After Sessions 5–6 · 55–70 minutes
Session 5 conceptual questions 1, 3 and 4; Session 6 conceptual question 3 and the applied adoption-grid exercise. Add the Week 5 R contrast and a five-minute interpretation check.
Course Content
Each session combines conceptual reasoning, empirical design, and a short base-R exercise. The topic numbers below follow the course programme. Use the combined teaching decks and the linked homework and R starters for each topic.
Which slides should I use? Sessions 2–3 have a continuation deck, building on the selection and matching background linked under Session 2. Sessions 5–6 share one teaching deck, and Sessions 7–8 share another. Revised topic-specific decks for Sessions 5–8 are available as reference material.
Econometric Refresher: Intuitions for Applied Work
Required RefresherQuestion: Which statistical ideas help us read an empirical evaluation?
Concepts: distributions, expectations, conditional means, regression, sampling uncertainty, identification and independence.
Recommended reading:
- Angrist, J. D., & Pischke, J.-S. (2010). The credibility revolution in empirical economics: How better research design is taking the con out of econometrics. Journal of Economic Perspectives, 24(2), 3–30.
Foundations of Policy Evaluation
Session 1Question: What exactly are we trying to learn when evaluating a policy?
Concepts: causal question, treatment, outcome, population, time horizon, potential outcomes, estimand, selection bias, internal and external validity.
Recommended paper:
- LaLonde, R. J. (1986). Evaluating the econometric evaluations of training programs with experimental data. American Economic Review, 76(4), 604–620.
Selection Bias, Regression, Matching, and Inference
Session 2Question: Can conditioning on observables reconstruct the missing counterfactual?
Concepts: naive comparisons, conditional independence, bad controls, propensity score, common support, balance diagnostics, clustered inference.
Recommended paper:
- Dehejia, R. H., & Wahba, S. (2002). Propensity score-matching methods for nonexperimental causal studies. Review of Economics and Statistics, 84(1), 151–161.
Randomised Experiments: Design and Power
Session 3Question: What does randomisation solve, and what problems remain?
Concepts: assignment mechanism, balance, implementation threats, minimum detectable effect, cluster design effect, experimental protocol.
Recommended paper:
- Miguel, E., & Kremer, M. (2004). Worms: Identifying impacts on education and health in the presence of treatment externalities. Econometrica, 72(1), 159–217.
Compliance, IV, LATE, and External Validity
Session 4Question: Can exogenous variation in treatment identify effects when treatment is endogenous?
Concepts: encouragement design, compliers, always-takers, never-takers, defiers, relevance, independence, exclusion, monotonicity, weak instruments, external validity.
Recommended paper:
- Angrist, J. D. (1990). Lifetime earnings and the Vietnam era draft lottery: Evidence from Social Security administrative records. American Economic Review, 80(3), 313–336.
Difference-in-Differences: 2x2 and TWFE
Session 5Question: Can untreated trends reveal the missing counterfactual trend?
Concepts: before-after comparison, treated-control comparison, ATT in the post period, parallel trends, TWFE, serial correlation, clustering.
Recommended paper:
- Card, D., and Krueger, A. B. (1994). Minimum wages and employment: A case study of the fast-food industry in New Jersey and Pennsylvania. American Economic Review, 84(4), 772–793. Author-hosted paper.
Event Studies and Staggered DiD
Session 6Question: What changes when treatment timing and treatment effects are heterogeneous?
Concepts: omitted period, dynamic effects, anticipation, weak support, never-treated and not-yet-treated controls, ATT(g,t), aggregation.
Methodological readings:
- Sun, L., and Abraham, S. (2021). Estimating dynamic treatment effects in event studies with heterogeneous treatment effects. Journal of Econometrics, 225(2), 175–199. Author manuscript.
- Callaway, B., and Sant’Anna, P. H. C. (2021). Difference-in-differences with multiple time periods. Journal of Econometrics, 225(2), 200–230. Author manuscript.
Regression Discontinuity
Session 7Question: Can an institutional threshold create a locally credible comparison?
Concepts: running variable, cutoff, bandwidth, local linear regression, density checks, covariate continuity, placebo cutoffs, fuzzy RDD, local IV estimand.
Empirical applications:
- Angrist, J. D., and Lavy, V. (1999). Using Maimonides’ rule to estimate the effect of class size on scholastic achievement. Quarterly Journal of Economics, 114(2), 533–575. University-hosted paper.
- Lee, D. S. (2008). Randomized experiments from non-random selection in U.S. House elections. Journal of Econometrics, 142(2), 675–697. Author-hosted paper. Optional close-election application.
Synthetic Control, Heterogeneity, and Replication
Session 8Question: How do we synthesise, interpret, and challenge causal evidence?
Concepts: donor pool, predictors, weights, pre-treatment fit, treatment gap, placebo tests, heterogeneity, transport, robustness, replication audit.
Combined teaching slides 7–8
SCM and synthesis reference slides
View optional homework
Download .md
R script
Recommended paper:
- Abadie, A., Diamond, A., and Hainmueller, J. (2010). Synthetic control methods for comparative case studies: Estimating the effect of California’s tobacco control program. Journal of the American Statistical Association, 105(490), 493–505. Author-hosted paper.
Replication case: McCrary’s comment (2002) and Levitt’s reply (2002) on electoral-cycle instruments, police hiring and crime.
Practice After the Combined Classes
Use one selected route after each combined class, rather than automatically completing two full homework sheets. Full sheets remain available for revision.
After Sessions 7–8 · 55–75 minutes
Session 7 conceptual questions 1, 3 and 5 and its applied threshold exercise; optional Session 8 conceptual question 2. Add the Week 7 R bandwidth and fuzzy-RDD calculations. The full Session 8 assignment remains optional.