# Checks (P4)
1. Clean re-run (results.json and exclusions.csv deleted, script re-run): byte-identical results.json — PASS.
2. SEs positive — PASS.
3. 95% CI contains tau_hat — PASS.
4. CI excludes 0 iff effect_significant_5pct — PASS (CI contains 0, not significant).
5. Counts: n_input_rows 1000 − 0 removed = n_analysed 1000 = 40 schools × 25 — PASS.
6. icc_hat in [0,1] — PASS (0.235).
7. Magnitude vs baseline: se_cluster/se_iid = 2.584 vs Moulton sqrt(1+(25−1)·ICC) = 2.577 — PASS (consistent).
8. Cross-check: Welch t-test on 40 school means gives p = 0.201 vs cluster p = 0.195 — PASS (same conclusion).