{
  "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 ATT -0.431** (SE 0.210) at 5 km, ** denotes p<0.05, and the % change row prints -35.0% (100(exp(-0.431)-1) = -35.0%), so every element matches.",
    "A2": "Table 7 prints the 1 km HIV ATT as +0.063 (SE 0.801), positive and insignificant, so the asserted negative sign and the claimed fall in HIV hospitalisations are contradicted.",
    "A3": "Table 6 converts the 3 km ATT of -0.283 into a -24.7% change (100(exp(-0.283)-1) = -24.65%), not 28.3%, which is the log-point coefficient misread as a percentage.",
    "A4": "With ATT -0.406 and SE 0.250 (no stars), the 95% CI is [-0.90, 0.08] (±1.96·SE), which includes zero; the asserted [-0.66, -0.16] is only ±1 SE ([-0.656, -0.156]).",
    "A5": "The extract reports ATTs only by buffer, estimator and diagnostic category (Tables 6, 7, A19) and never by sex, so a men-versus-women comparison can be neither confirmed nor refuted.",
    "A6": "Table 2 prints the ATE as -1.99 weeks (95% CI [-2.24, -1.75]) against a control mean of 17.3 weeks, and -1.99/17.3 = -11.5%, matching the 11.5% decrease stated in the results text."
  },
  "n_supported": 2,
  "n_not_supported": 3,
  "n_not_assessable": 1,
  "feasibility": "complete",
  "feasibility_reason": null,
  "data_issues": [
    {
      "issue": "arxiv-2609.27732v1 Table A19 prints the CS-DR 5 km ATT -0.431** with SE (0.218), while Table 6 prints the same specification with SE (0.210).",
      "handling": "A1 cites the main result; judged against Table 6 (SE 0.210). With SE 0.210, |t|=2.05 exceeds both 1.96 and 2.0; with SE 0.218, |t|=1.98 exceeds 1.96 but not 2.0, yet ** is printed in both tables, so the verdict rests on the printed significance stars."
    },
    {
      "issue": "arxiv-2609.27732v1 Table A19 note states 640 never-treated comparison regions; table body and Tables 6-7 state 608.",
      "handling": "Recorded only; no assertion depends on the control-region count."
    },
    {
      "issue": "arxiv-2609.27732v1 abstract says fatal hospitalisations fall by 35%, but Table 7 5 km fatal ATT -0.354 implies -29.8% (Section 6.1 says ~30%).",
      "handling": "Recorded only; no assertion concerns fatal outcomes."
    },
    {
      "issue": "arxiv-2609.27732v1 Figure 7 note mentions four sub-categories while the text and Table 7 report five.",
      "handling": "Recorded only; not decision-relevant."
    },
    {
      "issue": "arxiv-2606.14621v1 results text reports CIs of reductions as unsigned magnitudes ([1.75, 2.24]) while Table 2 uses signed values; rendering defect 'p ¡ 0.001'.",
      "handling": "Treated as equivalent; A6 judged against the signed Table 2 values."
    }
  ],
  "n_input_rows": {
    "assertions.json": 6,
    "README.md": 14,
    "sources/arxiv-2609.27732v1.txt": 134,
    "sources/arxiv-2606.14621v1.txt": 58
  },
  "n_analysed": 6,
  "random_seed": null,
  "baseline_verdicts_number_match": {
    "A1": "supported",
    "A2": "supported",
    "A3": "not_supported",
    "A4": "not_supported",
    "A5": "supported",
    "A6": "supported"
  },
  "baseline_agreement_with_final": 4,
  "derived_quantities": {
    "A1_pct_change_5km": -35.01,
    "A3_pct_change_3km": -24.65,
    "A4_ci95_1.96": [
      -0.896,
      0.084
    ],
    "A4_ci95_2.0": [
      -0.906,
      0.094
    ],
    "A4_ci_pm_1se": [
      -0.656,
      -0.156
    ],
    "A6_pct_of_control_mean": -11.5
  }
}