Editorial Standards

The core rule

Nothing is cited that has not been retrieved and read.

Not a title found in a search result. Not a claim repeated in a blog post. Not a study described in another study’s introduction. If we have not opened the source, it does not appear.

This sounds obvious and it is routinely violated across the entire fitness internet. A substantial fraction of what “everyone knows” about neck training traces back to a misremembering of something that, when you finally locate it, does not say that.

For scientific literature

Every entry in the research library records the authors, year, title, journal, and a DOI or PubMed identifier, and each is verified live before publication.

Each is also screened for retraction. A site whose entire proposition is accuracy cannot cite a withdrawn paper.

Where we have read only an abstract rather than a full text, the entry says so. This distinction matters enormously and is almost never made. An abstract omits sample size problems, dropout rates, and the limitations section — which is frequently where the most important sentence in a paper lives.

Study entries include limitations, not only findings. Where a result is preliminary, small, unreplicated, or comes from a low-tier journal, the entry says that.

For historical material

Historical quotation carries a specific hazard that took us a while to identify, and it is worth explaining because it caused a real error on this site.

Automated retrieval tools often pass a document through a summarising layer before returning it. That layer will silently reconstruct damaged text into clean, plausible, readable prose. The result looks like a verbatim quotation and is not one. We published a quotation from Farmer Burns that had been produced this way; the underlying scan actually reads A sirong peck in ABSOLUTELY NECESSART. The substance was right. The words were invented.

The rules that now follow from that:

1. Primary sources are read as raw text, not through any summarising retrieval layer.
2. The exact source is recorded — archive identifier, filename, and character offsets — so any quotation can be independently checked.
3. Optical character recognition artefacts are preserved in our working files. On the published page, hyphenation introduced by the scanner’s line-wrapping is silently repaired; nothing else is. Where a genuine OCR error is quoted, it is marked.
4. Provenance is checked, not assumed. The Burns scan turned out to be photographs of printouts from a defunct website, not the 1913 book. An archive listing is not a guarantee of what you are actually looking at.
5. Where a scan is too degraded to quote responsibly, we paraphrase and say so. Earle Liederman’s 1924 neck chapter is currently in this category.

What gets omitted

Claims that cannot be verified are left out. Not softened, not hedged with “some say” — omitted.

This is why some pages on this site say less than competing pages elsewhere. It is also why several pages record that a widely repeated claim has no locatable source. The “twenty-inch neck” figures attributed to early strongmen are a standing example: they circulate constantly and no primary source for them has yet been found.

Recording an absence is a finding. It is often more useful than the claim would have been.

On medical and injury content

Clinical material is educational. It is not diagnosis, treatment, or prescription, and it does not substitute for examination by a qualified professional.

Concussion and head-impact content uses risk-reduction language only. The evidence supports neck strength as one modifiable factor among several. It does not support prevention, and this site does not imply prevention by phrasing, placement, or omission.

Named individuals are described using the public record — “documented in public footage,” or “as reported by,” with the outlet named. This site does not describe anyone as using or endorsing a product without documentation, regardless of what may be true.

On commercial pressure

This site is owned by a manufacturer, which is disclosed on every page and explained in full on the disclosure page. Comparative claims about any product, ours included, must trace to the public record, the manufacturer’s own published material, an independent review or study, or our own demonstrable testing.

Where evidence is absent, the page says the evidence is absent — including where that absence is commercially inconvenient. There is currently no trial comparing head harnesses against ring-and-track devices, and this site says so on the device taxonomy page rather than filling the gap with confident prose.

Corrections

When something here is wrong, we fix it and say that we fixed it. The page records that a correction was made. We do not silently edit, because a source that quietly revises its history cannot be checked, and a source that cannot be checked is worthless.

Send corrections to [email protected]. Point out where we are wrong, and we will either fix it or explain publicly why we think we are right.