# K3 control – summary (30.10.2026)
All 404 rows checked against full texts; no wrong numbers. 2 major, 9 minor (flags/units/metadata).

| study | rows | numbers correct | critical | major | minor | fit for use |
|---|---|---|---|---|---|---|
| oba2006 | 37 | 37 | 0 | 0 | 2 | yes |
| wei2004 | 62 | 62 | 0 | 1 (Table III unit: per week printed vs per day in Methods; 061/062 → disputed) | 1 | yes; 061/062 disputed |
| singh1998 | 72 | 72 | 0 | 0 | 2 (white meat includes_processed should be unclear) | yes |
| bostick1994 | 74 | 74 | 0 | 1 (two main poultry sets) | 2 | yes after fix |
| cross2010 | 159 | 159 | 0 | 0 | 2 (adj_age unclear; no CUP row) | yes |

Points: Bostick main dish includes ham (partial correct); Wei main dish undefined (unclear correct; NHS 1984 questionnaire likely includes ham); Singh 157 = in model, 127 = known intake (both right); Cross unprocessed red meat only as HvL 1.13 (0.98–1.30); per-100 g rows are footnote model (not main).