Evidence standards
HobbyHQ covers everything from bobbin lace to open water swimming, and those claims do not all need the same kind of backing. Here is the rule we apply, and where we currently fall short of it.
“A regulation archery round at 70m uses a 122cm face” is a fact owned by a governing body, and only the governing body is an acceptable source for it. “Most people find a 4mm hook easiest to learn crochet on” is craft convention, and a reputable specialist publisher is a perfectly good source. Treating those two the same either makes the second impossible to write or makes the first unreliable.
So each kind of claim states what standard of source it needs, and that is enforced rather than trusted. A claim requiring a governing-body source cannot be backed by a blog alone.
| Kind of claim | Acceptable sources |
|---|---|
| RulesCompetition rules, dimensions, formats and classifications. | Governing body, Standards body |
| LegalLaw, licensing, permissions, land access and protected status. | Governing body, Standards body, Institutional |
| SafetyAnything where being wrong could hurt somebody. | Governing body, Standards body, Scientific, Institutional |
| TechnicalMaterial behaviour, tolerances, physical and chemical properties. | Standards body, Scientific, Manufacturer, Institutional |
| EquipmentEquipment specifications and measurements. | Standards body, Manufacturer, Governing body, Specialist publisher |
| TechniqueHow something is done, as practised. | Specialist publisher, Expert practitioner, Institutional, Governing body |
| HistoryOrigins, dates, development and attribution. | Institutional, Scientific, Specialist publisher |
| ConventionCustom, etiquette and what is normal in a community. | Specialist publisher, Expert practitioner, Governing body, Institutional, Standards body |
| MarketPrices, participation figures and club counts. | Governing body, Institutional, Specialist publisher, Scientific |
Every source carries a verification state, the date it was last checked, and — where an editor opened it — the date it was read and what the reading established. There are six states, because checking thirty-four sources properly produced six genuinely different outcomes, and reducing them to a simple yes or no would have thrown most of that away.
166 of 189 have been opened and read; 167 are in a state that lets a claim rest on them. The rest are recorded honestly: several publishers block automated requests, so their content was confirmed from their own material by another route, and a few could not be reached at all.
- Read — supports the claim151 sources
- An editor opened it and confirmed it says what we cite it for.
- Read — partly supports15 sources
- An editor opened it and it supports the claim only in part — usually an index page that points at the authoritative document rather than containing it.
- Corroborated indirectly16 sources
- The publisher blocks automated requests, so the cited page could not be opened directly, but the content was confirmed from that publisher’s own material by another route.
- Could not be reached7 sources
- Could not be reached at the last check. Retained where the organisation is genuinely the right authority, but nothing rests on it alone.
Checking also removes things. 11 sources were taken out — dead URLs, a domain that never existed, and landing pages standing in for the documents they index. They are listed on the sources page rather than deleted, because a bibliography that silently drops its failures looks exactly like one that never checked.
Claims, not just publishers
Sources answer “is this publisher any good?”. The claim register answers the question that matters more: does the sentence on the page rest on anything?
Every published safety or legal warning is recorded with the exact sentence, the sources behind it, the country or jurisdiction where it holds, and the date it was checked. A warning missing any of those cannot be published.
162 claims are registered. 34 of them are recorded as not established — places where we looked, could not confirm what an earlier draft asserted, and rewrote the page to say less rather than leaving the assertion standing. Those are the records worth reading.
- Search volumes. Our prioritisation model contains none. Nobody has run a keyword tool against these subjects, and an invented figure would be indistinguishable from a measured one.
- Health claims. Several activities here are widely described as good for mental or physical health. Where we have not assessed the underlying research, we say what is defensible — that many people find it useful — and explicitly decline to summarise a literature we have not read.
- Precise legal advice. We map which duties exist and who owns each one, with links to the primary sources. Law differs across the UK, is amended, and turns on facts we do not know about your situation.
- Invented history. Where an origin story is disputed or apocryphal — and several in this catalogue are — we say that it is disputed rather than picking the better anecdote.
- Rankings without criteria. No “best hobby”, no scored league table of activities, no “our pick” in comparisons.
Rules, prices and law all date. Each source records how quickly it is expected to change, so a review can be prioritised rather than done alphabetically. Anything competition-critical, and anything with a legal consequence, should be checked against the current edition of the source rather than against this site — we say so on the pages where it matters.
See every source we cite, with its tier, its verification state and what it supports — or tell us a source does not support what we said, which is the most useful report we receive.