# Contract — F4-conf-02
**Question.** For six assertions in a draft literature review, each citing one of two arXiv extracts, decide from the cited extract alone whether the assertion is supported / not_supported / not_assessable.
**Estimands.** Six categorical verdicts (A1..A6), each with one locator and a one-sentence reason. Meaningful under the rules: "supported" needs every element (sign, magnitude, uncertainty, significance, population) stated or arithmetically implied; any contradicted element -> not_supported; not_assessable only when the extract reports nothing bearing on the claim.
**Conventions.** % change = 100(exp(b)-1); 95% CI = b ± 1.96·SE (2.0 acceptable).
**Outputs.** out/results.json keys: verdicts, locators, reasons, n_supported, n_not_supported, n_not_assessable, task_id="F4-conf-02", feasibility, feasibility_reason, data_issues, n_input_rows, n_analysed, random_seed(null). out/report.md (Question; Data and exclusions; Method; Results; Uncertainty; Limitations and open issues), out/analysis.py, out/claims.json.
**Information rules.** Only the extracts are evidence; no memory of papers or outside knowledge.
**Invalidity.** Using outside info; mismatched locators; a source missing / unreadable would make the task infeasible.