Detailed course description:
- Introduction to causal inference, quick history
- Potential outcomes, counterfactuals, basic definitions
- RCT, randomization, power calculation, selection bias, heterogenous treatment bias, DAG notation
- Propensity score matching, nearest neighbor covariate matching
- Instrumental variables, 2SLS estimation
- Weak instruments, Just-identified IV, overidentification
- Formula IVs
- Weighting estimators of the local average treatment
- Regression discontinuity
- Optimal bandwidth selection, Robust confidence intervals,
- Local randomization RD analysis, local polynomial order
- Quantile treatment effects in the RD
- Window selection based on covariates
- Spatial RD
- Regression discontinuity in time
- 2×2 Difference-in-differences, parallel trends, triple difference
- DID with covariates, two-way fixed effects (TWFE)
- Nonbinary response (binary, fractional, nonnegative), ratio in ratios, ratio in odds-ratios
- Differential timing, Bacon decomposition
- Group-time average treatment effects
- Imputation estimators, Mundlak estimator, fuzzy difference-in-differences
- Synthetic control
- Synthetic difference-in-differences
- Causal latent variable models
Practical classes are taught in R and Stata. Participants are asked to install the software before the course.
Schedule:
13.01.2025 - 16.01.2025
15:00-21:30 GMT+1(Warsaw, Madrid, Berlin, Rome)
9:00-15:30 EST (Boston, New York, Washington)
24 hours contact hours