Build an evidence ledger that another researcher can audit

Connect every important claim to an original source, a definition, a date and a stated limitation.

For Public,Students,Researchers,Founders · Reviewed 2026-10-09

An evidence ledger is our proposed working table connecting a claim to the material used to support it. It separates direct observations, calculations, interpretations and unresolved questions. The Basel Committee’s principles on risk-data aggregation and reporting provide institutional background on data quality and reporting; this smaller research worksheet is an editorial adaptation, not a claim of regulatory compliance.

For each material claim, store the exact source URL, publisher, publication or observation date, relevant definition and a short explanation of what the evidence establishes. A company’s own filing can be a primary source for its reported figures. EDGAR provides a public search route for US securities filings, but a filing still needs to be read in context.

Add a calculation column whenever you transform a source number. State the numerator, denominator, units and assumptions. If two credible sources disagree, preserve both and explain whether their scope or timing differs. Ten articles repeating the same announcement do not necessarily supply ten independent pieces of evidence; trace them back to the underlying record.

Use a final column for limitations and the next verification step. An unsupported claim should remain unsupported until new evidence arrives, even when it sounds plausible or an AI system writes it confidently. This approach lets teammates update a briefing without losing the reasoning history. It also gives a reader a practical path from the conclusion back to its inputs, which makes the research more useful than a list of links alone.

Questions to investigate

1. Which source directly supports the claim rather than repeating it?
2. Can another reader reproduce the transformation from the original number?
3. What unresolved limitation could materially alter the conclusion?

Primary sources

• Basel Committee: Risk data aggregation and reporting principles
• US Securities and Exchange Commission: EDGAR filing search

Related research

• Audit market-data freshness before interpreting a chart

  Observation time, retrieval time, release time and revision history are different pieces of evidence.