Research standards

Prefer the original record where it is available. Distinguish official statistics, research, company statements and other source types, including the limits relevant to the claim.

Keep units, time periods, populations and definitions explicit. Do not turn a projection into an observation or compare incompatible measures without explaining the difference.

Explain the assumptions behind an inference and material evidence that could weaken it. Unknown information remains unknown.

Material corrections identify the affected claim and explain the change. Formatting updates do not make an old publication newly published.

These are publication standards, not a claim of external certification, universal independent peer review or a guarantee that no errors remain.