1# Verification checks (P4)23- Rerun from clean state (deleted out/results.json, reran analysis.py twice): byte-identical output (cmp) — PASS4- dcr_att_5km_se > 0 — PASS5- dcr_att_3km_se > 0 — PASS6- dcr_narcotics_att_3km_se > 0 — PASS7- dcr_fatal_att_1km_se > 0 — PASS8- dcr_hiv_att_5km_se > 0 — PASS9- ri_wage_ate_se_derived > 0 — PASS10- derived DCR CI contains estimate — PASS11- derived DCR CI excludes 0 (consistent with ** p<0.05) — PASS12- wage CI contains ATE — PASS13- weeks CI contains ATE and low<high — PASS14- dcr_obs_5km is int — PASS15- HIV 5km: |b/se|<1.645 consistent with no star — PASS16- narcotics 3km |b/se|>2.576 consistent with *** — PASS17- fatal 1km |b/se|>2.576 consistent with *** — PASS18- DCR 5km 1.96<|b/se|<2.576 consistent with ** — PASS19- locators: one per non-boolean key (23) — PASS20- p_value null and locator says not reported — PASS21- printed tokens decimals: 0.210, 0.279, -0.340 preserved in printed_tokens — PASS22- n_input_rows matches wc -l (10/134/58) — PASS23- control mean wage relative ATE ~3.5% as text states — PASS24- Exclusion ledger: 25 items requested, 0 removed, 25 analysed (= n_analysed) — PASS25- Note: JSON numbers cannot carry trailing zeros (0.210 is serialised as 0.21, -0.340 as -0.34); the printed strings are preserved verbatim in results.json["printed_tokens"]. Not a failure.26