<ul><li>Article discusses Regression Discontinuity Design (RDD) as a method for causal inference in scenarios where traditional experimental methods are not feasible.</li><li>RDD exploits cutoffs on a 'running' variable to estimate causal effects, assuming continuity holds.</li><li>The article explains the core assumption of continuity using examples like legal drinking age laws.</li><li>It emphasizes the importance of maintaining continuity to ensure the validity of RDD.</li><li>The article delves into instrumental variables and the front-door criterion in RDD to identify causal effects.</li><li>Practical application of RDD is illustrated in the context of e-commerce listing positions and their impact on performance.</li><li>Modeling choices in RDD, such as parametric vs. non-parametric approaches, polynomial degree, and bandwidth, are discussed.</li><li>Placebo testing is highlighted as a method to validate results, along with the importance of continuity assumption and density continuity testing.</li><li>The article concludes by stressing the careful application of RDD and provides additional resources for further learning.</li></ul>

Regression Discontinuity Design: How It Works and When to Use It

Discover more