A hands-on refresher on experimental design and causal inference for tech company interviews. Heavy on simulations, numerical examples, and intuition. Applications drawn from ride-sharing platforms, online advertising, and academic research.
| # | Module | Status |
|---|---|---|
| 1 | The Experimental Ideal — potential outcomes, ATE/ATT/ATU, selection bias, randomization | ✓ done |
| 2 | SUTVA and When It Breaks — interference, marketplace effects, network spillovers | ✓ done |
| 3 | Designing Around Interference — cluster randomization, switchback, geo experiments | ✓ done |
| 4 | Power and Sample Size — MDE, simulation-based power, clustering, ICC | upcoming |
| 5 | Analyzing Experiments — regression adjustment, CUPED, ITT vs LATE, PAPs | ✓ done |
| 6 | Multiple Testing & Subgroups — Bonferroni, BH, pre-specified subgroups, forking paths | upcoming |
| 7 | External Validity — site selection, transportability, temporal validity | ✓ done |
| 8 | Beyond the A/B Test — modern DiD (Goodman-Bacon, CS, Honest DiD), synthetic control / SDID, causal forests for HTE | ✓ done |