Editorial policy

Why this page exists

Honey is one of the most claim-saturated foods there is. "Manuka kills superbugs", "raw honey cures allergies", "honey never expires", "honey is healthier than sugar" — each has a real mechanism somewhere behind it and each is routinely stated far past what the evidence supports. A honey site without an explicit evidence policy does not stay neutral by default; it drifts into repeating whatever the sellers say.

The evidence ladder

Every substantive claim on this site is placed on one of five tiers, and the tier is always rendered as visible text rather than as a colour or an icon.

The rule that does the most work

Claims on the top two tiers must also state what they are not claiming. This is the field that prevents most of the damage, because the failure mode here is rarely saying something false — it is saying something true that a reader correctly summarises into something false.

"Honey reduces cough frequency in children over one year, in a self-limiting illness" is supportable. The sentence a reader constructs from it — "honey treats coughs" — is not. Writing the boundary down is cheaper than defending the overreach later.

What we do not do

• No numeric confidence scores. A single 0-1 number invites false precision and conceals exactly the uncertainty the model exists to surface. There is none anywhere in the data model and none may be added.
• No medical advice. The health section describes what the evidence supports and where it stops. It is not a substitute for a clinician.
• No brands, no shop, no affiliate links. Honey records carry no brand, no price, no jar size and no seller, and the schema has no field for any of them. The moment a honey record can name a brand, every sensory field becomes marketing copy.
• No ratings or review markup. We publish no scores, so no Review or AggregateRating structured data is emitted, and no Product markup on honey pages — that would invite a rich result promising a transaction that does not exist.
• No superlatives about ourselves."Definitive", "most comprehensive" and "number one" are unsubstantiated claims, and a site whose whole argument is evidence discipline cannot open with one.

Traditional use is recorded, not endorsed

A great deal of what is interesting about honey is cultural, and removing it would make the site poorer and less honest. So traditional uses are recorded as traditional uses — the tier says that people have done something and for how long, and explicitly does not say that it works. Cultural belief gets the same treatment.

Unsupported claims are published, not deleted

The weakest tier exists and is rendered. Removing a popular false claim from the page does not remove it from the reader's head — it just means they read the correction somewhere less careful. Where a belief has a real origin, the page says where it came from, because "this is false" persuades nobody who has heard the claim their whole life.

Numbers are ranges

Every physical property on this site is stated as a range. "Acacia honey" is not one substance — it varies by region, by season and by year — so a single figure would assert a precision nobody has. Where a value is genuinely unknown, the page says so rather than leaving a blank, because an empty cell reads as zero or as an oversight.

Source independence

Different URLs are never assumed to be independent sources. Honey health claims are republished across thousands of retailer pages tracing back to one small study, and independence-blindness is precisely how a marketing line acquires the appearance of consensus. Sources declaring themselves republished, aggregated or translated must name their original.

Corrections

Records carry a review date. If something here is wrong, we would like to know — see contact. Corrections are made to the record, and the review date moves.