Method & sources · method
How we grade evidence (method v1)
Every bio-solution is graded on two independent axes. They are never merged into one grade.
| Axis | Question | Scale |
|---|---|---|
| D — adoption | Has the alternative actually displaced the old process, and where? | D0–D3 |
| E — source quality | How reliable are the sources behind the claims? | E0–E5 |
Source levels#
| Level | Source type |
|---|---|
| E0 | Claim without a source |
| E1 | The company’s own website or post; other sources until reviewed by hand |
| E2 | Trade press, industry associations |
| E3 | Official statistics, government and intergovernmental registers |
| E4 | Peer-reviewed study or independent test |
| E5 | Market outcome: registration plus repeat purchases (not used yet) |
Rule for the E grade#
- Only claims whose quote was checked against the source text count.
- For D2–D3, displacement claims count; for D0–D1, performance and displacement claims count.
- One publisher (source domain) is one independent source, at its best level.
- The E grade is the second-strongest independent source, so one strong source cannot carry a grade alone.
- With fewer than two independent sources at E2 or above, the grade is E–, and the adoption level (D) counts as unconfirmed.
Limits of v1#
- No time window: the data holds the year of the event, not the publication date of the source.
- Sources typed “other” count as E1 until reviewed by hand (strict mode).
- Russia coverage is still weak: most sources are foreign.
Version: v1, 2026-10-17. Changes to the method are published here with a date.