Advanced Topics in Applied Empirical Methods: "Causality in Cross Sections"

I. Syllabus:

The lecture presents strategies for estimating the effect of a binary treatment on some outcome when data are observational. For the identification of this (causal) effect, focus is on the Neyman-Rubin potential outcomes framework and, in particular, the widely-used Propensity Score (PS) techniques. Judea Pearl's causal diagrams serve as an introduction. Some auxiliary methods -like spline fitting, and regularized regression for covariate selection- are also discussed.

  1. From Randomized Experiments to Observational Data
  2. Causal Diagrams
  3. Causality Frameworks
  4. The Potential Outcomes Framework
  5. The Propensity Score I (theory and estimation)
  6. The Propensity Score II (regularization)
  7. Weighting
  8. Matching I: Exact, Nearest-Neighbor, Caliper and Radius
  9. Matching II: Kernel and Optimal Balance
  10. Stratification and Propensity Score Covariates
  11. Difference in Differences and Regression Discontinuity


II. Prerequisites:

  • tba


III. Downloads:


IV. Literature:

  • Imbens, G. W. & Rubin, D. B. (2015), Causal Inference for Statistics, Social, and Biomedical Science. Cambridge University Press, New York, NY.)
  • Pearl, J. (2009), Causality: Models, Reasoning and Inference. Cambridge University Press, New York, NY.
  • Morgan, S. L. & Winship, C. (2015), Counterfactuals and Causal Inference. Cambridge University Press, New York, NY.
  • Guo, S. & Fraser, M. W. (2014), Propensity Score Analysis. Sage Publications, Thousand Oaks, CA.


VI. Lecture:

VII. PC-Tutorial:

Access to the computer lab requires one-time registration with a Stu-Account.

The first sessions of the tutorial reinforce the intuition of causal diagrams and review the mathematics of the potential outcomes framework. In later sessions, simulations are carried out in R to explore the properties of the PS weighting, matching and stratification estimators, as well as two other estimation strategies that are popular in the econometric literature: difference in differences and regression discontinuity.

  1. Introduction to R
  2. Causal Diagrams
  3. The Potential Outcomes Framework
  4. Propensity Score Estimation
  5. Weighting)
  6. Matching
  7. Difference in Differences and Regression Discontinuity