# Contract — F4-conf-01

Question. Extract, exactly as printed, specified treatment-effect estimates and their uncertainty from two source extracts (Armendariz 2026 DCR staggered DiD, arxiv-2609.27732v1, Tables 6/7; Li et al. 2026 RESEA RCT, arxiv-2606.14621v1, Table 2), with locators; derive (a) a 95% CI from the Table 6 5 km SE and (b) an SE from the Table 2 wage-ATE 95% CI, stating the convention; recognise unreported quantities (exact p-value of HIV 5 km ATT; SE of wage ATE). Estimands. The printed numbers themselves (no re-estimation). Derived quantities are meaningful only under approximate normality of the estimator and a symmetric Wald interval (critical value 1.96). Outputs. out/results.json (keys listed in TASK.md incl. locators, derivation_notes, task_id, feasibility, feasibility_reason, data_issues, n_input_rows, n_analysed, random_seed=null); out/report.md (6 sections); out/analysis.py (regenerates results.json from data/ and prints matched source lines); out/claims.json. Information rules. Only the two extracts are admissible; values as printed (sign, decimals; "17,250" -> 17250); use the table named in the key when a value appears in several tables; derived keys from printed values of named table; null + "not reported" locator when not printed. Invalidity. Using remembered values; using A19 or Table 3 instead of Table 6/Table 2; rounding printed values; deriving a p-value and reporting it as printed.