Fact-Checking Policy

What We Check

• Names, dates and institutional affiliations, including whether a role or organization is correctly identified in the relevant context
• Statistics, measurements and calculations, including units, comparisons and the basis for numerical claims
• Quotations and attributed statements, including whether the wording and attribution preserve the source’s intended meaning
• Study design, publication status and limitations that materially affect the strength or generalizability of a finding
• Official mission, regulatory and technical claims, including whether the cited document supports the statement made
• Links to papers, datasets and public records, so readers can reach the underlying source where it is publicly available

Preferred Sources

Peer-reviewed papers, clearly identified preprints, datasets, government records, official institutional documents, technical documentation and attributable experts are used where relevant. Source selection depends on the claim being checked: a press release may establish what an institution announced, but it is treated as a first-party statement rather than independent proof of scientific validity or real-world effect.

Research Status and Uncertainty

Articles should identify whether research is peer reviewed, preliminary, observational, laboratory-based, animal-based or clinical when that distinction affects interpretation. Correlation should not be described as causation without adequate support, and an early result should not be presented as settled evidence merely because it is new or widely discussed.

Uncertainty, missing data and important methodological limits should be stated where they materially affect the conclusion, including limits on sample size, measurement, comparison groups, generalization or the difference between a model and an observed outcome.

Health and High-Impact Claims

Medical, health, public-safety and other high-impact claims require additional caution, appropriate sources and wording that does not turn population-level evidence into a personalized conclusion. Science Report does not provide personalized diagnosis, treatment, legal, financial, tax, investment or mental-health advice.

Anonymous Sources

Named and attributable sources are preferred because they allow readers to assess expertise, access and possible conflicts. Unnamed sources are used rarely, only when there is clear public interest, the source’s identity is known to editorial leadership and material claims can be corroborated where possible.

Post-Publication Review

If reliable new evidence or a credible error report becomes available, the article may be reviewed to determine whether published information is inaccurate, incomplete, unclear or no longer current, and may then be corrected, clarified or updated under the Corrections Policy.