{
  "task_id": "F4-conf-02",
  "verdicts": {
    "A1": "supported",
    "A2": "not_supported",
    "A3": "not_supported",
    "A4": "not_supported",
    "A5": "not_assessable",
    "A6": "supported"
  },
  "locators": {
    "A1": "arxiv-2609.27732v1 | Table 6 | row \"ATT\" | column \"5 km\"",
    "A2": "arxiv-2609.27732v1 | Table 7 | row \"HIV\" | column \"1 km\"",
    "A3": "arxiv-2609.27732v1 | Table 6 | row \"% change\" | column \"3 km\"",
    "A4": "arxiv-2609.27732v1 | Table 7 | row \"Hepatitis\" | column \"5 km\"",
    "A5": "not reported in arxiv-2609.27732v1",
    "A6": "arxiv-2606.14621v1 | Table 2 | row \"Weeks on UI\" | column \"Average treatment effect estimate\""
  },
  "reasons": {
    "A1": "Table 6 prints the 5 km ATT as -0.431** (SE 0.210, ** = p<0.05) with a % change of -35.0% (100*(exp(-0.431)-1) = -35.0%), matching every element of the assertion.",
    "A2": "Table 7 prints the 1 km HIV ATT as 0.063, i.e. positive (SE 0.801), so the asserted negative sign and the claimed fall in HIV hospitalisations are contradicted.",
    "A3": "The ATT of -0.283 is in log points and Table 6 converts it to a -24.7% change (100*(exp(-0.283)-1) = -24.65%), not a 28.3% reduction; the estimate is also not significant (SE 0.279).",
    "A4": "With ATT -0.406 and SE 0.250 (no stars), the 95% CI is about [-0.90, 0.08], which includes zero; the asserted [-0.66, -0.16] is only the estimate ± 1 SE.",
    "A5": "The extract reports no estimates by sex (Tables 6, 7 and A19 report only buffer, outcome-category and estimator breakdowns, none by sex), so a men-versus-women comparison can be neither confirmed nor refuted.",
    "A6": "Table 2 prints the ATE for weeks on UI as -1.99 weeks (95% CI [-2.24, -1.75]) with a control mean of 17.3 weeks, and -1.99/17.3 = -11.5%, matching the text's 11.5% decrease."
  },
  "n_supported": 2,
  "n_not_supported": 3,
  "n_not_assessable": 1,
  "feasibility": "complete",
  "feasibility_reason": null,
  "data_issues": [
    {
      "issue": "arxiv-2609.27732v1: the CS-DR 5 km overall ATT -0.431** has SE 0.210 in Table 6 but SE 0.218 in Appendix Table A19 (same estimator, buffer, window and N=17,250).",
      "handling": "A1 is judged on Table 6 (the cited main result); the 5% significance claim holds under both SEs (|z| = 2.05 and 1.98 > 1.96), so the verdict is unaffected."
    },
    {
      "issue": "arxiv-2609.27732v1: Table A19 lists 608 control regions but its note says 640 never-treated comparison regions.",
      "handling": "Not decisive for any assertion; recorded only."
    },
    {
      "issue": "arxiv-2609.27732v1: the abstract says fatal outcomes fall by 35%, whereas Table 7 gives a 5 km fatal ATT of -0.354 (100*(exp(b)-1) = -29.8%) and Section 6.1 says approximately 30%.",
      "handling": "Not decisive for any assertion (none concerns fatal outcomes); recorded only."
    },
    {
      "issue": "arxiv-2609.27732v1: Table A19 printed % changes for Synthetic DiD (-30.3%) and Stacked DiD (-13.9%) differ by 0.1 pp from 100*(exp(b)-1) (-30.4%, -13.8%), likely rounding from unrounded b.",
      "handling": "Not decisive for any assertion; recorded only."
    },
    {
      "issue": "arxiv-2606.14621v1: the results text reports the weeks-on-UI CI as magnitudes ([1.75 weeks, 2.24 weeks]) while Table 2 prints signed bounds [-2.24, -1.75]; the text also contains a rendering defect 'p ¡ 0.001'.",
      "handling": "Treated as equivalent sign conventions; A6 judged on the signed Table 2 values."
    }
  ],
  "n_input_rows": {
    "assertions.json": 6,
    "README.md": 14,
    "arxiv-2609.27732v1.txt": 134,
    "arxiv-2606.14621v1.txt": 58
  },
  "n_analysed": 6,
  "random_seed": null,
  "baseline_share_numbers_found_verbatim": {
    "A1": 1.0,
    "A2": 1.0,
    "A3": 0.75,
    "A4": 0.75,
    "A5": null,
    "A6": 1.0
  }
}