Primary sources first
Official datasets, regulatory filings, public records, academic literature and direct evidence wherever possible.
Nataraja LabsResearch FoundationMandates ↗How we research
A conclusion is only as serious as the chain connecting source, transformation, interpretation, uncertainty and review.

THE STANDARD
The Foundation crosses disciplinary boundaries because reality requires it—not to escape the standards of any discipline. Every enquiry must make its sources, transformations, assumptions, uncertainty and institutional interests visible enough to be examined.
OPERATING PRINCIPLES
Official datasets, regulatory filings, public records, academic literature and direct evidence wherever possible.
Claims should remain traceable to the evidence, transformations and assumptions that produced them.
Questions are examined across technical, economic, institutional, historical and social boundaries.
Measurement reveals patterns; institutions, incentives, history and human behaviour explain their meaning.
Uncertainty is represented explicitly rather than hidden behind false precision.
Material conclusions should be challenged through competing explanations and adversarial review.
Substantive errors should be corrected visibly and without institutional defensiveness.
Funding may support inquiry. It must not purchase conclusions.

ORDER OF EVIDENCE
Official records, original datasets, filings, legislation, judgements, technical documentation, field observation and direct testimony.
Peer-reviewed research, authoritative monographs, archival scholarship and established technical literature.
Company disclosures, policy documents, professional records and material produced by accountable organisations.
Expert analysis, journalism and commentary used with attribution and never permitted to conceal the underlying evidence.
LANGUAGE OF CONFIDENCE
These categories are not mathematical probabilities. They are editorial disciplines intended to prevent presentation from implying more certainty than the record can support.
Directly supported by evidence considered sufficiently reliable for the claim.
The best present account, with uncertainty and competing explanations disclosed.
Material evidence or interpretation remains unresolved between credible positions.
A consequential claim exists in the field but has not been independently established.
CORRECTIONS & REVISION
Material errors should be corrected visibly. Revised datasets, methods and publications should preserve version history where practicable. Changes affecting the interpretation of a finding should be distinguished from typographic or presentational corrections.
Authors remain responsible for accuracy. Review strengthens that responsibility; it does not replace it.
Submit a research correction ↗