Data audit (P1)
| file | SHA-256 | lines (wc -l) | notes |
|---|---|---|---|
| README.md | c6d375725356176fa7600ab663d9306c5f37c1097b619792bcbed166ccc792b2 | 17 | provenance table |
| reference_list.json | 7ab6cbf2606f90f3174215214feddd73533d29cda4befdc77ffdb595193ab5da | (5 items) | keys id, entry, cited_for, source_file; no duplicate ids; every source_file exists |
| sources/arxiv-2505.04555v3.txt | b1731a9b…95c1 | 66 | SHA matches README |
| sources/arxiv-2510.03658v1.txt | f8782937…1127 | 91 | SHA matches README |
| sources/arxiv-2603.03260v2.txt | 3cb9a4f3…1551 | 107 | SHA matches README |
| sources/arxiv-2606.14621v1.txt | 27c463c0…a238 | 58 | SHA matches README |
| sources/arxiv-2609.27732v1.txt | ee2d444d…a8b | 134 | SHA matches README |
All files are UTF-8 with no CR characters. Each extract has Title, Authors and a Bibliographic record header, and a body title line beginning with "# ". Submission dates (from the Bibliographic record lines): 2609.27732 v1 23 Sep 2026; 2606.14621 v1 12 Jun 2026; 2603.03260 v1 3 Mar 2026, v2 2 Apr 2026; 2505.04555 v1 7 May 2025, v2 27 May 2025, v3 16 Mar 2026; 2510.03658 v1 4 Oct 2025. Anomalies (also listed under data_issues in results.json):
- 2609.27732: SE of the 5 km CS-DR ATT is 0.210 in Table 6 and 0.218 in Table A19. The A19 note says 640 control regions, while the other tables and notes say 608. The abstract says fatal outcomes fall 35%, while Sec 6.1 says ~30%.
- 2505.04555: the own-wage elasticity SE (0.091) equals the affected-employment SE, which is implausible for a ratio estimate.
- 2603.03260: the Table 3 header is garbled. The editor's note gives the column order.
- 2510.03658: footnote markers are rendered inline (for example "percent^9^9 9"). This is cosmetic. Information availability: no input violates the rules. The "Current version" dates printed in the HTML renderings are excluded by the extract editor. Only header and body content is used.