Test report forgery: difference between revisions
Diff·revision 21 → 22·11:27, 27 Aug 2025
Difference between revision 21 and revision 22 of Test report forgery. 2 lines changed; the page grew by 336 bytes.
| Revision 21 — 12:22, 9 Aug 2025 PanelPlannerPam (talk) expand §Forms and their detectability 5,055 bytes ±0 | Revision 22 — 11:27, 27 Aug 2025 eGFR_Esme (talk) add the citation for the blind-sampling discussion 5,391 bytes +336 | ||
|---|---|---|---|
| 44 | The practical position that follows is not that documents are worthless. A supplier that issues lot-specific certificates naming its methods, retrievable years later against a lot identifier, has built something that is expensive to fake consistently over time — and consistency over time is a stronger signal than any single document. That is a statement about a records system rather than about any one certificate.{{r|iso9001}} | 44 | The practical position that follows is not that documents are worthless. A supplier that issues lot-specific certificates naming its methods, retrievable years later against a lot identifier, has built something that is expensive to fake consistently over time — and consistency over time is a stronger signal than any single document. That is a statement about a records system rather than about any one certificate.{{r|iso9001}} |
| 45 | 45 | ||
| + | 46 | The reverse also holds: a single unverifiable document proves little in either direction, and treating its absence as evidence of wrongdoing is as unsound as treating its presence as evidence of quality. See [[Purity claim inflation]] for the adjacent failure in which the document is genuine but the figure is presented misleadingly. | |
| + | 47 | ||
| 46 | == References == | 48 | == References == |
| 47 | {{reflist}} | 49 | {{reflist}} |