Authority
How close is the source to the underlying event, record, or responsibility?
Methodology / Trust Index
Communicate confidence without false precision
The Clarify Data Trust Index is a proposed, explainable review framework for describing evidence quality. It is not a universal truth score.
Working principle: A useful confidence label identifies the weakest evidence dimension and the next review action. It should never hide uncertainty behind an opaque percentage.
Core dimensions
Evaluate the parts of confidence separately
Freshness
Is the source current enough for the decision window and change rate?
Directness
Does the evidence support the exact claim or only a related inference?
Agreement
Do independent sources align after definitions, dates, and scopes are normalized?
Provenance
Can a reviewer trace the result through source versions and transformations?
Review status
Has an accountable person assessed applicability, conflicts, and intended use?
Method
From evidence record to review label
Define the claim and use
The same evidence can be sufficient for an internal lead and insufficient for a consequential public claim.
Assess each dimension
Record the supporting evidence and reason for each dimension rather than a bare score.
Surface the weakest point
A stale primary source and a current indirect source have different remediation paths.
Assign a review state
Use clear language such as ready for editorial use, needs source-date review, or requires specialist review.
Reassess on change
Expire or reopen the result when sources, claims, contexts, or decision requirements change.
Responsible interpretation
What the methodology does not claim
Not a probability of truth
A composite number should not imply statistically calibrated certainty unless that calibration has been demonstrated for the specific use.
Not a replacement for expertise
Clinical, legal, regulatory, financial, safety, and other high-impact interpretation remains with qualified professionals.
Not permanent
Confidence can change when evidence is corrected, superseded, contradicted, or applied to a different context.
A practical next step
Bring a real question. Leave with a clearer evidence path.
Show us the dataset, claim, report, or market question your team needs to trust. We will map the relevant workflow and the review points it requires.