Part II · Causal Frameworks & Experiments
Module 06. Selection on Observables: Matching, IPW & Propensity Scores
When treatment depends only on observables, rebuild the experiment by matching and reweighting.
Welfare economics: LaLonde's NSW vs CPS and targeted assistance programmes about 6 hours3 lessons
Module progress0/3 lessons · 0%
Lessons
6.1
Selection on Observables and the Propensity Score
- · State the two identifying assumptions: conditional independence (unconfoundedness) and overlap
- · Rebuild LaLonde's test: the NSW trainees against a survey (CPS) comparison group
- · Estimate a propensity score with a logit and check overlap visually and numerically
120 min 6 Colab cellsVideo coming soon
6.2
Matching: Propensity Score and Mahalanobis Distance
- · Implement nearest-neighbour propensity score matching for the ATT
- · Implement Mahalanobis (covariate) matching and compare it with score matching
- · Check balance after matching with standardised differences
120 min 6 Colab cellsVideo coming soon
6.3
Inverse Probability Weighting, Trimming and Diagnostics
- · Explain how inverse probability weights create a pseudo-population in which treatment is random
- · Estimate the ATE and the ATT with IPW, using normalised (Hájek) weights
- · Diagnose extreme weights and stabilise them by trimming
120 min 6 Colab cellsVideo coming soon