how valid instruments can overcome omitted variable bias. Theoretical grounding and practical application will help building intuition for using IV in real-world contexts.

Day 5: Difference-in-Differences (DiD) 

The final day examines the Difference-in-Differences approach, a powerful tool for evaluating effects using variations across time and groups. Participants study the parallel trends assumption, robustness checks, and model extensions. Applying DiD to real world data, they learn to estimate treatment effects in dynamic policy settings. 

This description is subject to change at the discretion of the Instructor

2 Credits

For completion of all work before and during the course, as outlined by the Instructor, and 90% participation and attendance of the course.

2 Additional Credits

Course specific extra assignments. These can include submitting assignments before the course, daily assignments, and/or a final assignment to be completed after the course as decided by the Instructor.