R0002/2026-03-13/C002/H1

Research	R0002 — Research Standards for AI-Assisted Writing
Run	2026-03-13
Claim	C002
Hypothesis	H1

Statement

All four sub-claims are accurate: GRADE separates evidence quality from recommendation strength, has four certainty levels, five downgrade criteria, and three upgrade criteria.

Status

Current: Supported

Supporting Evidence

Evidence	Summary
SRC01-E01	Explicitly states GRADE separates evidence quality from recommendation strength
SRC01-E02	Confirms four certainty levels: High, Moderate, Low, Very low
SRC01-E03	Confirms five downgrade criteria
SRC02-E01	Confirms three upgrade criteria
SRC02-E02	Independently confirms all structural elements

Contradicting Evidence

Evidence	Summary

[No evidence contradicts this hypothesis.]

Reasoning

H1 is strongly supported by both the seminal GRADE publication and the CDC implementation handbook. The 2008 BMJ paper explicitly states that GRADE deliberately separates evidence quality from recommendation strength, making the "core insight" characterization editorially reasonable. All numerical counts (four levels, five downgrade, three upgrade) are confirmed by multiple sources.

Relationship to Other Hypotheses

H1 is mutually exclusive with H2 (which claims the "core insight" is editorial interpretation only) and H3 (which claims numerical counts are wrong). H1's strong support effectively eliminates both alternatives.