1# Verification checks (P4)23- Clean re-run of out/analysis.py: results.json byte-identical (cmp) -> PASS4- all SEs positive -> PASS5- DCR CI contains estimate -> PASS6- DCR CI symmetric half-width = 1.96*SE -> PASS7- DCR CI excludes 0 consistent with ** star -> PASS8- RI wage CI contains estimate -> PASS9- RI weeks CI contains estimate and lo<hi -> PASS10- RI wage CI ~symmetric (midpoint within 1 of estimate) -> PASS11- normal p from derived SE rounds to printed 0.022 -> PASS12- no star mismatches across 22 DCR cells -> PASS13- HIV 5km not significant (no star) and normal p>0.10 -> PASS14- p_value null with 'not reported' locator -> PASS15- dcr_obs_5km integer -> PASS16- locators: 23 keys (all but 2 booleans) -> PASS17- n_input_rows matches wc -l (10/134/58) -> PASS18- exclusion ledger: 25-0=25=n_analysed -> PASS19- ATE magnitude vs control mean 3.5% (text) -> PASS20