{
  "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** with SE (0.210) and % change -35.0% (100*(exp(-0.431)-1) = -35.0%), so estimate, SE, 5% significance and the 35.0% reduction all match.",
    "A2": "Table 7 prints the 1 km HIV ATT as +0.063 (SE 0.801), so the asserted negative sign (-0.063) and the claimed fall in HIV hospitalisations are contradicted.",
    "A3": "Table 6 converts the 3 km ATT of -0.283 to a -24.7% change (100*(exp(-0.283)-1) = -24.6%), so the asserted 28.3% reduction wrongly reads log points as percent.",
    "A4": "Table 7 prints the 5 km hepatitis ATT as -0.406 (SE 0.250) with no star, giving a 95% CI of [-0.896, 0.084] that includes zero; the asserted [-0.66, -0.16] is only +/-1 SE.",
    "A5": "The extract reports ATTs only by buffer (Table 6) and diagnostic category (Table 7, Table A19) and contains no estimate by sex, so the men-versus-women comparison can be neither confirmed nor refuted.",
    "A6": "Table 2 prints the ATE as -1.99 weeks with 95% CI [-2.24, -1.75] and 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: Table A19 prints the CS-DR 5 km ATT -0.431** with SE (0.218), whereas Table 6 prints the same estimate with SE (0.210).",
      "handling": "Table 6 (the main results table, the one A1 cites) is used as the locator for A1; A1's 5% significance also holds with SE 0.218 at z=1.96 (|t|=1.98), and the source prints ** in both tables."
    },
    {
      "issue": "arxiv-2609.27732v1: Table A19 note says 640 never-treated comparison regions, while Table A19 body, Table 6 and the Figure 6 note say 608.",
      "handling": "Recorded only; no assertion depends on the control-region count."
    },
    {
      "issue": "arxiv-2609.27732v1: abstract states fatal hospitalisations fall by 35%, but Table 7 5 km fatal ATT -0.354 implies -29.8% and Section 6.1 says approximately 30%.",
      "handling": "Recorded only; no assertion concerns fatal hospitalisations. Tables are treated as primary evidence."
    },
    {
      "issue": "arxiv-2609.27732v1: Section 6 text calls HIV estimates 'close to zero' although Table 7 HIV ATTs at 3 km/5 km are -0.239/-0.340 (insignificant).",
      "handling": "A2 is decided by the Table 7 1 km cell (+0.063), not by the narrative."
    },
    {
      "issue": "arxiv-2609.27732v1: Table 6 prints the 3 km % change as -24.7%, while 100*(exp(-0.283)-1) = -24.65% rounds to -24.6% (a 0.05 pp rounding difference, presumably from the unrounded coefficient).",
      "handling": "Immaterial for A3: both -24.6% and -24.7% are far from the asserted -28.3%; the printed -24.7% is quoted."
    },
    {
      "issue": "arxiv-2606.14621v1: results text gives the weeks-on-UI CI as unsigned magnitudes [1.75 weeks, 2.24 weeks] while Table 2 prints the signed CI [-2.24, -1.75].",
      "handling": "Treated as the same interval expressed as a reduction; A6 is decided by Table 2."
    },
    {
      "issue": "arxiv-2606.14621v1: rendering defect 'p ¡ 0.001' in the short-term weeks-on-UI sentence.",
      "handling": "Read as p < 0.001 (consistent with Table 2); not decisive for any assertion."
    }
  ],
  "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_numeric_verbatim_match": {
    "A1": true,
    "A2": true,
    "A3": false,
    "A4": false,
    "A5": true,
    "A6": true
  },
  "computation_details": {
    "A1": {
      "att": -0.431,
      "se": 0.21,
      "stars": 2,
      "printed_pct": -35.0,
      "computed_pct": -35.01,
      "z": -2.052,
      "se_tableA19_CSDR": 0.218,
      "z_with_A19_se": -1.977
    },
    "A2": {
      "hiv_att_1km": 0.063,
      "hiv_se_1km": 0.801,
      "asserted": -0.063
    },
    "A3": {
      "att_3km": -0.283,
      "printed_pct_3km": -24.7,
      "computed_pct_3km": -24.65,
      "asserted_pct": -28.3
    },
    "A4": {
      "att_5km": -0.406,
      "se_5km": 0.25,
      "ci95_z1.96": [
        -0.896,
        0.084
      ],
      "ci95_z2.0": [
        -0.906,
        0.094
      ],
      "pm_1se": [
        -0.656,
        -0.156
      ],
      "excludes_zero_z1.96": false,
      "excludes_zero_z2.0": false,
      "stars_printed": 0
    },
    "A5": {
      "n_lines_mentioning_sex_or_gender": 0
    },
    "A6": {
      "ate": -1.99,
      "ci95": [
        -2.24,
        -1.75
      ],
      "control_mean": 17.3,
      "computed_rel_pct": -11.5
    }
  }
}