claim-confidence-registry

What problem does it solve?

Prevents overclaiming and lack of transparency by systematically mapping manuscript statements to the concrete outputs that support them and by assigning clear confidence classes with required caveats.

Core Features & Use Cases

• Claim-to-Evidence Mapping: Locates exact table or cell references in analysis outputs that support each manuscript claim.
• Confidence Classification: Applies a four-class taxonomy (robust/tentative/exploratory/unsupported) with rules for downgrading overclaims.
• Flagging & Caveats: Detects common overclaiming patterns (CV presented as confirmatory, small-N subgroup claims, null-result language) and generates required cautionary language for publication.
• Use Case: Validate Results and Discussion claims before submission by producing a CSV/JSON claim matrix linking claims to source outputs and caveats.

Quick Start

Generate a claim matrix mapping each manuscript claim to its supporting output, assign a confidence class, and export the results as CSV or JSON.