# Contract — F4-conf-01

Question. Extract, exactly as printed, specified treatment-effect estimates and their uncertainty from two arXiv extracts (Armendariz 2026, DCR staggered DiD, Tables 6/7; Li et al. 2026, RESEA RCT, Table 2), attach locators, derive (a) a 95% CI from a printed SE (DCR 5 km ATT, Table 6) and (b) an SE from a printed 95% CI (RESEA wage ATE, Table 2), and identify which requested quantity is not reported (exact p-value of DCR HIV 5 km ATT).

Estimands. The printed numbers themselves (no re-estimation). Derived quantities are meaningful only under an approximate-normal sampling distribution of the estimator (symmetric Wald interval, critical value 1.96).

Outputs. out/results.json (25 content keys listed in TASK.md + locators, derivation_notes, task_id, feasibility, feasibility_reason, data_issues, n_input_rows, n_analysed, random_seed), out/report.md (Question; Data and exclusions; Method; Results; Uncertainty; Limitations and open issues), out/analysis.py (regenerates results.json from data/ only, prints matched lines), out/claims.json.

Information rules. Only the two extracts are admissible; no memory of the papers. Values as printed (sign, decimals; thousands separators removed). Where a value appears in several tables, use the table named in the key (e.g. Table 6 SE 0.210, not Table A19 0.218; Table 2 not Table 3). Derived keys state convention. Unreported quantities → null with "not reported in " locator.

Invalidity / impossibility. Missing tables/cells in extracts; ambiguous cell attribution; using non-named table; inventing an exact p-value from stars.