The ABCs of Regression Discontinuity Designs and Instrumental Variables

About

The course aims to provide students with the necessary skills for critically understanding, managing, and analyzing data and designs that aim at causal identification. The course begins by introducing key concepts, benefits and pitfalls of longitudinal and panel data analysis, and an introduction to the counterfactual framework and selection bias. The entire course is built upon intuition-based teaching and pays special emphasis on using real-world examples and data to introduce students to different concepts. This includes implementing two classical causal inference designs: regression discontinuity (RDD) and instrumental variable (IV) design. The course includes frontal classes and practical sessions to give students hands-on experience working with real-world data. This will provide students with practical guides to then be able to implement and critically assess causal inference designs. We will pay special attention not only to each design’s assumptions but also to the different checks and threats that researchers should be implementing while using these designs.
This course is designed for researchers with a basic understanding of quantitative analysis but willing to delve into the various statistical and econometric techniques related to causal inference analysis. By the end of the course, researchers will be able to:

1. Understand and critically engage with terminology and concepts behind causal identification.
2. Understand and perform causal inference designs in Stata or R.
3. Critically assess checks and threats to causal inference designs.

The class structure will be a lecture followed by a lab session covering the materials taught in the theoretical session. Lab sessions consist of solving hands-on exercises covering each day’s topics. We will mainly replicate published papers on both sociology and political science and look at their caveats, but students are welcome to submit their data and work on that during the lab session. In fact, in the last lab session, students are welcome to submit and discuss their projects and we can cover practical extensions of other causal inference designs on demand.

Course leader

Mar C. Espadafor - Deputy Director at the Directorate for Public Policy

Fee info

Fee 700 EUR, for professionals