# Verification checks (P4), 2026-10-29

| check | outcome |
|---|---|
| Clean re-run (deleted results.json, table1.csv, figure1.png, exclusions.csv; re-ran analysis.py) → results.json byte-identical (`cmp`) | PASS |
| CI contains estimate: 4.854 < 5.267 < 5.680 | PASS |
| effect_se, baseline_metric, model_metric > 0 | PASS |
| Exclusion ledger: 2349 − (396+1736+20) = 197 = n_analysed | PASS |
| n_analysed + n_excluded_missing = 217 countries in 2023 | PASS |
| pearson_r² = r_squared (simple OLS identity) | PASS |
| r_squared = 1 − (model_metric/baseline_metric)² | PASS |
| model_metric < baseline_metric | PASS |
| table1 n_countries sums to n_analysed (197) | PASS |
| CI half-width = 1.95996 × HC3 SE | PASS |
| conclusion code consistent with CI sign | PASS |
| analysis_year = max year with ≥150 complete countries (all years qualify; 2023 has 197) | PASS |
| Magnitude: fitted LE at ln(GDPpc)=9.5 ≈ 72.6 years, plausible; slope in typical literature range (~4–6 yrs/log point) | PASS |
| Figure visually inspected: scatter + OLS line, labelled axes | PASS |
| data/ unmodified (SHA-256 identical to README) | PASS |