| 1 |
Clean re-run (rm results.json; run twice) gives identical SHA-256 (d41b75df…ae4) |
PASS |
| 2 |
n_supported+n_not_supported+n_not_assessable = 6 = n_analysed = len(assertions) (2+3+1) |
PASS |
| 3 |
Every verdict in allowed set; one locator per assertion in the required form |
PASS |
| 4 |
Input SHA-256 match data/README.md and bundle.json |
PASS |
| 5 |
n_input_rows equals wc -l (README 14, DCR 134, RESEA 58) and JSON item count 6 |
PASS |
| 6 |
A1: 100(exp(-0.431)-1) = -35.01% matches printed -35.0%; t=-2.05 consistent with ** |
PASS |
| 7 |
A3: 100(exp(-0.283)-1) = -24.65% matches printed -24.7% (source internally consistent) |
PASS |
| 8 |
A4: CI contains estimate (-0.896 < -0.406 < 0.084); width check: 1.96 SE CI includes 0 under 1.96 and 2.0 |
PASS |
| 9 |
A6: 1.99/17.3 = 11.50%; ATE CI [-2.24,-1.75] contains -1.99; Table 3 col 1 duplicates it |
PASS |
| 10 |
A5: regex search for men/women/male/female/sex/gender in DCR extract returns no lines |
PASS |
| 11 |
Baseline (naive verbatim-number match) agrees on 4/6; disagreements A2 (sign ignored) and A5 (no numbers) are explained by baseline's blindness to sign and to absent subgroups |
PASS (expected) |
| 12 |
Internal source inconsistencies (Table 6 vs A19 SE; 608 vs 640 controls) recorded in data_issues; A1 robust |
PASS (reported) |