Every piece of content is scored across seven structural dimensions. These dimensions evaluate how content is constructed and sourced, not its ideological or political orientation. The methodology is public and auditable.
A gauge implies an instrument reading. A grade implies a judge. The reader interprets. The tool measures.
Scores update continuously as the evidence environment changes. Every state is stored. No score is ever overwritten.
All scoring methodology, evidence, and challenge outcomes are permanently accessible and publicly auditable.
How many genuinely independent sources confirm the core claims? Sources that all derive from the same original report do not count as multiple corroborating sources.
Signal Extraction
How traceable is the sourcing chain? Are claims attributed to named sources? Is the attribution chain visible or obscured?
Signal Extraction
What proportion of content is inferential, and how explicitly is speculation marked? Has speculation been upgraded to fact as the story traveled?
Signal Extraction
What is the quality and independence of cited sources? Do they have structural conflicts of interest? Are they appropriately qualified for the claims they are cited on?
Signal Extraction
Does the content build or degrade the reader's evaluative capacity? Does it present multiple perspectives, or does it pre-digest conclusions?
Signal Extraction
Does framing exploit emotional reactivity rather than inform? Is the headline accurately representative of the body?
Signal Extraction
Are the conditions under which the content was produced transparent? Is funding disclosed? Are conflicts of interest visible?
Signal Extraction
Rather than rating sources as credible or not, the Source Index tracks capture states , behavioral trajectories, not static labels.
Definition: Structural conditions present that create possibility of capture.
Triggers: Ownership relationship, funding relationship, or personnel relationship creating conditions for capture.
Does not mean source is unreliable; means conditions for unreliability are present.
Definition: Behavioral patterns consistent with capture are observable.
Triggers: Pattern detection across a corpus: systematic absence of inconvenient information, directional skew in source selection, correction rate below baseline.
Identified through pattern detection, not editorial judgment about individual pieces.
Definition: Independently verified conflicts of interest documented and confirmed.
Triggers: Verifiable funding disclosure, documented coordination with covered party, formal correction record demonstrating systematic inaccuracy.
Requires external verification; not inferred from pattern alone.
Any source classified at Potential, Probable, or Verified can submit a formal challenge. Challenges are publicly logged immediately. Review outcomes, including full reasoning, are permanently preserved. A source that has successfully challenged a classification has a different record than one that has never been classified. A source that has repeatedly failed challenges has a different record than either.
