Research Guides
A HEAD Check Is Not a Content Review
A link checker reports success. Can you now say the article contains a working citation, the correct canonical tag or the promised explanation? Not if the checker only requested response metadata. Before using a technical result in a research report, establish what request was actually made.
This is a focused worksheet for authorized checks of public pages. It does not describe a built-in Colquhoun.shop feature or certify a site's search visibility.
1. Identify the method behind the result
RFC 9110 specifies that a HEAD response has no content. It is used to obtain metadata about a selected representation without transferring its data. Servers should generally return the headers they would send for GET, but may omit fields whose values are determined only while generating content. RFC 9110, section 9.3.2: HEAD
Record whether your tool used HEAD or GET. Also record the requested URL, time, final destination and whether it followed redirects. “URL checked” is too vague when another reviewer needs to reproduce the observation.
2. Match the evidence to the question
A HEAD result can support a narrow statement about the status and headers observed for that request. It does not supply HTML to inspect. Questions about an article's wording, an HTML canonical element or an in-page citation require content evidence.
For an invented audit, separate “HEAD returned 200” from “GET returned the expected article and its citation opened the intended source.” These are two observations, not interchangeable labels for a single pass. Even retrieving HTML does not establish that every script-dependent feature works in a browser.
3. Investigate disagreement without inventing a cause
If HEAD and GET results differ during an authorized comparison, retain both records. Check that the URL, redirect policy and relevant request context were comparable. Do not immediately conclude that the page is broken for visitors or that the server is deliberately misleading a crawler.
A missing header in HEAD is not automatically proof that it is missing from GET. Conversely, a successful HEAD response is not proof that GET delivers the expected body. State the mismatch first, then investigate with a small, bounded check. Stop on access restrictions or rate limits rather than changing identities or repeatedly retrying.
4. Review the actual content only when needed
Use a permitted GET request or ordinary page visit for the specific content question. Confirm that the response is the intended article rather than a generic page, challenge or sign-in screen. Save the minimum public evidence needed to explain the finding, not session secrets or unrelated personal information.
If the request is rate-limited, mark the content assessment incomplete. The companion rate-limit coverage worksheet explains why an uncollected page should not become a confirmed content failure.
5. Report the request and the conclusion separately
- Question: availability metadata, HTML content or interactive behavior?
- Request: method, public URL, timestamp and redirect policy.
- Observed: status, relevant headers and content inspected, if any.
- Conclusion: the narrow finding those observations support.
- Not tested: remaining content, rendering or interaction checks.
The useful outcome is not a longer checklist of green checks. It is a report where readers can distinguish a metadata probe from a content review and know what still needs verification.