1{2 "baseline_verdicts_literal_number_match": {3 "A1": "supported",4 "A2": "supported",5 "A3": "not_supported",6 "A4": "not_supported",7 "A5": "supported",8 "A6": "supported"9 },10 "computation_details": {11 "A1": {12 "att": "-0.431**",13 "pct_computed": -35.01,14 "pct_printed": "-35.0%",15 "se": "(0.210)",16 "z": -2.05217 },18 "A2": {19 "att_1km": "0.063",20 "se_1km": "(0.801)"21 },22 "A3": {23 "att_3km": -0.283,24 "pct_computed": -24.65,25 "pct_printed": "-24.7%"26 },27 "A4": {28 "att_5km": -0.406,29 "ci95_1.96": [30 -0.896,31 0.08432 ],33 "ci95_2.00": [34 -0.906,35 0.09436 ],37 "pm_1se": [38 -0.656,39 -0.15640 ],41 "se_5km": 0.25,42 "stars": "none"43 },44 "A5": {45 "sex_term_hits": 046 },47 "A6": {48 "ate": -1.99,49 "ci": [50 -2.24,51 -1.7552 ],53 "control_mean": 17.3,54 "pct_of_control": 11.5,55 "text_has_11.5": true56 }57 },58 "data_issues": [59 {60 "handling": "A1 decided on Table 6 (the cited overall-ATT table); the 5% significance claim holds under either SE (z=-2.05 or -1.98), so the verdict is unaffected; flagged.",61 "issue": "DCR Table A19 prints SE (0.218) for the CS-DR 5 km ATT -0.431, while Table 6 prints (0.210) for the same estimate; the A19 note says 640 never-treated regions while A19's body and Tables 6-7 say 608."62 },63 {64 "handling": "Not decisive for any assertion; tables used as the authoritative values; flagged.",65 "issue": "DCR abstract states fatal hospitalisations fall by 35% and narcotics rise 'by a comparable amount', but Table 7 (5 km: -0.354, +0.390) implies -29.8% and +47.7%, and Section 6.1 says ~30% and 47%."66 },67 {68 "handling": "A2 decided on the printed sign of the 1 km Table 7 cell (+0.063), not on the prose.",69 "issue": "DCR text describes HIV estimates as 'close to zero' although the 3 km and 5 km cells are -0.239 and -0.340 (insignificant)."70 },71 {72 "handling": "Not decisive; noted only.",73 "issue": "DCR Figure 7 note refers to four diagnostic sub-categories whereas Table 7 and the text list five."74 },75 {76 "handling": "Treated as equivalent representations; A6 decided on Table 2 signed values.",77 "issue": "RESEA results text prints CIs of reductions as positive magnitudes ([1.75, 2.24]) while Table 2 prints signed bounds [-2.24, -1.75]; the text also contains a rendering defect 'p ¡ 0.001' for p < 0.001."78 }79 ],80 "feasibility": "complete",81 "feasibility_reason": null,82 "locators": {83 "A1": "arxiv-2609.27732v1 | Table 6 | row \"ATT\" | column \"5 km\"",84 "A2": "arxiv-2609.27732v1 | Table 7 | row \"HIV\" | column \"1 km\"",85 "A3": "arxiv-2609.27732v1 | Table 6 | row \"% change\" | column \"3 km\"",86 "A4": "arxiv-2609.27732v1 | Table 7 | row \"Hepatitis\" | column \"5 km\"",87 "A5": "not reported in arxiv-2609.27732v1",88 "A6": "arxiv-2606.14621v1 | Table 2 | row \"Weeks on UI\" | column \"Average treatment effect estimate\""89 },90 "n_analysed": 6,91 "n_input_rows": {92 "README.md": 14,93 "arxiv-2606.14621v1.txt": 58,94 "arxiv-2609.27732v1.txt": 134,95 "assertions.json": 696 },97 "n_not_assessable": 1,98 "n_not_supported": 3,99 "n_supported": 2,100 "random_seed": null,101 "reasons": {102 "A1": "Table 6 prints the 5 km ATT as -0.431** with SE (0.210) (** = p<0.05) and a % change of -35.0%, which equals 100(exp(-0.431)-1) = -35.0%, so sign, magnitude, SE, significance and conversion all match.",103 "A2": "Table 7 prints the 1 km HIV ATT as +0.063 (SE 0.801), positive and insignificant, so the assertion's negative sign (-0.063) and the claimed fall in HIV hospitalisations are contradicted.",104 "A3": "Table 6 converts the 3 km ATT of -0.283 to a -24.7% change (100(exp(-0.283)-1) = -24.6%), not 28.3%, so the magnitude is wrong (log points were read as percent); the estimate is also insignificant (SE 0.279).",105 "A4": "Table 7 prints the 5 km hepatitis ATT as -0.406 with SE (0.250) and no star, giving a 95% CI of [-0.896, 0.084] (±1.96 SE), which includes zero; the asserted [-0.66, -0.16] is only ±1 SE.",106 "A5": "The extract reports ATTs only by buffer, diagnostic category and estimator (Tables 6, 7, A19) and contains no estimate by sex, so the men-versus-women comparison can be neither confirmed nor refuted.",107 "A6": "Table 2 prints the ATE on weeks on UI as -1.99 weeks (95% CI [-2.24, -1.75], p<0.001) with control mean 17.3 weeks, and 1.99/17.3 = 11.5%, matching the Section 3 text."108 },109 "task_id": "F4-conf-02",110 "verdicts": {111 "A1": "supported",112 "A2": "not_supported",113 "A3": "not_supported",114 "A4": "not_supported",115 "A5": "not_assessable",116 "A6": "supported"117 }118}