# Data audit (P1)
| file | SHA-256 | lines (wc -l) / items | notes |
|---|---|---|---|
| data/assertions.json | 2965ab3acc6f326653169b83a5345698b51fac88a556de65faf89210aea0cad4 | 6 items (id, cites, text) | ids A1–A6 unique; A1–A5 cite arxiv-2609.27732v1, A6 cites arxiv-2606.14621v1; no missing fields |
| data/README.md | 5348bd49912099735108dbbe19ae09fa72bd600cb3d12b878658e6b9806198a7 | 14 lines | provenance; extract hashes listed match computed hashes |
| data/sources/arxiv-2609.27732v1.txt | ee2d444d2f78f2af2ca2cf69b59a61fcd7ac5e10ea29e8ddcd67917bf07d6a8b | 134 lines | matches README hash. Tables 6, 7, A19; log points; SE in parentheses; stars ***/**/* = 1/5/10% |
| data/sources/arxiv-2606.14621v1.txt | 27c463c0a144d9d5a3916dc641f15b6029f6c68b5e574ad9074e7df1aeb0a238 | 58 lines | matches README hash. Table 2/3 re-set one row per outcome; weeks, $, pp |

No dates/time series to range-check (the papers' study periods: DCR 1998–2022; RESEA Feb 2022–Sep 2023, as stated in the extracts).

## Anomalies / internal inconsistencies (all recorded in results.json `data_issues`)
1. DCR: CS-DR 5 km ATT -0.431** has SE (0.210) in Table 6 but (0.218) in Table A19. Not decisive for A1 (significant at 5% under both).
2. DCR: Table A19 says 608 control regions, its note says 640.
3. DCR: abstract says fatal outcomes fall 35%; Table 7 5 km fatal ATT -0.354 → -29.8%; Section 6.1 says ~30%.
4. DCR: Table A19 % change for Synthetic DiD / Stacked DiD differ by 0.1 pp from 100(exp(b)-1) of printed b (rounding).
5. RESEA: text gives CI as magnitudes [1.75, 2.24] vs signed table [-2.24, -1.75]; rendering defect "p ¡ 0.001".

## Information-availability
No input violates the rules; only the extracts are used. No outside knowledge used. No exclusions applied (all 6 assertions assessed).