# Contract — F4-conf-02
**Question.** For each of six assertions (data/assertions.json), each citing one of two arXiv extracts, decide from the cited extract alone whether it is `supported` (every element — sign, magnitude, uncertainty, significance, population — stated or arithmetically implied), `not_supported` (some element contradicted / does not follow), or `not_assessable` (extract reports no quantity bearing on the claim).
**Estimands.** Six categorical verdicts, each with one locator and a one-sentence reason. 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 [{issue,handling}], n_input_rows {file: lines/items}, n_analysed, random_seed (null). out/report.md (Question; Data and exclusions; Method; Results; Uncertainty; Limitations and open issues). out/analysis.py (regenerates results.json from data/, prints source lines relied on). out/claims.json.
**Information rules.** Only the extracts are admissible; no memory of papers or outside knowledge. Report printed numbers verbatim.
**Invalidity.** Using outside info; a verdict not grounded in an extract line; locator absent; mixing partially correct claims into "supported".