Editorial standards
Evidence first. Tone second.
PerBiscuit aims to make portion calculations clear without making the evidence sound stronger than it is.
Scope
This policy applies to public copy, interface labels, metadata, structured data, calculations, articles and corrections published by PerBiscuit.
Evidence and data classification
Every quantity should be identifiable as product-label data, third-party database data, user-reported PerBiscuit data, research data, a PerBiscuit calculation, a PerBiscuit assumption or an illustrative example.
Assumptions and calculations must not be described as observed consumption. A selected portion represents one input; it does not establish a population pattern.
Serving terminology
PerBiscuit uses label serving or serving used on the label. On US Nutrition Facts labels, serving size is a regulated calculation unit and is not dietary advice.
Comparison quantities are labelled as selected, reported, assumed or illustrative. Additional nutrition values are described relative to the label serving, not as undisclosed information.
Claims and interpretation
- Facts, interpretation and opinion are separated.
- Uncertain findings use proportionate language such as may, suggests or in the available data.
- PerBiscuit does not assign motives to manufacturers without direct evidence.
- Company conduct, ingredient disclosure, processing classification and portion agreement are separate questions.
- One product or portion is not presented as determining an individual's health.
Sources
Important scientific and regulatory claims should use primary sources. Product claims should identify the product, market, source and evidence-check date. Existing factual citations are preserved unless they are obsolete or do not support the claim.
Writing and humour
Copy uses plain English, short paragraphs and descriptive headings. Humour is occasional and restrained. It is excluded from methods, limitations, medical claims, source notes and regulatory explanations.
Review and transparency
Articles show authorship, review status, sources, methodology, evidence-check date, review dates, limitations and a corrections link. A missing reviewer or disclosure is marked pending rather than inferred.
Corrections
Material errors are corrected promptly and described on the corrections page. Editorial disagreements that do not change the evidence may be noted without presenting them as factual errors.
Policy history
First published and last reviewed July 14, 2026. Next scheduled review: October 14, 2026.